From bwalton at opencsw.org Wed Feb 1 14:46:09 2012 From: bwalton at opencsw.org (Ben Walton) Date: Wed, 01 Feb 2012 08:46:09 -0500 Subject: [csw-maintainers] Archiving the OpenCSW lists at MARC? In-Reply-To: References: Message-ID: <1328103782-sup-4358@pinkfloyd.chass.utoronto.ca> Excerpts from Hank Leininger's message of Wed Feb 01 00:37:52 -0500 2012: Hi Hank, > I maintain the MARC mailing list archives, http://marc.info/ A very useful service! Thanks for running this. > I'd like to add some of the OpenCSW lists. Would you guys have any > objections? I don't have any, but I've CC'd the maintainers list. I don't expect they'll object either, but let's give them a chance to comment. > I'm thinking announce, devel, and users; any others you would > recommend? For instance, I'd be happy to add bug-notifications too, > but users of the bug-tracker website might not expect their > submissions to end up on a searchable archive, even though they are > already public at > http://lists.opencsw.org/pipermail/bug-notifications/ This sounds reasonable to me. Thanks -Ben -- Ben Walton Systems Programmer - CHASS University of Toronto C:416.407.5610 | W:416.978.4302 From dam at opencsw.org Wed Feb 1 14:52:08 2012 From: dam at opencsw.org (Dagobert Michelsen) Date: Wed, 1 Feb 2012 14:52:08 +0100 Subject: [csw-maintainers] Archiving the OpenCSW lists at MARC? In-Reply-To: <1328103782-sup-4358@pinkfloyd.chass.utoronto.ca> References: <1328103782-sup-4358@pinkfloyd.chass.utoronto.ca> Message-ID: <6649789B-A2BD-4F31-8445-4EC6BE9B89EA@opencsw.org> Hi, Am 01.02.2012 um 14:46 schrieb Ben Walton: > Excerpts from Hank Leininger's message of Wed Feb 01 00:37:52 -0500 2012: >> I maintain the MARC mailing list archives, http://marc.info/ > > A very useful service! Thanks for running this. > >> I'd like to add some of the OpenCSW lists. Would you guys have any >> objections? > > I don't have any, but I've CC'd the maintainers list. I don't expect > they'll object either, but let's give them a chance to comment. Excellent idea! Would it be possible to add the links of the archives to the footer of the email? Something like Archived under http://.... to be able to immediately reference a mail when searched through the local mailbox? Best regards -- Dago -- "You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process." - xkcd #896 From bwalton at opencsw.org Wed Feb 1 15:35:22 2012 From: bwalton at opencsw.org (Ben Walton) Date: Wed, 01 Feb 2012 09:35:22 -0500 Subject: [csw-maintainers] OpenCSW @ serverfault.com In-Reply-To: References: Message-ID: <1328106896-sup-6863@pinkfloyd.chass.utoronto.ca> Excerpts from Maciej (Matchek) Blizi?ski's message of Sat Jan 28 06:08:44 -0500 2012: > It's good to see our packages mentioned in howtos on the web. I > remember Passenger mentioned here already, I'm wondering if anybody > was working on a package with it? I thought someone was, maybe Dago? I use passenger, but not on Solaris. Thanks -Ben -- Ben Walton Systems Programmer - CHASS University of Toronto C:416.407.5610 | W:416.978.4302 From dam at opencsw.org Wed Feb 1 15:40:08 2012 From: dam at opencsw.org (Dagobert Michelsen) Date: Wed, 1 Feb 2012 15:40:08 +0100 Subject: [csw-maintainers] OpenCSW @ serverfault.com In-Reply-To: <1328106896-sup-6863@pinkfloyd.chass.utoronto.ca> References: <1328106896-sup-6863@pinkfloyd.chass.utoronto.ca> Message-ID: Hi Ben, Am 01.02.2012 um 15:35 schrieb Ben Walton: > Excerpts from Maciej (Matchek) Blizi?ski's message of Sat Jan 28 06:08:44 -0500 2012: >> It's good to see our packages mentioned in howtos on the web. I >> remember Passenger mentioned here already, I'm wondering if anybody >> was working on a package with it? > > I thought someone was, maybe Dago? I use passenger, but not on > Solaris. Indeed, but it was quite hard to compile and then we decided to use mongrel instead. Maybe it is easier now in the newer version. Best regards -- Dago -- "You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process." - xkcd #896 From ihsan at opencsw.org Wed Feb 1 16:28:21 2012 From: ihsan at opencsw.org (=?UTF-8?B?xLBoc2FuwqBEb8SfYW4=?=) Date: Wed, 01 Feb 2012 16:28:21 +0100 Subject: [csw-maintainers] Archiving the OpenCSW lists at MARC? In-Reply-To: References: Message-ID: <4F295A15.90707@opencsw.org> Hello Hank, Am 01.02.2012 06:37, schrieb Hank Leininger: > I maintain the MARC mailing list archives, http://marc.info/ > > I'd like to add some of the OpenCSW lists. Would you guys have any > objections? That would be really great. > I'm thinking announce, devel, and users; any others you would > recommend? For instance, I'd be happy to add bug-notifications too, > but users of the bug-tracker website might not expect their > submissions to end up on a searchable archive, even though they are > already public at > http://lists.opencsw.org/pipermail/bug-notifications/ I think this should be fine. The bug database is already searchable through Google and other search engines. Ihsan -- ihsan at dogan.ch http://blog.dogan.ch/ -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 260 bytes Desc: OpenPGP digital signature URL: From grzemba at contac-dt.de Wed Feb 1 08:30:18 2012 From: grzemba at contac-dt.de (Carsten Grzemba) Date: Wed, 01 Feb 2012 08:30:18 +0100 Subject: [csw-maintainers] Blogger says installation instructions are unclear In-Reply-To: <7490a8cd2918.4f28e9d8@contac-dt.de> References: <7490a8cd2918.4f28e9d8@contac-dt.de> Message-ID: <7490ac5a2dd7.4f28f81a@contac-dt.de> I am relatively new to OpenCSW and have gone through in several installations recent days. My experience on Solaris 10 was: # pkgadd -d http://get.opencsw.org/now has not worked for me. Because I have to use a proxy? Download and install it is simple and ok. There are detailed instructions on the next page: * it should or must updated v2.4 to v2.6.1 by pkgutil -u * the mirror must changed, if you do not wont use catalog testing * run pkgutil -U Perhaps we could put these two web pages together because Configuration and Upgrade is significant. Carsten Am 31.01.12, schrieb Peter FELECAN : > Peter Bonivart writes: > > > I've always thought it was overkill to have both announcements and > > news, one should be enough. We could then move the text from "Get > > started" right onto the first page in its own box. Most similar sites > > have bootstrap info right in your face, we should as well. > > Overkill and quite stale. I fully agree with your proposal. BTW, if we > cannot feed correctly the "news" or "announcements" sections it's way > better to remove them altogether. > -- > Peter > _______________________________________________ > maintainers mailing list > maintainers at lists.opencsw.org > https://lists.opencsw.org/mailman/listinfo/maintainers > .:: This mailing list's archive is public. ::. > > -- Carsten Grzemba Tel.:?? +49 3677 64740 Mobil: +49 171 9749479 Fax::?? +49 3677 6474111 Email: carsten.grzemba at contac-dt.de contac Datentechnik GmbH -------------- next part -------------- An HTML attachment was scrubbed... URL: From gadavis at opencsw.org Wed Feb 1 22:34:28 2012 From: gadavis at opencsw.org (Geoff Davis) Date: Wed, 1 Feb 2012 13:34:28 -0800 Subject: [csw-maintainers] OpenCSW @ serverfault.com In-Reply-To: <1328106896-sup-6863@pinkfloyd.chass.utoronto.ca> References: <1328106896-sup-6863@pinkfloyd.chass.utoronto.ca> Message-ID: <005FBAC7-DD61-4073-B725-447DDE0690C5@opencsw.org> I ended up responding to the question after I figured out how to hack my way into a working build. I suppose I could throw together a package. However, I don't think I'm going to actually end up using Passenger in production and thus have little incentive to maintain it going forward. On Feb 1, 2012, at 6:35 AM, Ben Walton wrote: > Excerpts from Maciej (Matchek) Blizi?ski's message of Sat Jan 28 06:08:44 -0500 2012: > >> It's good to see our packages mentioned in howtos on the web. I >> remember Passenger mentioned here already, I'm wondering if anybody >> was working on a package with it? > > I thought someone was, maybe Dago? I use passenger, but not on > Solaris. > > Thanks > -Ben > -- > Ben Walton > Systems Programmer - CHASS > University of Toronto > C:416.407.5610 | W:416.978.4302 > > _______________________________________________ > maintainers mailing list > maintainers at lists.opencsw.org > https://lists.opencsw.org/mailman/listinfo/maintainers > .:: This mailing list's archive is public. ::. From maciej at opencsw.org Thu Feb 2 11:17:29 2012 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Thu, 2 Feb 2012 10:17:29 +0000 Subject: [csw-maintainers] The dublin release: remaining tasks In-Reply-To: References: <20111013090121.GA15669@bender.opencsw.org> <20111013101714.GB15669@bender.opencsw.org> Message-ID: 2012/1/31 Maciej (Matchek) Blizi?ski : > openldap: Not done I built the legacy libs yesterday. Dago built the 2.4 libs too, but they aren't in unstable yet. Dago, should we push the 2.4 openldap libraries to unstable today? Then we could respin at least some of the openldap dependent packages to make them link against the specific libraries and not the legacy runtime package (although that's optional for the dublin release). From maciej at opencsw.org Thu Feb 2 17:28:16 2012 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Thu, 2 Feb 2012 16:28:16 +0000 Subject: [csw-maintainers] OpenCSW Google+ page Message-ID: I've created a Google+ page for our project: https://plus.google.com/101553438503824697034 Those of you who use Google+, can circle this profile, and share posts with it. The page is currently managed mostly by me, but Ben and Ihsan also have the necessary rights. Dago and Sebastian will have the rights once they accept the invitations. If you have any links that are relevant to OpenCSW, but you don't want to send them to maintainers@ for any reason, you can share them with our OpenCSW page, and I'll reshare them. Especially if you've built a new set of packages and you would like to make the world aware of them, sharing that information on Google+ is a good idea. Maciej From maciej at opencsw.org Thu Feb 2 17:35:43 2012 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Thu, 2 Feb 2012 16:35:43 +0000 Subject: [csw-maintainers] error: This intrinsic requires target that supports appropriate instruction set Message-ID: Hello maintainers, I'm getting the following error when compiling ghostscript: /opt/SUNWspro/bin/cc -DHAVE_MKSTEMP -DHAVE_FILE64 -DHAVE_MKSTEMP64 -DHAVE_FONTCONFIG -DHAVE_LIBIDN -DHAVE_SETLOCALE -DHAVE_SSE2 -DHAVE_DBUS -KPIC -O2 -I/opt/csw/include -DHAVE_STDINT_H -DGX_COLOR_INDEX_TYPE="unsigned long long" -xO3 -m32 -xarch=pentium_proa -DGS_DEVS_SHARED -DGS_DEVS_SHARED_DIR=\"/opt/csw/lib/ghostscript/9.04\" -I./obj -I./base -DWHICH_CMS="lcms" -o ./obj/gxidata.o -c ./base/gxidata.c "./base/gxht_thresh.c", line 288: warning: attribute "align" is unknown, ignored "./base/gxht_thresh.c", line 130: [threshold_16_SSE]:ube: error: This intrinsic requires target that supports appropriate instruction set "./base/gxht_thresh.c", line 133: [threshold_16_SSE]:ube: error: This intrinsic requires target that supports appropriate instruction set "./base/gxht_thresh.c", line 134: [threshold_16_SSE]:ube: error: This intrinsic requires target that supports appropriate instruction set "./base/gxht_thresh.c", line 137: [threshold_16_SSE]:ube: error: This intrinsic requires target that supports appropriate instruction set "./base/gxht_thresh.c", line 138: [threshold_16_SSE]:ube: error: This intrinsic requires target that supports appropriate instruction set "./base/gxht_thresh.c", line 140: [threshold_16_SSE]:ube: error: This intrinsic requires target that supports appropriate instruction set "./base/gxht_thresh.c", line 142: [threshold_16_SSE]:ube: error: This intrinsic requires target that supports appropriate instruction set cc: ube failed for ./base/gxht_thresh.c gmake: *** [obj/gxht_thresh.o] Error 2 gmake: *** Waiting for unfinished jobs.... gmake: Leaving directory `/home/maciej/src/opencsw/pkg/ghostscript/trunk/work/solaris10-i386/build-isa-pentium_pro+mmx/ghostscript-9.04' gmake[1]: *** [build-work/solaris10-i386/build-isa-pentium_pro+mmx/ghostscript-9.04/Makefile] Error 2 gmake[1]: Leaving directory `/home/maciej/src/opencsw/pkg/ghostscript/trunk' gmake: *** [merge-isa-pentium_pro+mmx] Error 2 This happens on unstable9x and unstable10x. Before I sink a lot of time into this, did you see this kind of a problem before? Maciej From dam at opencsw.org Thu Feb 2 21:30:10 2012 From: dam at opencsw.org (Dagobert Michelsen) Date: Thu, 2 Feb 2012 21:30:10 +0100 Subject: [csw-maintainers] OpenCSW Google+ page In-Reply-To: References: Message-ID: Hi Maciej, Am 02.02.2012 um 17:28 schrieb Maciej (Matchek) Blizi?ski: > I've created a Google+ page for our project: > > https://plus.google.com/101553438503824697034 > > Those of you who use Google+, can circle this profile, and share posts with it. > > The page is currently managed mostly by me, but Ben and Ihsan also > have the necessary rights. Dago and Sebastian will have the rights > once they accept the invitations. I didn't get an invitation. What do I need to accept? Best regards -- Dago From raos at opencsw.org Fri Feb 3 10:32:17 2012 From: raos at opencsw.org (Rafael Ostertag) Date: Fri, 3 Feb 2012 10:32:17 +0100 Subject: [csw-maintainers] error: This intrinsic requires target that supports appropriate instruction set In-Reply-To: References: Message-ID: <20120203093217.GD12665@bender.opencsw.org> Hi Maciej On Thu, Feb 02, 2012 at 04:35:43PM +0000, Maciej (Matchek) Blizi??ski wrote: > Hello maintainers, > > I'm getting the following error when compiling ghostscript: > > /opt/SUNWspro/bin/cc -DHAVE_MKSTEMP -DHAVE_FILE64 -DHAVE_MKSTEMP64 > -DHAVE_FONTCONFIG -DHAVE_LIBIDN -DHAVE_SETLOCALE -DHAVE_SSE2 > -DHAVE_DBUS -KPIC -O2 -I/opt/csw/include -DHAVE_STDINT_H > -DGX_COLOR_INDEX_TYPE="unsigned long long" -xO3 -m32 > -xarch=pentium_proa -DGS_DEVS_SHARED > -DGS_DEVS_SHARED_DIR=\"/opt/csw/lib/ghostscript/9.04\" -I./obj > -I./base -DWHICH_CMS="lcms" -o ./obj/gxidata.o -c ./base/gxidata.c > "./base/gxht_thresh.c", line 288: warning: attribute "align" is > unknown, ignored > "./base/gxht_thresh.c", line 130: [threshold_16_SSE]:ube: error: This > intrinsic requires target that supports appropriate instruction set > "./base/gxht_thresh.c", line 133: [threshold_16_SSE]:ube: error: This > intrinsic requires target that supports appropriate instruction set > "./base/gxht_thresh.c", line 134: [threshold_16_SSE]:ube: error: This > intrinsic requires target that supports appropriate instruction set > "./base/gxht_thresh.c", line 137: [threshold_16_SSE]:ube: error: This > intrinsic requires target that supports appropriate instruction set > "./base/gxht_thresh.c", line 138: [threshold_16_SSE]:ube: error: This > intrinsic requires target that supports appropriate instruction set > "./base/gxht_thresh.c", line 140: [threshold_16_SSE]:ube: error: This > intrinsic requires target that supports appropriate instruction set > "./base/gxht_thresh.c", line 142: [threshold_16_SSE]:ube: error: This > intrinsic requires target that supports appropriate instruction set > cc: ube failed for ./base/gxht_thresh.c > gmake: *** [obj/gxht_thresh.o] Error 2 > gmake: *** Waiting for unfinished jobs.... > gmake: Leaving directory > `/home/maciej/src/opencsw/pkg/ghostscript/trunk/work/solaris10-i386/build-isa-pentium_pro+mmx/ghostscript-9.04' > gmake[1]: *** [build-work/solaris10-i386/build-isa-pentium_pro+mmx/ghostscript-9.04/Makefile] > Error 2 > gmake[1]: Leaving directory `/home/maciej/src/opencsw/pkg/ghostscript/trunk' > gmake: *** [merge-isa-pentium_pro+mmx] Error 2 > > This happens on unstable9x and unstable10x. > > Before I sink a lot of time into this, did you see this kind of a > problem before? Well, to me it appears that -DHAVE_SEE2 and -xarch=pentium_proa do not go well together. AFAIK, Pentium Pro does not support SSE2. Maybe that's the culprit. Cheers rafi From skayser at opencsw.org Sat Feb 4 06:46:13 2012 From: skayser at opencsw.org (Sebastian Kayser) Date: Sat, 4 Feb 2012 06:46:13 +0100 (CET) Subject: [csw-maintainers] Archiving the OpenCSW lists at MARC? In-Reply-To: <6649789B-A2BD-4F31-8445-4EC6BE9B89EA@opencsw.org> References: <1328103782-sup-4358@pinkfloyd.chass.utoronto.ca> <6649789B-A2BD-4F31-8445-4EC6BE9B89EA@opencsw.org> Message-ID: <52080.207.81.171.169.1328334373.squirrel@ssl.skayser.de> Hey Dago, Dagobert Michelsen schrieb: > Am 01.02.2012 um 14:46 schrieb Ben Walton: >> Excerpts from Hank Leininger's message of Wed Feb 01 00:37:52 -0500 >> 2012: >>> I maintain the MARC mailing list archives, http://marc.info/ >> >> A very useful service! Thanks for running this. >> >>> I'd like to add some of the OpenCSW lists. Would you guys have any >>> objections? >> >> I don't have any, but I've CC'd the maintainers list. I don't expect >> they'll object either, but let's give them a chance to comment. > > Excellent idea! Would it be possible to add the links of the archives to > the > footer of the email? Something like > Archived under http://.... > to be able to immediately reference a mail when searched through the local > mailbox? I asked a similar question on mailman-users a while ago. There seems to be way to hack this sort of feature into Mailman 2. Mailman 3 (no idea on the ETA) should have it integrated. http://mail.python.org/pipermail/mailman-users/2011-October/072379.html http://wiki.list.org/x/RAKJ Sebastian From bonivart at opencsw.org Sun Feb 5 17:20:29 2012 From: bonivart at opencsw.org (Peter Bonivart) Date: Sun, 5 Feb 2012 17:20:29 +0100 Subject: [csw-maintainers] Website suggestion Message-ID: My annotation skills are not that great but it's simple to explain: - we don't need both announcements and news - take the announcement "box", divide it into two - use one for a summary of the project which we lack on the first page (why have to click further to understand what OpenCSW is?) - use the other for "get started" info which similar projects also put right on the first page - I'm ok with the activity overview, news and releases & updates, let's start with changing what needs changing Do you agree? Who can change it? /peter -------------- next part -------------- A non-text attachment was scrubbed... Name: opencsw1.png Type: image/png Size: 160206 bytes Desc: not available URL: From maciej at opencsw.org Sun Feb 5 18:05:58 2012 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Sun, 5 Feb 2012 17:05:58 +0000 Subject: [csw-maintainers] The dublin release: remaining tasks In-Reply-To: References: <20111013090121.GA15669@bender.opencsw.org> <20111013101714.GB15669@bender.opencsw.org> Message-ID: With OpenLDAP rebuilt, we should be now ready to integrate from unstable to testing (== dublin). Does anyone know about anything seriously broken in unstable? Or any other reason not to integrate right now? If so, please respond. If no one responds by Tuesday evening, we will do the big push. Maciej From bwalton at opencsw.org Sun Feb 5 18:19:34 2012 From: bwalton at opencsw.org (Ben Walton) Date: Sun, 05 Feb 2012 12:19:34 -0500 Subject: [csw-maintainers] The dublin release: remaining tasks In-Reply-To: References: <20111013090121.GA15669@bender.opencsw.org> <20111013101714.GB15669@bender.opencsw.org> Message-ID: <1328462253-sup-3570@pinkfloyd.chass.utoronto.ca> Excerpts from Maciej (Matchek) Blizi?ski's message of Sun Feb 05 12:05:58 -0500 2012: > With OpenLDAP rebuilt, we should be now ready to integrate from > unstable to testing (== dublin). Does anyone know about anything > seriously broken in unstable? Or any other reason not to integrate > right now? If so, please respond. If no one responds by Tuesday > evening, we will do the big push. Php 5.3.9 is vulnerable to a remote exploit due to a botched fix for the hash collission problems that existed in 5.3.8. I'm working on 5.3.10 right now. Rafi noticed that the test suite had forked a ton of processes (eg: run amok) on unstable9x though. I've just initiated the test suite again and will try to keep an eye on it. Other than that, things are in good shape to the best of my knowledge. Thanks -Ben -- Ben Walton Systems Programmer - CHASS University of Toronto C:416.407.5610 | W:416.978.4302 From maciej at opencsw.org Mon Feb 6 13:01:38 2012 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Mon, 6 Feb 2012 12:01:38 +0000 Subject: [csw-maintainers] Website suggestion In-Reply-To: References: Message-ID: 2012/2/5 Peter Bonivart : > My annotation skills are not that great but it's simple to explain: > > - we don't need both announcements and news > - take the announcement "box", divide it into two > - use one for a summary of the project which we lack on the first page > (why have to click further to understand what OpenCSW is?) > - use the other for "get started" info which similar projects also put > right on the first page > - I'm ok with the activity overview, news and releases & updates, > let's start with changing what needs changing > > Do you agree? Who can change it? I agree, I'm certain most maintainers do too. The code is here: https://opencsw.svn.sourceforge.net/svnroot/opencsw/web/wordpresscsw/ You can submit your changes to it. Dago, Ben, Sebastian, I, and other people are able to deploy the code in production. Maciej From dam at opencsw.org Mon Feb 6 22:06:59 2012 From: dam at opencsw.org (Dagobert Michelsen) Date: Mon, 6 Feb 2012 22:06:59 +0100 Subject: [csw-maintainers] Website suggestion In-Reply-To: References: Message-ID: Hi, Am 06.02.2012 um 13:01 schrieb Maciej (Matchek) Blizi?ski: > 2012/2/5 Peter Bonivart : >> My annotation skills are not that great but it's simple to explain: >> >> - we don't need both announcements and news >> - take the announcement "box", divide it into two >> - use one for a summary of the project which we lack on the first page >> (why have to click further to understand what OpenCSW is?) >> - use the other for "get started" info which similar projects also put >> right on the first page >> - I'm ok with the activity overview, news and releases & updates, >> let's start with changing what needs changing >> >> Do you agree? Who can change it? > > I agree, I'm certain most maintainers do too. > > The code is here: > https://opencsw.svn.sourceforge.net/svnroot/opencsw/web/wordpresscsw/ > > You can submit your changes to it. > > Dago, Ben, Sebastian, I, and other people are able to deploy the code > in production. There is also the proposed layout at http://www-mockup.opencsw.org/ but I like the newly proposed one better. Best regards -- Dago -- "You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process." - xkcd #896 From pfelecan at opencsw.org Tue Feb 7 09:43:43 2012 From: pfelecan at opencsw.org (Peter FELECAN) Date: Tue, 07 Feb 2012 09:43:43 +0100 Subject: [csw-maintainers] Website suggestion In-Reply-To: (Dagobert Michelsen's message of "Mon, 6 Feb 2012 22:06:59 +0100") References: Message-ID: Dagobert Michelsen writes: > There is also the proposed layout at > http://www-mockup.opencsw.org/ > but I like the newly proposed one better. Which is? -- Peter From maciej at opencsw.org Tue Feb 7 10:56:06 2012 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Tue, 7 Feb 2012 09:56:06 +0000 Subject: [csw-maintainers] Website suggestion In-Reply-To: References: Message-ID: 2012/2/7 Peter FELECAN : > Dagobert Michelsen writes: > >> There is also the proposed layout at >> ? http://www-mockup.opencsw.org/ >> but I like the newly proposed one better. > > Which is? The two pane layout shown in an attachment in the first email in this thread. The most important thing is for us to be able to modify the website. Then we'll keep modifying it until we're happy. From pfelecan at opencsw.org Tue Feb 7 18:48:18 2012 From: pfelecan at opencsw.org (Peter FELECAN) Date: Tue, 07 Feb 2012 18:48:18 +0100 Subject: [csw-maintainers] Website suggestion In-Reply-To: ("Maciej (Matchek) =?utf-8?Q?Blizi=C5=84ski=22's?= message of "Tue, 7 Feb 2012 09:56:06 +0000") References: Message-ID: "Maciej (Matchek) Blizi?ski" writes: > 2012/2/7 Peter FELECAN : >> Dagobert Michelsen writes: >> >>> There is also the proposed layout at >>> ? http://www-mockup.opencsw.org/ >>> but I like the newly proposed one better. >> >> Which is? > > The two pane layout shown in an attachment in the first email in this thread. Oh, that! As I use a text only mail reader I haven't opened the attachment... What a efficient secretary we have! Tx > The most important thing is for us to be able to modify the website. > Then we'll keep modifying it until we're happy. The layout proposed by Peter is very alright emphasizing the right stuff and suppressing the clunky cruft. -- Peter From maciej at opencsw.org Thu Feb 9 12:30:19 2012 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Thu, 9 Feb 2012 11:30:19 +0000 Subject: [csw-maintainers] The dublin release: remaining tasks In-Reply-To: <1328462253-sup-3570@pinkfloyd.chass.utoronto.ca> References: <20111013090121.GA15669@bender.opencsw.org> <20111013101714.GB15669@bender.opencsw.org> <1328462253-sup-3570@pinkfloyd.chass.utoronto.ca> Message-ID: 2012/2/5 Ben Walton : > Excerpts from Maciej (Matchek) Blizi?ski's message of Sun Feb 05 12:05:58 -0500 2012: > >> With OpenLDAP rebuilt, we should be now ready to integrate from >> unstable to testing (== dublin). Does anyone know about anything >> seriously broken in unstable? Or any other reason not to integrate >> right now? If so, please respond. If no one responds by Tuesday >> evening, we will do the big push. > > Php 5.3.9 is vulnerable to a remote exploit due to a botched fix for > the hash collission problems that existed in 5.3.8. ?I'm working on > 5.3.10 right now. ?Rafi noticed that the test suite had forked a ton > of processes (eg: run amok) on unstable9x though. ?I've just initiated > the test suite again and will try to keep an eye on it. Keep me posted about the php progress. We probably don't want to release 5.3.9 into dublin. Is version 5.2.9 also vulnerable? Maciej From maciej at opencsw.org Thu Feb 9 13:24:07 2012 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Thu, 9 Feb 2012 12:24:07 +0000 Subject: [csw-maintainers] The dublin release: remaining tasks In-Reply-To: References: <20111013090121.GA15669@bender.opencsw.org> <20111013101714.GB15669@bender.opencsw.org> <1328462253-sup-3570@pinkfloyd.chass.utoronto.ca> Message-ID: Heads up: I'm integrating unstable to dublin, skipping the php5 package family. From maciej at opencsw.org Thu Feb 9 13:51:11 2012 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Thu, 9 Feb 2012 12:51:11 +0000 Subject: [csw-maintainers] Update to the testing / dublin / current catalog In-Reply-To: References: Message-ID: [a repost from users@] Hello everyone, You probably have noticed, that the testing / dublin / current catalog has not been receiving updates for some time now. All the work and churn has been confined to the unstable catalog, saving you hassle with the constant stream of updates. However, since the works are progressing at the normal pace, the testing / dublin / current catalog (Solaris 9 and up) will soon get a large, accumulated update. The update migrates all updated packages from the unstable catalog to testing / dublin. If you want additional safety during the upgrade, you can temporarily switch one of your machines to the unstable catalog and run pkgutil -y -u. If everything goes well, you can update the other machines using the dublin catalog. To learn more about the unstable and testing catalogs, you can read this news item: http://www.opencsw.org/2012/01/which-catalog-should-i-use/ We expect upgrades to go smoothly, but in case of any problems, don't hesitate to contact us: either here, or at #opencsw on Freenode. Regards, Maciej Blizi?ski From maciej at opencsw.org Fri Feb 10 11:53:03 2012 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Fri, 10 Feb 2012 10:53:03 +0000 Subject: [csw-maintainers] Statistics page out of date Message-ID: http://www.opencsw.org/get-it/package-statistics/ This page shows updates up to September 2012. Who knows how it works and what do we do to bring it up to date? Maciej From maciej at opencsw.org Fri Feb 10 12:00:55 2012 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Fri, 10 Feb 2012 11:00:55 +0000 Subject: [csw-maintainers] The dublin release: remaining tasks In-Reply-To: References: <20111013090121.GA15669@bender.opencsw.org> <20111013101714.GB15669@bender.opencsw.org> <1328462253-sup-3570@pinkfloyd.chass.utoronto.ca> Message-ID: The change is now live. If there are any bugs, we're likely to hear about them now, rather than some other time. Let's keep an eye in the IRC channel, users mailing list and mantis. From skayser at opencsw.org Fri Feb 10 12:29:05 2012 From: skayser at opencsw.org (Sebastian Kayser) Date: Fri, 10 Feb 2012 12:29:05 +0100 Subject: [csw-maintainers] Statistics page out of date In-Reply-To: References: Message-ID: <20120210112905.GP19520@sebastiankayser.de> * Maciej (Matchek) Blizi??ski wrote: > http://www.opencsw.org/get-it/package-statistics/ > > This page shows updates up to September 2012. Who knows how it works > and what do we do to bring it up to date? Not sure which jobs are responsible for the page, but I gave it a quick stab. William has a couple of cronjobs running on mockup and his $HOME was exceeding quota, which in turn made cronjobs fall flat. I temporarily freed up some space and triggered his statistics cronjob on mockup which now fails with perl symbol resolution errors. $ cd /home/wbonnet/packageStatistics ; /home/wbonnet/packageStatistics/processUpdateStatistics.sh ... ld.so.1: perl: fatal: relocation error: file /opt/csw/lib/perl/csw/auto/DBI/DBI.so: symbol Perl_Tstack_sp_ptr: referenced symbol not found /home/wbonnet/packageStatistics/processUpdateStatistics.sh: line 114: 2360 Killed ./updatePackagesStatistics.pl --command=updateFromPkgGetCatalog --catalog=${CATALOG_TODAY_NAME} ${1} ${2} ${3} ${4} Someone more knowledgable please take it from here. Sebastian From skayser at opencsw.org Fri Feb 10 12:56:09 2012 From: skayser at opencsw.org (Sebastian Kayser) Date: Fri, 10 Feb 2012 12:56:09 +0100 Subject: [csw-maintainers] Statistics page out of date In-Reply-To: <20120210112905.GP19520@sebastiankayser.de> References: <20120210112905.GP19520@sebastiankayser.de> Message-ID: <20120210115609.GQ19520@sebastiankayser.de> * Sebastian Kayser wrote: > * Maciej (Matchek) Blizi??ski wrote: > > http://www.opencsw.org/get-it/package-statistics/ > > > > This page shows updates up to September 2012. Who knows how it works > > and what do we do to bring it up to date? > > Not sure which jobs are responsible for the page, ... > Someone more knowledgable please take it from here. Also, if the benevolent helper is possibly looking for yet another good deed, the infrastructure docs @ http://opencsw-orga.wikidot.com/ appreciate all the love that they can get. Stats jobs yet unmentioned. Sebastian From bonivart at opencsw.org Fri Feb 10 13:19:52 2012 From: bonivart at opencsw.org (Peter Bonivart) Date: Fri, 10 Feb 2012 13:19:52 +0100 Subject: [csw-maintainers] Statistics page out of date In-Reply-To: <20120210112905.GP19520@sebastiankayser.de> References: <20120210112905.GP19520@sebastiankayser.de> Message-ID: On Fri, Feb 10, 2012 at 12:29 PM, Sebastian Kayser wrote: > $ cd /home/wbonnet/packageStatistics ; /home/wbonnet/packageStatistics/processUpdateStatistics.sh ... > ld.so.1: perl: fatal: relocation error: file /opt/csw/lib/perl/csw/auto/DBI/DBI.so: symbol Perl_Tstack_sp_ptr: referenced symbol not found > /home/wbonnet/packageStatistics/processUpdateStatistics.sh: line 114: ?2360 Killed ? ? ? ? ? ? ? ? ?./updatePackagesStatistics.pl --command=updateFromPkgGetCatalog --catalog=${CATALOG_TODAY_NAME} ${1} ${2} ${3} ${4} This looks like the DBI module is out of sync with the main Perl package. An update should resolve it. /peter From bwalton at opencsw.org Sat Feb 11 01:54:11 2012 From: bwalton at opencsw.org (Ben Walton) Date: Fri, 10 Feb 2012 19:54:11 -0500 Subject: [csw-maintainers] The dublin release: remaining tasks In-Reply-To: References: <20111013090121.GA15669@bender.opencsw.org> <20111013101714.GB15669@bender.opencsw.org> <1328462253-sup-3570@pinkfloyd.chass.utoronto.ca> Message-ID: <1328921545-sup-4762@pinkfloyd.chass.utoronto.ca> Excerpts from Maciej (Matchek) Blizi?ski's message of Thu Feb 09 06:30:19 -0500 2012: > Keep me posted about the php progress. We probably don't want to > release 5.3.9 into dublin. Is version 5.2.9 also vulnerable? AFAIK, 5.2.9 is vulnerable to older bugs (possibly including the hash table issue that prompted the 5.3.9 fixes). I just pushed 5.3.10 to unstable. A day late and a dollar short, but it's there now! :) After a bit of incubation in unstable, we could promote it and then mark dublin as an official release? That would let us move on to kiel! Thanks -Ben -- Ben Walton Systems Programmer - CHASS University of Toronto C:416.407.5610 | W:416.978.4302 From maciej at opencsw.org Sat Feb 11 16:34:41 2012 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Sat, 11 Feb 2012 15:34:41 +0000 Subject: [csw-maintainers] The dublin release: remaining tasks In-Reply-To: <1328921545-sup-4762@pinkfloyd.chass.utoronto.ca> References: <20111013090121.GA15669@bender.opencsw.org> <20111013101714.GB15669@bender.opencsw.org> <1328462253-sup-3570@pinkfloyd.chass.utoronto.ca> <1328921545-sup-4762@pinkfloyd.chass.utoronto.ca> Message-ID: 2012/2/11 Ben Walton : > Excerpts from Maciej (Matchek) Blizi?ski's message of Thu Feb 09 06:30:19 -0500 2012: > >> Keep me posted about the php progress. We probably don't want to >> release 5.3.9 into dublin. Is version 5.2.9 also vulnerable? > > AFAIK, 5.2.9 is vulnerable to older bugs (possibly including the hash > table issue that prompted the 5.3.9 fixes). ?I just pushed 5.3.10 to > unstable. ?A day late and a dollar short, but it's there now! :) Cool. > After a bit of incubation in unstable, we could promote it and then > mark dublin as an official release? ?That would let us move on to > kiel! +1 From yann at pleiades.fr.eu.org Sat Feb 11 17:21:03 2012 From: yann at pleiades.fr.eu.org (Yann Rouillard) Date: Sat, 11 Feb 2012 17:21:03 +0100 Subject: [csw-maintainers] crypto-hardware acceleration for openssl Message-ID: <4F36956F.8010303@pleiades.fr.eu.org> Hi everybody, I put openssl 0.9.8 with crypto-hardware acceleration support in my experimental repository so then can be easily tested: pkgutil -t http://buildfarm.opencsw.org/opencsw/experimental/yann -U pkgutil -t http://buildfarm.opencsw.org/opencsw/experimental/yann -i openssl libssl0_9_8 libssl_dev I would welcome additionnal testing of the package. If everything goes well, I am planning to release the package in unstable, so the performances of opencsw ssl are not anymore lower than the ones provided with Solaris. You can test the performance with: openssl speed -engine pkcs11 rsa (or dsa or ...) I would also welcome some test with apache. Some instructions to setup Apache are in this file: http://www.oracle.com/technetwork/server-storage/archive/a11-014-crypto-accelerators-439765.pdf I don't have administrator access on an Ultrasparc T1/T2 server (I made the openssl test on unstable10s). For information, here is the results of the speed test under unstable10s: RSA with hardware acceleration: sign verify sign/s verify/s rsa 512 bits 0.000038s 0.000176s 26196.6 5697.5 rsa 1024 bits 0.000267s 0.000182s 3741.9 5502.1 rsa 2048 bits 0.000292s 0.000186s 3425.5 5386.1 rsa 4096 bits 0.698000s 0.018638s 1.4 53.7 RSA without hardware acceleration: sign verify sign/s verify/s rsa 512 bits 0.003153s 0.000253s 317.2 3947.4 rsa 1024 bits 0.014025s 0.000656s 71.3 1523.8 rsa 2048 bits 0.077385s 0.002087s 12.9 479.2 rsa 4096 bits 0.494762s 0.007289s 2.0 137.2 DSA with hardware acceleration: sign verify sign/s verify/s dsa 512 bits 0.002403s 0.002793s 416.2 358.1 dsa 1024 bits 0.006339s 0.007634s 157.8 131.0 dsa 2048 bits 0.020367s 0.023923s 49.1 41.8 DSA without hardware acceleration: sign verify sign/s verify/s dsa 512 bits 0.000063s 0.000053s 15764.1 18955.6 dsa 1024 bits 0.000063s 0.000053s 15842.6 19026.8 dsa 2048 bits 0.033831s 0.067315s 29.6 14.9 Thanks in advance, Yann From yann at pleiades.fr.eu.org Sat Feb 11 18:10:20 2012 From: yann at pleiades.fr.eu.org (Yann Rouillard) Date: Sat, 11 Feb 2012 18:10:20 +0100 Subject: [csw-maintainers] High Performance SSH (HPN - SSH) Message-ID: <4F36A0FC.9010906@pleiades.fr.eu.org> Hi everybody, I also compiled and put in my experimental repository an (experimental) openssh package compiled with the High Performance patches applied: http://www.psc.edu/networking/projects/hpn-ssh/ You can easily install them with the following commands: pkgutil -t http://buildfarm.opencsw.org/opencsw/experimental/yann -U pkgutil -t http://buildfarm.opencsw.org/opencsw/experimental/yann -i openssh openssh_client I made some tests and noted a sligh improvement in a localhost scp test with (30 MB/s -> 35 Mb/s). However according to the website, HPN SSH should not make a lot of difference in a local network but rather across the internet. If someone could help to do some more test, it would be great. Note that there is currently a problem with the multi-threaded cipher and forking so I will probably not enable this cipher in the end unless this problem is solved. Yann -------------- next part -------------- An HTML attachment was scrubbed... URL: From maciej at opencsw.org Sun Feb 12 16:53:50 2012 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Sun, 12 Feb 2012 15:53:50 +0000 Subject: [csw-maintainers] crypto-hardware acceleration for openssl In-Reply-To: <4F36956F.8010303@pleiades.fr.eu.org> References: <4F36956F.8010303@pleiades.fr.eu.org> Message-ID: 2012/2/11 Yann Rouillard : > I put openssl 0.9.8 with crypto-hardware acceleration support in my > experimental repository so then can be easily tested: > pkgutil -t http://buildfarm.opencsw.org/opencsw/experimental/yann -U > pkgutil -t http://buildfarm.opencsw.org/opencsw/experimental/yann -i openssl > libssl0_9_8 libssl_dev > > I would welcome additionnal testing of the package. > If everything goes well, I am planning to release the package in unstable, > so the performances of opencsw ssl are not anymore lower than the ones > provided with Solaris. These numbers look very good! It would be awesome to include these packages in the dublin catalog if we can get enough testing. I did a smoke test with: openssl s_client -connect www.google.com:443 ...and it worked. Has anyone else done any testing of these packages? Maciej From dam at opencsw.org Mon Feb 13 10:13:00 2012 From: dam at opencsw.org (Dagobert Michelsen) Date: Mon, 13 Feb 2012 10:13:00 +0100 Subject: [csw-maintainers] High Performance SSH (HPN - SSH) In-Reply-To: <4F36A0FC.9010906@pleiades.fr.eu.org> References: <4F36A0FC.9010906@pleiades.fr.eu.org> Message-ID: <1642F0A1-B2FF-421F-9F32-5A0291F4B0FF@opencsw.org> Hi Yann, Am 11.02.2012 um 18:10 schrieb Yann Rouillard: > I also compiled and put in my experimental repository an (experimental) openssh package compiled with the High Performance patches applied: > http://www.psc.edu/networking/projects/hpn-ssh/ > > You can easily install them with the following commands: > pkgutil -t http://buildfarm.opencsw.org/opencsw/experimental/yann -U > pkgutil -t http://buildfarm.opencsw.org/opencsw/experimental/yann -i openssh openssh_client > > I made some tests and noted a sligh improvement in a localhost scp test with (30 MB/s -> 35 Mb/s). However according to the website, HPN SSH should not make a lot of difference in a local network but rather across the internet. If someone could help to do some more test, it would be great. > > Note that there is currently a problem with the multi-threaded cipher and forking so I will probably not enable this cipher in the end unless this problem is solved. This is excellent news, especially with the recent enhancements to the OpenSSL package! Thanks Yann :-) Best regards -- Dago -- "You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process." - xkcd #896 -------------- next part -------------- An HTML attachment was scrubbed... URL: From pfelecan at opencsw.org Mon Feb 13 10:50:22 2012 From: pfelecan at opencsw.org (Peter FELECAN) Date: Mon, 13 Feb 2012 10:50:22 +0100 Subject: [csw-maintainers] crypto-hardware acceleration for openssl In-Reply-To: ("Maciej (Matchek) =?utf-8?Q?Blizi=C5=84ski=22's?= message of "Sun, 12 Feb 2012 15:53:50 +0000") References: <4F36956F.8010303@pleiades.fr.eu.org> Message-ID: "Maciej (Matchek) Blizi?ski" writes: > It would be awesome to include these packages in the dublin catalog if > we can get enough testing. *Enough testing* is the key concept. why hurry? Let it be in unstable and, if after 4 weeks there is no bug report, promote the packages to the next catalog. It seems to me that this is nearer to the life cycle on which we agreed. BTW, isn't there work on automatic promotion? -- Peter From bwalton at opencsw.org Mon Feb 13 15:16:55 2012 From: bwalton at opencsw.org (Ben Walton) Date: Mon, 13 Feb 2012 09:16:55 -0500 Subject: [csw-maintainers] crypto-hardware acceleration for openssl In-Reply-To: References: <4F36956F.8010303@pleiades.fr.eu.org> Message-ID: <1329142538-sup-5460@pinkfloyd.chass.utoronto.ca> Excerpts from Peter FELECAN's message of Mon Feb 13 04:50:22 -0500 2012: > on which we agreed. BTW, isn't there work on automatic promotion? Yes, but it's stalled at the moment. I ran out of cycles and didn't pick it back up yet. If someone else has the time, feel free to step in. It is on my todo list still but my work has been extra busy for some time leaving me fairly drained in the evenings. Thanks -Ben -- Ben Walton Systems Programmer - CHASS University of Toronto C:416.407.5610 | W:416.978.4302 From dam at opencsw.org Mon Feb 13 17:44:38 2012 From: dam at opencsw.org (Dagobert Michelsen) Date: Mon, 13 Feb 2012 17:44:38 +0100 Subject: [csw-maintainers] Reorganizing the webpage Message-ID: Hi folks, I just an email from someone I recommended OpenCSW to and got the response that OpenCSW looks dead when reading the news section at http://www.opencsw.org/category/recent-updates/ It is really time to move the mockup feed to the main site. Best regards -- Dago -- "You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process." - xkcd #896 From william at wbonnet.net Mon Feb 13 23:07:19 2012 From: william at wbonnet.net (William Bonnet) Date: Mon, 13 Feb 2012 23:07:19 +0100 Subject: [csw-maintainers] Statistics page out of date In-Reply-To: References: <20120210112905.GP19520@sebastiankayser.de> Message-ID: <4F398997.2070505@wbonnet.net> Hi all > On Fri, Feb 10, 2012 at 12:29 PM, Sebastian Kayser wrote: >> $ cd /home/wbonnet/packageStatistics ; /home/wbonnet/packageStatistics/processUpdateStatistics.sh ... >> ld.so.1: perl: fatal: relocation error: file /opt/csw/lib/perl/csw/auto/DBI/DBI.so: symbol Perl_Tstack_sp_ptr: referenced symbol not found >> /home/wbonnet/packageStatistics/processUpdateStatistics.sh: line 114: 2360 Killed ./updatePackagesStatistics.pl --command=updateFromPkgGetCatalog --catalog=${CATALOG_TODAY_NAME} ${1} ${2} ${3} ${4} > > This looks like the DBI module is out of sync with the main Perl > package. An update should resolve it. I was having a look to this problem and no longer see my crontab. Was it deactivated ? The archives directory (/home/wbonnet/packageStatistics/archives) has no files between september and now. Which means the cron job has not been running, or the archives were purged. I can rebuilt the statistics easily if you have an archive of catalog during this period of time. cheers W. -- William http://www.wbonnet.net http://www.opencsw.org Community SoftWare for Solaris From dam at opencsw.org Tue Feb 14 11:17:30 2012 From: dam at opencsw.org (Dagobert Michelsen) Date: Tue, 14 Feb 2012 11:17:30 +0100 Subject: [csw-maintainers] Statistics page out of date In-Reply-To: <4F398997.2070505@wbonnet.net> References: <20120210112905.GP19520@sebastiankayser.de> <4F398997.2070505@wbonnet.net> Message-ID: Hi William, Am 13.02.2012 um 23:07 schrieb William Bonnet: >> On Fri, Feb 10, 2012 at 12:29 PM, Sebastian Kayser wrote: >>> $ cd /home/wbonnet/packageStatistics ; /home/wbonnet/packageStatistics/processUpdateStatistics.sh ... >>> ld.so.1: perl: fatal: relocation error: file /opt/csw/lib/perl/csw/auto/DBI/DBI.so: symbol Perl_Tstack_sp_ptr: referenced symbol not found >>> /home/wbonnet/packageStatistics/processUpdateStatistics.sh: line 114: 2360 Killed ./updatePackagesStatistics.pl --command=updateFromPkgGetCatalog --catalog=${CATALOG_TODAY_NAME} ${1} ${2} ${3} ${4} >> >> This looks like the DBI module is out of sync with the main Perl >> package. An update should resolve it. > > I was having a look to this problem and no longer see my crontab. Was it deactivated ? I don't know, but Ihsan changed a few things in the past. Maybe we should transfer the scripts to a functional user like "webjobs" or something and monitor that. Additionally, the MySQL host was separated so the connect string is different (I adjust that for the stats). > The archives directory (/home/wbonnet/packageStatistics/archives) has no files between september and now. Which means the cron job has not been running, or the archives were purged. > > I can rebuilt the statistics easily if you have an archive of catalog during this period of time. Cool, the catalogs are all available on web at web [web]:/home/web/bin/catalogs I can copy them over if you tell me which catalogs are needed. Can you please also take a look at the qa/upstream_watch scripts? Could you also aid in documenting the layout and jobs here? http://opencsw-orga.wikidot.com/ Best regards -- Dago -- "You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process." - xkcd #896 From maciej at opencsw.org Tue Feb 14 18:11:16 2012 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Tue, 14 Feb 2012 17:11:16 +0000 Subject: [csw-maintainers] MySQL package names Message-ID: Quick summary of recent developments of the MySQL packages: - Proper support for multiple MySQL versions is hard - Therefore, we're better off succeeding to support one version of MySQL rather than trying and failing to support multiple versions - We have 5.0, which is old, 5.1 is old, even 5.5 is frankly quite old now. - We probably want to release 5.5, replacing 5.0 and skipping 5.1. - Our packages are currently named CSWmysql5 + friends When we release 5.5, we'll want to remove 5.0 from the catalog, or mark it incompatible. I think that in dublin, we'll ship 5.0, and in kiel we will ship only 5.5. Question: Do we want to remove the CSWmysql5 package family and create CSWmysql55 package family? Or is it better to put 5.5 into packages named CSWmysql5*? Maciej From bwalton at opencsw.org Tue Feb 14 18:23:18 2012 From: bwalton at opencsw.org (Ben Walton) Date: Tue, 14 Feb 2012 12:23:18 -0500 Subject: [csw-maintainers] MySQL package names In-Reply-To: References: Message-ID: <1329240135-sup-6569@pinkfloyd.chass.utoronto.ca> Excerpts from Maciej (Matchek) Blizi?ski's message of Tue Feb 14 12:11:16 -0500 2012: Hi Maciej, > Question: Do we want to remove the CSWmysql5 package family and > create CSWmysql55 package family? Or is it better to put 5.5 into > packages named CSWmysql5*? I think I'd stick with the current naming if there aren't going to be side-by-side versions. This could always be changed at a later date if required. Your outlined plan sounds good to me. Thanks -Ben -- Ben Walton Systems Programmer - CHASS University of Toronto C:416.407.5610 | W:416.978.4302 From pfelecan at opencsw.org Tue Feb 14 19:59:24 2012 From: pfelecan at opencsw.org (Peter FELECAN) Date: Tue, 14 Feb 2012 19:59:24 +0100 Subject: [csw-maintainers] MySQL package names In-Reply-To: ("Maciej (Matchek) =?utf-8?Q?Blizi=C5=84ski=22's?= message of "Tue, 14 Feb 2012 17:11:16 +0000") References: Message-ID: "Maciej (Matchek) Blizi?ski" writes: > Question: Do we want to remove the CSWmysql5 package family and create > CSWmysql55 package family? Or is it better to put 5.5 into packages > named CSWmysql5*? Support only 1 set per major release, in this case CSWmysql5 contains 5.5 and the related ones have the same prefix. my 2 oboloi -- Peter From maciej at opencsw.org Wed Feb 15 20:16:12 2012 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Wed, 15 Feb 2012 19:16:12 +0000 Subject: [csw-maintainers] MySQL package names In-Reply-To: References: Message-ID: 2012/2/14 Peter FELECAN : > "Maciej (Matchek) Blizi?ski" writes: > >> Question: Do we want to remove the CSWmysql5 package family and create >> CSWmysql55 package family? Or is it better to put 5.5 into packages >> named CSWmysql5*? > > Support only 1 set per major release, in this case CSWmysql5 contains > 5.5 and the related ones have the same prefix. Cool. This makes things easier. It makes them easier by such extent, that I already have a working set of MySQL 5.5 packages. maciej at testing10s [build10st]:~ > mysql Welcome to the MySQL monitor. Commands end with ; or \g. Your MySQL connection id is 1 Server version: 5.5.20 MySQL Community Server (GPL) Copyright (c) 2000, 2011, Oracle and/or its affiliates. All rights reserved. Oracle is a registered trademark of Oracle Corporation and/or its affiliates. Other names may be trademarks of their respective owners. Type 'help;' or '\h' for help. Type '\c' to clear the current input statement. mysql> From pfelecan at opencsw.org Wed Feb 15 20:26:05 2012 From: pfelecan at opencsw.org (Peter FELECAN) Date: Wed, 15 Feb 2012 20:26:05 +0100 Subject: [csw-maintainers] MySQL package names In-Reply-To: ("Maciej (Matchek) =?utf-8?Q?Blizi=C5=84ski=22's?= message of "Wed, 15 Feb 2012 19:16:12 +0000") References: Message-ID: "Maciej (Matchek) Blizi?ski" writes: > Cool. This makes things easier. It makes them easier by such extent, > that I already have a working set of MySQL 5.5 packages. Great! -- Peter From maciej at opencsw.org Thu Feb 16 15:40:21 2012 From: maciej at opencsw.org (Maciej (Matchek) Blizinski) Date: Thu, 16 Feb 2012 14:40:21 +0000 Subject: [csw-maintainers] Dublin release highlights Message-ID: Hello maintainers, With the upcoming dublin release, we'd like to write an article outlining its main highlights. I'd like to ask everyone to send in their highlights of the dublin release. Something that was significant and you'd like it mentioned. For example, I'd put forward the GCC update: version upgrade and moving it from /opt/csw/gcc4 to /opt/csw. We also had a collective effort to repackage shared libraries. I'm sure there's more and that every maintainer has something significant that they did in the last year. Please send them in, either on or off the list, by the end of the next week, which is the 24th of February. Maciej From maciej at opencsw.org Thu Feb 16 15:43:40 2012 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Thu, 16 Feb 2012 14:43:40 +0000 Subject: [csw-maintainers] Wintercamp 2011 In-Reply-To: <4F169DE1.10904@lutter.sk> References: <4F169DE1.10904@lutter.sk> Message-ID: Everyone, please update the doodle. http://doodle.com/scawm5zdt3x5tvgs If you don't intend to come, feel free to check off all the weekends as "No", it'll help us know whether you're planning to not participate, or did not respond. Thanks! Maciej From rupert at opencsw.org Fri Feb 17 10:39:41 2012 From: rupert at opencsw.org (rupert THURNER) Date: Fri, 17 Feb 2012 10:39:41 +0100 Subject: [csw-maintainers] Dublin release highlights In-Reply-To: References: Message-ID: subversion 1.7x is now linked with gnome keyring mercurial 2.1 now supports large files and phases On Thu, Feb 16, 2012 at 15:40, Maciej (Matchek) Blizinski wrote: > Hello maintainers, > > With the upcoming dublin release, we'd like to write an article > outlining its main highlights. I'd like to ask everyone to send in > their highlights of the dublin release. Something that was significant > and you'd like it mentioned. For example, I'd put forward the GCC > update: version upgrade and moving it from /opt/csw/gcc4 to /opt/csw. > We also had a collective effort to repackage shared libraries. I'm > sure there's more and that every maintainer has something significant > that they did in the last year. Please send them in, either on or off > the list, by the end of the next week, which is the 24th of February. > > Maciej > _______________________________________________ > maintainers mailing list > maintainers at lists.opencsw.org > https://lists.opencsw.org/mailman/listinfo/maintainers > .:: This mailing list's archive is public. ::. From ja at opencsw.org Sat Feb 18 19:14:58 2012 From: ja at opencsw.org (Juergen Arndt) Date: Sat, 18 Feb 2012 19:14:58 +0100 Subject: [csw-maintainers] Nagios 3.3.1 in experimental Message-ID: Hi all, Nagios 3.3.1 is available in experimental. Feedback is welcome. Regards, Juergen -- Juergen Arndt From maciej at opencsw.org Sun Feb 19 10:08:38 2012 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Sun, 19 Feb 2012 09:08:38 +0000 Subject: [csw-maintainers] Packages in /opt/csw/X11 Message-ID: A number of packages are compiled in /opt/csw/X11. For example, libxcb. I know that we were once rebuilding all X Window system related packages into /opt/csw/X11, then Phil discovered that 3D acceleration can't be supported that way, and we stopped. And then any activity in this area died. So what's the current status. libxcb is still installed into /opt/csw/X11. Do we move it back to /opt/csw, or leave it in /opt/csw/X11? Maciej From maciej at opencsw.org Sun Feb 19 13:06:30 2012 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Sun, 19 Feb 2012 12:06:30 +0000 Subject: [csw-maintainers] Website suggestion In-Reply-To: References: Message-ID: I looked at www-mockup today. The /var/www/www-mockup.opencsw.org/htdocs directory is not controlled by subversion or git. How do we make and migrate changes between www-mockup and www? From dam at opencsw.org Sun Feb 19 15:29:15 2012 From: dam at opencsw.org (Dagobert Michelsen) Date: Sun, 19 Feb 2012 15:29:15 +0100 Subject: [csw-maintainers] Packages in /opt/csw/X11 In-Reply-To: References: Message-ID: Hi Maciej, Am 19.02.2012 um 10:08 schrieb Maciej (Matchek) Blizi?ski: > A number of packages are compiled in /opt/csw/X11. For example, libxcb. From http://www.opencsw.org/packages/libxcb/ I see no dependencies actually using it. With the deprecation of Solaris 9 we should be pretty clear in dropping CSW X11. William: You have the most insight in this package conglomerat. Do you concur or do you see a reason to carry on with our custom X11 packages? Best regards -- Dago From dam at opencsw.org Sun Feb 19 15:30:17 2012 From: dam at opencsw.org (Dagobert Michelsen) Date: Sun, 19 Feb 2012 15:30:17 +0100 Subject: [csw-maintainers] Website suggestion In-Reply-To: References: Message-ID: Hi Maciej, Am 19.02.2012 um 13:06 schrieb Maciej (Matchek) Blizi?ski: > I looked at www-mockup today. The > /var/www/www-mockup.opencsw.org/htdocs directory is not controlled by > subversion or git. How do we make and migrate changes between > www-mockup and www? I would say make a copy of it on the machine, check the current version out and copy back the old stuff but keep the repo information. That way we can see the differences and carry on with a clean integration path. Best regards -- Dago From bwalton at opencsw.org Mon Feb 20 15:36:37 2012 From: bwalton at opencsw.org (Ben Walton) Date: Mon, 20 Feb 2012 09:36:37 -0500 Subject: [csw-maintainers] ruby 1.9 madness Message-ID: <1329747521-sup-9372@pinkfloyd.chass.utoronto.ca> Hi All, I'm hoping to get some input on a rather nasty issue with packaging ruby 1.9. The current version of 1.9 is 1.9.3p0. When built, by default, I end up with $libdir/ruby/1.9.1 and sonames like .so.1.9.1. This is because even though the language is at 1.9.3p0, the API is at 1.9.1 still so older libraries are still compatible. This[1] thread outlines how the ruby devs plan for changes in the API version. I _do_ have the option to override the API number with a configure option. Thinking about how to handle library transitions and the installation of gems, things could be messy for users (and us in terms of forced repackaging) if the release 1.9.4 sees a change in the API. So, it seems to me that I can either override the API number at build time and use three digit versioning in the package names to allow graceful transitions through the version/api matrix or use three digit naming but make those digits line up with the API versio giving us ruby191 for a package delivering ruby -v == '1.9.3'. (The latter is what Debian does currently.) A dummy ruby19 package could always point at the latest three digit version. I think I prefer the former and it doesn't sound like Debian is happy with their current solution[2], but neither is really appealing. What do you folks thing about this mess? Thanks -Ben [1] http://www.ruby-forum.com/topic/176847 [2] http://www.lucas-nussbaum.net/blog/?p=681 -- Ben Walton Systems Programmer - CHASS University of Toronto C:416.407.5610 | W:416.978.4302 From maciej at opencsw.org Tue Feb 21 09:41:43 2012 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Tue, 21 Feb 2012 08:41:43 +0000 Subject: [csw-maintainers] ruby 1.9 madness In-Reply-To: <1329747521-sup-9372@pinkfloyd.chass.utoronto.ca> References: <1329747521-sup-9372@pinkfloyd.chass.utoronto.ca> Message-ID: 2012/2/20 Ben Walton : > So, it seems to me that I can either override the API number at build > time and use three digit versioning in the package names to allow > graceful transitions through the version/api matrix or use three digit > naming but make those digits line up with the API versio giving us > ruby191 for a package delivering ruby -v == '1.9.3'. ?(The latter is > what Debian does currently.) ?A dummy ruby19 package could always > point at the latest three digit version. > > I think I prefer the former and it doesn't sound like Debian is happy > with their current solution[2], but neither is really appealing. It looks like the main problem with the ruby191 package containing ruby -v == '1.9.3' is that it confuses users, but it works well on the technical level. I assume we're talking about the main ruby package, the one that contains the interpreter executable. You could have a "ruby19" wrapper package pointing to the newest one, and keep the binaries in a package watermarked with the API version, but done in such a way that it's clear to users that this has to do with the version handling. For example, give it some horrible name like "ruby19_api1", which will obscure the resemblance between strings "1.9.3" and "1.9.1". Maciej From bwalton at opencsw.org Wed Feb 22 01:59:49 2012 From: bwalton at opencsw.org (Ben Walton) Date: Tue, 21 Feb 2012 19:59:49 -0500 Subject: [csw-maintainers] ruby 1.9 madness In-Reply-To: References: <1329747521-sup-9372@pinkfloyd.chass.utoronto.ca> Message-ID: <1329871232-sup-6199@pinkfloyd.chass.utoronto.ca> Excerpts from Maciej (Matchek) Blizi?ski's message of Tue Feb 21 03:41:43 -0500 2012: > It looks like the main problem with the ruby191 package containing > ruby -v == '1.9.3' is that it confuses users, but it works well on > the technical level. I assume we're talking about the main ruby > package, the one that contains the interpreter executable. Yes, things are mostly under the covers so you won't see it unless you look. The things that land in bin/ are where users will spot the oddities...although I do have the option to adjust the suffix, making it match the output of -v even though directory names in lib/ might be using the api version instead. I think using a weird nomenclature on the package name and having it pulled in via metapackage might be the nicest method if I suffix the binaries with the api number instead of the version number. What I find strange is that the intent of this scheme was to allow having multiple versions of ruby, at different api levels installed side by side...It seems to me that it would have been much easier if they just used the package version number to accomplish this. Thanks -Ben -- Ben Walton Systems Programmer - CHASS University of Toronto C:416.407.5610 | W:416.978.4302 From jcraig at opencsw.org Wed Feb 22 16:10:30 2012 From: jcraig at opencsw.org (Jonathan Craig) Date: Wed, 22 Feb 2012 10:10:30 -0500 Subject: [csw-maintainers] ruby 1.9 madness In-Reply-To: <1329871232-sup-6199@pinkfloyd.chass.utoronto.ca> References: <1329747521-sup-9372@pinkfloyd.chass.utoronto.ca> <1329871232-sup-6199@pinkfloyd.chass.utoronto.ca> Message-ID: On Tue, Feb 21, 2012 at 7:59 PM, Ben Walton wrote: > What I find strange is that the intent of this scheme was to allow > having multiple versions of ruby, at different api levels installed > side by side...It seems to me that it would have been much easier if > they just used the package version number to accomplish this. > I'm sure it looked brilliant on paper. I guess this allows multiple versions of ruby to share .so libraries on a host. I wonder how ruby version manager (rvm) factors into this. Does it handle packaging differently, and are developers moving towards it rather than OS specific packaging? Regardless, I agree with you in that we should follow the lead set by Linux and use the API version rather than the Ruby version number. Its awkward to have these decoupled but I suppose ruby enthusiasts should be used to it, or get used to it. From maciej at opencsw.org Wed Feb 22 16:56:02 2012 From: maciej at opencsw.org (Maciej (Matchek) Blizinski) Date: Wed, 22 Feb 2012 15:56:02 +0000 Subject: [csw-maintainers] REST: Getting paths and packages by basename Message-ID: I added a new RESTful URL. It allows you to give it a basename, e.g. "libgcc_s.so.1" and get the list of paths under which this file occurs. For example: $ time curl -s " http://buildfarm.opencsw.org/pkgdb/rest/catalogs/unstable/sparc/SunOS5.10/pkgnames-and-paths-by-basename?basename=bash" | python -c "import json, pprint, sys; pprint.pprint(json.loads(sys.stdin.read()))" {u'/opt/csw/bin': [u'CSWbash'], u'/opt/csw/share/doc': [u'CSWbash'], u'/opt/csw/share/quilt/compat': [u'CSWquilt'], u'/usr/bin': [u'SUNWbash']} real 0m0.558s user 0m0.030s sys 0m0.010s (the python part is only to prettify the results) This call is used by checkpkg to identify which shared libraries could a binary link to. This is one of the more powerful queries that the checkpkg database allows. It can also get quite slow if you're searching for something occurring in many packages. Maciej -------------- next part -------------- An HTML attachment was scrubbed... URL: From trygvis at opencsw.org Thu Feb 23 01:19:01 2012 From: trygvis at opencsw.org (=?ISO-8859-1?Q?Trygve_Laugst=F8l?=) Date: Thu, 23 Feb 2012 01:19:01 +0100 Subject: [csw-maintainers] REST: Getting paths and packages by basename In-Reply-To: References: Message-ID: <4F4585F5.60304@opencsw.org> On 02/22/2012 04:56 PM, Maciej (Matchek) Blizinski wrote: > I added a new RESTful URL. It allows you to give it a basename, e.g. > "libgcc_s.so.1" and get the list of paths under which this file occurs. > For example: > > $ time curl -s > "http://buildfarm.opencsw.org/pkgdb/rest/catalogs/unstable/sparc/SunOS5.10/pkgnames-and-paths-by-basename?basename=bash" > | python -c "import json, pprint, sys; > pprint.pprint(json.loads(sys.stdin.read()))" > {u'/opt/csw/bin': [u'CSWbash'], > u'/opt/csw/share/doc': [u'CSWbash'], > u'/opt/csw/share/quilt/compat': [u'CSWquilt'], > u'/usr/bin': [u'SUNWbash']} I see your Python and raise with: $ time curl -s "http://buildfarm.opencsw.org/pkgdb/rest/catalogs/unstable/sparc/SunOS5.10/pkgnames-and-paths-by-basename?basename=bash" | python -mjson.tool { "/opt/csw/bin": [ "CSWbash" ], "/opt/csw/share/doc": [ "CSWbash" ], "/opt/csw/share/quilt/compat": [ "CSWquilt" ], "/usr/bin": [ "SUNWbash" ] } real 0m0.391s user 0m0.028s sys 0m0.008s > real 0m0.558s > user 0m0.030s > sys 0m0.010s > > (the python part is only to prettify the results) > > This call is used by checkpkg to identify which shared libraries could a > binary link to. This is one of the more powerful queries that the > checkpkg database allows. It can also get quite slow if you're searching > for something occurring in many packages. > > Maciej > > > _______________________________________________ > maintainers mailing list > maintainers at lists.opencsw.org > https://lists.opencsw.org/mailman/listinfo/maintainers > .:: This mailing list's archive is public. ::. -- Trygve From maciej at opencsw.org Thu Feb 23 09:23:33 2012 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Thu, 23 Feb 2012 08:23:33 +0000 Subject: [csw-maintainers] REST: Getting paths and packages by basename In-Reply-To: <4F4585F5.60304@opencsw.org> References: <4F4585F5.60304@opencsw.org> Message-ID: 2012/2/23 Trygve Laugst?l : > (...) | python -mjson.tool That's brilliant, I didn't know this trick! From ihsan at opencsw.org Thu Feb 23 13:40:58 2012 From: ihsan at opencsw.org (=?UTF-8?B?xLBoc2FuwqBEb8SfYW4=?=) Date: Thu, 23 Feb 2012 13:40:58 +0100 Subject: [csw-maintainers] checkpkg errors Message-ID: <4F4633DA.6070506@opencsw.org> Hi, I was packaging a few new Perl modules and run into this issue with checkpkg: CHECKPKG_OVERRIDES_CSWpm-email-messageid += pkginfo-bad-catalogname|pm_email-messageid CHECKPKG_OVERRIDES_CSWpm-email-messageid += catalogname-does-not-match-pkgname|pkgname=CSWpm-email-messageid|catalogname=pm_email-messageid|expected-catalogname=pm_email_messageid I missed probably a policy change. What would be the correct catalog name? Ihsan -- ihsan at dogan.ch http://blog.dogan.ch/ From dam at opencsw.org Thu Feb 23 13:43:24 2012 From: dam at opencsw.org (Dagobert Michelsen) Date: Thu, 23 Feb 2012 13:43:24 +0100 Subject: [csw-maintainers] checkpkg errors In-Reply-To: <4F4633DA.6070506@opencsw.org> References: <4F4633DA.6070506@opencsw.org> Message-ID: <6548EAB5-209A-4EAB-8A8F-D421C51626BB@opencsw.org> Hi Ihsan, Am 23.02.2012 um 13:40 schrieb ?hsan Do?an: > I was packaging a few new Perl modules and run into this issue with > checkpkg: > > CHECKPKG_OVERRIDES_CSWpm-email-messageid += > pkginfo-bad-catalogname|pm_email-messageid catalog names not contain '-', but '_'. It is best to leave CATALOGNAME empty if it is the default. > CHECKPKG_OVERRIDES_CSWpm-email-messageid += > catalogname-does-not-match-pkgname|pkgname=CSWpm-email-messageid|catalogname=pm_email-messageid|expected-catalogname=pm_email_messageid > > I missed probably a policy change. What would be the correct catalog name? You may want to test in cpan/ ./makemake Email::MessageID which generates the Makefile automatically among other things. Best regards -- Dago -- "You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process." - xkcd #896 From ihsan at opencsw.org Thu Feb 23 14:03:27 2012 From: ihsan at opencsw.org (=?UTF-8?B?xLBoc2FuwqBEb8SfYW4=?=) Date: Thu, 23 Feb 2012 14:03:27 +0100 Subject: [csw-maintainers] checkpkg errors In-Reply-To: <6548EAB5-209A-4EAB-8A8F-D421C51626BB@opencsw.org> References: <4F4633DA.6070506@opencsw.org> <6548EAB5-209A-4EAB-8A8F-D421C51626BB@opencsw.org> Message-ID: <4F46391F.5040408@opencsw.org> Hi Dago, Am 23.02.2012 13:43, schrieb Dagobert Michelsen: >> I was packaging a few new Perl modules and run into this issue with >> checkpkg: >> >> CHECKPKG_OVERRIDES_CSWpm-email-messageid += >> pkginfo-bad-catalogname|pm_email-messageid > > catalog names not contain '-', but '_'. It is best to leave CATALOGNAME empty if > it is the default. Thanks. I've used the defaults provided by Gar. Ihsan -- ihsan at dogan.ch http://blog.dogan.ch/ From dam at opencsw.org Thu Feb 23 14:32:08 2012 From: dam at opencsw.org (Dagobert Michelsen) Date: Thu, 23 Feb 2012 14:32:08 +0100 Subject: [csw-maintainers] Inconsistency on package database Message-ID: Hi, I just found an ugly inconsistency in the web database: If you look at the dependencies of the Hobbit package at http://www.opencsw.org/packages/hobbit/ you see four dependencies: chkconfig A system tool for maintaining the /etc/rcX.d hierarchy. common Installs some useful standard CSW filesystem symlinks and files fping fping is a ping(1) like program accepting any number of hosts on the command line and producing outp pcre Perl-compatible regular expression tools When you disassemble the package you see some more: dam at login [login]:/home/dam/tmp/CSWhobbit/install > more depend P CSWcommon common - common files and dirs for CSW packages P CSWrrd rrdtool - Round Robin Database graphing tool P CSWfping fping - ping(1) replacement, optimised for scripting. P CSWoldaprt openldap_rt - OpenLDAP runtime libraries P CSWpng png - library for Portable Network Graphics format (PNG) P CSWpcre pcre - Perl-compatible regular expression library P CSWossl openssl - The Open Source toolkit for SSL and TLS. P CSWchkconfig chkconfig - Tool for maintaining the /etc/rcX.d hierarchy. P SUNWcsu Core Solaris, (Usr) P SUNWcsl Core Solaris, (Shared Libs) While the SUNW deps can be safely ignored by the package database the missing deps to the other packages is really bad: these are obsoleted packages transformed to _stub which are no longer in the package database. I think either the display needs to be adjusted to view by what it has been obsoleted or the stub packages need to be re-added to the database. Thoughts? Best regards -- Dago -- "You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process." - xkcd #896 From dam at opencsw.org Thu Feb 23 14:41:11 2012 From: dam at opencsw.org (Dagobert Michelsen) Date: Thu, 23 Feb 2012 14:41:11 +0100 Subject: [csw-maintainers] Inconsistency on package database In-Reply-To: References: Message-ID: <7C3BAD6A-C8EF-4CC3-9B99-0BAFE73CE6B8@opencsw.org> Hi, Am 23.02.2012 um 14:32 schrieb Dagobert Michelsen: > I just found an ugly inconsistency in the web database: If you look at the > dependencies of the Hobbit package at > http://www.opencsw.org/packages/hobbit/ > you see four dependencies: > chkconfig A system tool for maintaining the /etc/rcX.d hierarchy. > common Installs some useful standard CSW filesystem symlinks and files > fping fping is a ping(1) like program accepting any number of hosts on the command line and producing outp > pcre Perl-compatible regular expression tools > > When you disassemble the package you see some more: > > dam at login [login]:/home/dam/tmp/CSWhobbit/install > more depend > P CSWcommon common - common files and dirs for CSW packages > P CSWrrd rrdtool - Round Robin Database graphing tool > P CSWfping fping - ping(1) replacement, optimised for scripting. > P CSWoldaprt openldap_rt - OpenLDAP runtime libraries > P CSWpng png - library for Portable Network Graphics format (PNG) > P CSWpcre pcre - Perl-compatible regular expression library > P CSWossl openssl - The Open Source toolkit for SSL and TLS. > P CSWchkconfig chkconfig - Tool for maintaining the /etc/rcX.d hierarchy. > P SUNWcsu Core Solaris, (Usr) > P SUNWcsl Core Solaris, (Shared Libs) > > While the SUNW deps can be safely ignored by the package database the missing > deps to the other packages is really bad: these are obsoleted packages > transformed to _stub which are no longer in the package database. > I think either the display needs to be adjusted to view by what it has been obsoleted > or the stub packages need to be re-added to the database. One more thing: the bugs associated with an obsoleted package like samba_common are still in the bugtracker, but are no longer visible for the maintainer by showing the maintainer bug status as the obsoleted package is not associated with the maintainer although it was built in the batch. See again samba_common for my bug report page: http://www.opencsw.org/buglist/maintainer.cgi?maintainer='dam' and in the toplist of all bugs: http://www.opencsw.org/buglist/buglist.cgi Best regards -- Dago -- "You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process." - xkcd #896 From ihsan at opencsw.org Thu Feb 23 15:16:03 2012 From: ihsan at opencsw.org (=?UTF-8?B?xLBoc2FuwqBEb8SfYW4=?=) Date: Thu, 23 Feb 2012 15:16:03 +0100 Subject: [csw-maintainers] submitting packages Message-ID: <4F464A23.1000209@opencsw.org> Hi, Looks like I've missed something else: As far submitpkg is gone, how do I submit new packages? :-) Ihsan -- ihsan at dogan.ch http://blog.dogan.ch/ From dam at opencsw.org Thu Feb 23 15:18:26 2012 From: dam at opencsw.org (Dagobert Michelsen) Date: Thu, 23 Feb 2012 15:18:26 +0100 Subject: [csw-maintainers] submitting packages In-Reply-To: <4F464A23.1000209@opencsw.org> References: <4F464A23.1000209@opencsw.org> Message-ID: Hi Ihsan, Am 23.02.2012 um 15:16 schrieb ?hsan Do?an: > Looks like I've missed something else: As far submitpkg is gone, how do > I submit new packages? :-) csw-upload-pkg ... Best regards -- Dago -- "You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process." - xkcd #896 From ihsan at opencsw.org Thu Feb 23 15:36:00 2012 From: ihsan at opencsw.org (=?UTF-8?B?xLBoc2FuwqBEb8SfYW4=?=) Date: Thu, 23 Feb 2012 15:36:00 +0100 Subject: [csw-maintainers] submitting packages In-Reply-To: References: <4F464A23.1000209@opencsw.org> Message-ID: <4F464ED0.7020305@opencsw.org> Hi Dago, Am 23.02.2012 15:18, schrieb Dagobert Michelsen: >> Looks like I've missed something else: As far submitpkg is gone, how do >> I submit new packages? :-) > > csw-upload-pkg ... Thanks. Ihsan -- ihsan at dogan.ch http://blog.dogan.ch/ From bwalton at opencsw.org Fri Feb 24 14:46:25 2012 From: bwalton at opencsw.org (Ben Walton) Date: Fri, 24 Feb 2012 08:46:25 -0500 Subject: [csw-maintainers] Inconsistency on package database In-Reply-To: References: Message-ID: <1330091058-sup-811@pinkfloyd.chass.utoronto.ca> Excerpts from Dagobert Michelsen's message of Thu Feb 23 08:32:08 -0500 2012: Hi Dago, > While the SUNW deps can be safely ignored by the package database > the missing deps to the other packages is really bad: these are > obsoleted packages transformed to _stub which are no longer in the > package database. I think either the display needs to be adjusted > to view by what it has been obsoleted or the stub packages need to > be re-added to the database. Yes, this is a problem. I'll look at it over the weekend. It should be fairly easy to either add all _stubs and setup the deps again or to use the stub info to register the transitive dependency instead. Is there a preference for this? Thanks -Ben -- Ben Walton Systems Programmer - CHASS University of Toronto C:416.407.5610 | W:416.978.4302 From bwalton at opencsw.org Fri Feb 24 14:47:49 2012 From: bwalton at opencsw.org (Ben Walton) Date: Fri, 24 Feb 2012 08:47:49 -0500 Subject: [csw-maintainers] Inconsistency on package database In-Reply-To: <7C3BAD6A-C8EF-4CC3-9B99-0BAFE73CE6B8@opencsw.org> References: <7C3BAD6A-C8EF-4CC3-9B99-0BAFE73CE6B8@opencsw.org> Message-ID: <1330091219-sup-8511@pinkfloyd.chass.utoronto.ca> Excerpts from Dagobert Michelsen's message of Thu Feb 23 08:41:11 -0500 2012: Hi Dago, > One more thing: the bugs associated with an obsoleted package like > samba_common are still in the bugtracker, but are no longer visible > for the maintainer by showing the maintainer bug status as the > obsoleted package is not associated with the maintainer although it > was built in the batch. See again samba_common for my bug report > page: http://www.opencsw.org/buglist/maintainer.cgi?maintainer='dam' > and in the toplist of all bugs: > http://www.opencsw.org/buglist/buglist.cgi When packages are stubbed out, the bugs are supposed to be moved to one of the other projects that obsoletes them. Is this not what's happening or am I misunderstanding the problem you've noticed? Thanks -Ben -- Ben Walton Systems Programmer - CHASS University of Toronto C:416.407.5610 | W:416.978.4302 From dam at opencsw.org Fri Feb 24 14:52:59 2012 From: dam at opencsw.org (Dagobert Michelsen) Date: Fri, 24 Feb 2012 14:52:59 +0100 Subject: [csw-maintainers] Inconsistency on package database In-Reply-To: <1330091219-sup-8511@pinkfloyd.chass.utoronto.ca> References: <7C3BAD6A-C8EF-4CC3-9B99-0BAFE73CE6B8@opencsw.org> <1330091219-sup-8511@pinkfloyd.chass.utoronto.ca> Message-ID: <996596CC-33B7-4705-832A-92000B9DEFF0@opencsw.org> Hi Ben, Am 24.02.2012 um 14:47 schrieb Ben Walton: > Excerpts from Dagobert Michelsen's message of Thu Feb 23 08:41:11 -0500 2012: >> One more thing: the bugs associated with an obsoleted package like >> samba_common are still in the bugtracker, but are no longer visible >> for the maintainer by showing the maintainer bug status as the >> obsoleted package is not associated with the maintainer although it >> was built in the batch. See again samba_common for my bug report >> page: http://www.opencsw.org/buglist/maintainer.cgi?maintainer='dam' >> and in the toplist of all bugs: >> http://www.opencsw.org/buglist/buglist.cgi > > When packages are stubbed out, the bugs are supposed to be moved to > one of the other projects that obsoletes them. Is this not what's > happening or am I misunderstanding the problem you've noticed? Correct, this is not happening. Please browse on the bugtracker directly to samba by visiting https://www.opencsw.org/mantis/ and then use the popup to select "samba_common". Then open the search filter and set "Hide Status" to "none", then lots of bugs appear which I closed yesterday. From your description I would have expected that they would have been moved to "samba", as samba_common was obsoleted by this one. However, please note that a package may be obsoleted by more than one package, so it may be best to just leave the bugs and re-register the _stub packages. I tend to leave artifacts in the view so they are clearly visible instead of applying different levels of magic. Best regards -- Dago -- "You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process." - xkcd #896 From maciej at opencsw.org Sat Feb 25 14:18:03 2012 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Sat, 25 Feb 2012 13:18:03 +0000 Subject: [csw-maintainers] Catalog checks via REST In-Reply-To: References: <033689A1-FD32-4DCB-9CE3-6C68D3AF6B4B@opencsw.org> <65B032CA-3C2D-4030-B1AA-5CF7C1086E42@opencsw.org> Message-ID: 2012/1/15 Peter Bonivart : > I'm not talking about pushing the catalogs, I'm talking about getting > info about a bad upload during the upload, the same thing you have > brought up earlier. So you're talking about synchronous checks. > If the database contained everything needed for a catalog to be > generated directly from it one could be created temporarily, checked > by chkcat and the result could be returned to csw-upload-pkg. With the > new chkcat I think this could be done in 1-2 seconds per catalog max. It depends on the database performance, it could be more than only 1-2 seconds. The performance of a normalized database is usually worse than performance of cached data or denormalized database. I wouldn't like to denormalize ours - denormalizing adds complexity and makes code maintenance harder. The other option is caching. > This has nothing to do with the actual creation of official catalogs, > they could still be generated from a cron job. I plan to update the database schema and add more fields to the tables. I try to keep the number of tables and number of fields to the minimum (less code to maintain!), but I'll make sure that we in the next revision the database has enough information directly in the tables (as opposed to pickled objects). But I've been planning to do that for maybe a year now and there was always something for me to do which was more urgent. If you feel strongly about this and really want to see this happen soon, you need to take it on yourself. Maciej From maciej at opencsw.org Sun Feb 26 15:18:12 2012 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Sun, 26 Feb 2012 14:18:12 +0000 Subject: [csw-maintainers] C++ libraries built with GCC Message-ID: If we build any C++ libraries, we need to consider which compiler we build them with. For example, if we want to build something that depends on Boost, and we want (or need) to build it with GCC, we have to build Boost with GCC as well. But Studio and GCC produce incompatible libraries because of different name mangling schemes. Therefore, I suggest that we designate a separate directory for GCC C++ libraries. To keep things simple, --prefix=/opt/csw/gcc, and --libdir=/opt/csw/gcc/lib. Package names would be CSWlibfoo0-gcc and libfoo0_gcc. An alternative would be to patch sonames. Thoughts? Maciej From bwalton at opencsw.org Sun Feb 26 18:32:13 2012 From: bwalton at opencsw.org (Ben Walton) Date: Sun, 26 Feb 2012 12:32:13 -0500 Subject: [csw-maintainers] Inconsistency on package database In-Reply-To: <996596CC-33B7-4705-832A-92000B9DEFF0@opencsw.org> References: <7C3BAD6A-C8EF-4CC3-9B99-0BAFE73CE6B8@opencsw.org> <1330091219-sup-8511@pinkfloyd.chass.utoronto.ca> <996596CC-33B7-4705-832A-92000B9DEFF0@opencsw.org> Message-ID: <1330274181-sup-310@pinkfloyd.chass.utoronto.ca> Excerpts from Dagobert Michelsen's message of Fri Feb 24 08:52:59 -0500 2012: Hi Dago, > > When packages are stubbed out, the bugs are supposed to be moved > > to one of the other projects that obsoletes them. Is this not > > what's happening or am I misunderstanding the problem you've > > noticed? > > Correct, this is not happening. Please browse on the bugtracker > directly to samba by visiting https://www.opencsw.org/mantis/ and > then use the popup to select "samba_common". Then open the search > filter and set "Hide Status" to "none", then lots of bugs appear > which I closed yesterday. From your description I would have > expected that they would have been moved to "samba", as samba_common > was obsoleted by this one. Ok, I'm looking into this now. There are some heuristics applied here to determine where the bugs are moved. I do take into account the fact that once stub can pull in multiple other new packages. I don't actually see samba_common* in the catalog currently. Might it have been removed instead of stubbed out? That would line up with what I see in the db. > However, please note that a package may be obsoleted by more than > one package, so it may be best to just leave the bugs and > re-register the _stub packages. I tend to leave artifacts in the > view so they are clearly visible instead of applying different > levels of magic. The idea behind the current function is that dropped packages are hidden from view, leaving the bugs as is. It's a very simple change if people would like the removal of a package from the catalog to result in a no-op as far as mantis is concerned though. If we want to keep stubs in the db instead of moving the bugs to the new packages, that's a larger change. I don't mind doing this if it's the preferred behaviour though. In the case of stub handling, I attempt to move bugs to one of the packages in the dependency list of the stub using the following mechanism: 1. Remove CSWcommon from the list. 2. Reverse the dependency list to hopefully skim out automatically added things (CSWcas-*, etc). Some of these auto things could be moved to step 1. 3. Pick the first package. I see that there is some potential for more corner case bugs in this area though, so I'm looking at how to improve it... If we want to drop some of the rename detection for stub handling, many things can become simpler. In hindsight, this might be a better option for the web db...We could note the catalog name change in the database and preserve the dependencies. The _stub packages could then be filtered from the 'whole list' view but made available when clicking through from an individual package. If that sounds ok, I'll make the required changes. I think it might also be a good exercise to empty all of the package tables (in my test db) and start from a clean slate as if the original catalog was empty. The tables could then be imported to the webdb on www.opencsw.org. If anyone is putting real effort into making a nice package browser based on the buildfarm db though, let me know as I'd rather spend effort on that than polishing this turd if it's going to be replaced. :) Thanks -Ben -- Ben Walton Systems Programmer - CHASS University of Toronto C:416.407.5610 | W:416.978.4302 From bwalton at opencsw.org Sun Feb 26 18:52:10 2012 From: bwalton at opencsw.org (Ben Walton) Date: Sun, 26 Feb 2012 12:52:10 -0500 Subject: [csw-maintainers] C++ libraries built with GCC In-Reply-To: References: Message-ID: <1330278296-sup-7303@pinkfloyd.chass.utoronto.ca> Excerpts from Maciej (Matchek) Blizi?ski's message of Sun Feb 26 09:18:12 -0500 2012: Hi Maciej, > If we build any C++ libraries, we need to consider which compiler we > build them with. For example, if we want to build something that > depends on Boost, and we want (or need) to build it with GCC, we > have to build Boost with GCC as well. But Studio and GCC produce > incompatible libraries because of different name mangling schemes. > > Therefore, I suggest that we designate a separate directory for GCC > C++ libraries. To keep things simple, --prefix=/opt/csw/gcc, and > --libdir=/opt/csw/gcc/lib. Package names would be CSWlibfoo0-gcc > and libfoo0_gcc. The idea is good and we should definitely accommodate this nicely. I don't like /opt/csw/gcc, but I haven't got a better name that doesn't also suffer from: 1. Top level directory proliferation. 2. Confusion with /opt/csw/gcc[34]. If nobody has a better path name, I'd be inclined to go with /opt/csw/gcc as suggested. > An alternative would be to patch sonames. Patching the sonames feels much more invasive and would be harder (more work per package) to accommodate. Thanks -Ben -- Ben Walton Systems Programmer - CHASS University of Toronto C:416.407.5610 | W:416.978.4302 From bwalton at opencsw.org Sun Feb 26 19:15:37 2012 From: bwalton at opencsw.org (Ben Walton) Date: Sun, 26 Feb 2012 13:15:37 -0500 Subject: [csw-maintainers] Inconsistency on package database In-Reply-To: <1330274181-sup-310@pinkfloyd.chass.utoronto.ca> References: <7C3BAD6A-C8EF-4CC3-9B99-0BAFE73CE6B8@opencsw.org> <1330091219-sup-8511@pinkfloyd.chass.utoronto.ca> <996596CC-33B7-4705-832A-92000B9DEFF0@opencsw.org> <1330274181-sup-310@pinkfloyd.chass.utoronto.ca> Message-ID: <1330280059-sup-7355@pinkfloyd.chass.utoronto.ca> Excerpts from Ben Walton's message of Sun Feb 26 12:32:13 -0500 2012: > If anyone is putting real effort into making a nice package browser > based on the buildfarm db though, let me know as I'd rather spend > effort on that than polishing this turd if it's going to be > replaced. :) Just to clarify, I'm referring to our limited legacy db schema and single arch/rel view of the package db here. My code isn't perfect by any stretch but I don't think it's a turd... *g* Thanks -Ben -- Ben Walton Systems Programmer - CHASS University of Toronto C:416.407.5610 | W:416.978.4302 From maciej at opencsw.org Sun Feb 26 20:19:29 2012 From: maciej at opencsw.org (Maciej =?utf-8?Q?Blizi=C5=84ski?=) Date: Sun, 26 Feb 2012 19:19:29 +0000 Subject: [csw-maintainers] C++ libraries built with GCC In-Reply-To: <1330278296-sup-7303@pinkfloyd.chass.utoronto.ca> References: <1330278296-sup-7303@pinkfloyd.chass.utoronto.ca> Message-ID: <20120226191929.GA8597@quince.home.blizinski.pl> Ben Walton wrote: > Excerpts from Maciej (Matchek) Blizi?ski's message of Sun Feb 26 09:18:12 -0500 2012: > > Hi Maciej, > > The idea is good and we should definitely accommodate this nicely. I > don't like /opt/csw/gcc, but I haven't got a better name that doesn't > also suffer from: > > 1. Top level directory proliferation. > 2. Confusion with /opt/csw/gcc[34]. Maybe /opt/csw/lib/gcc would be a better location? I've experimented with this in the past, with mixed results. Software packages generally like their layout being modified by --prefix, and not by --libdir. > If nobody has a better path name, I'd be inclined to go with > /opt/csw/gcc as suggested. > > > An alternative would be to patch sonames. > > Patching the sonames feels much more invasive and would be harder > (more work per package) to accommodate. True. On the other hand, the results would be better accommodated to our needs. We could keep both GCC and Studio libraries under /opt/csw/lib and package name checking would not need to handle GCC libs as a special case. Maciej -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 198 bytes Desc: Digital signature URL: From maciej at opencsw.org Sun Feb 26 20:23:30 2012 From: maciej at opencsw.org (Maciej =?utf-8?Q?Blizi=C5=84ski?=) Date: Sun, 26 Feb 2012 19:23:30 +0000 Subject: [csw-maintainers] Inconsistency on package database In-Reply-To: <1330280059-sup-7355@pinkfloyd.chass.utoronto.ca> References: <7C3BAD6A-C8EF-4CC3-9B99-0BAFE73CE6B8@opencsw.org> <1330091219-sup-8511@pinkfloyd.chass.utoronto.ca> <996596CC-33B7-4705-832A-92000B9DEFF0@opencsw.org> <1330274181-sup-310@pinkfloyd.chass.utoronto.ca> <1330280059-sup-7355@pinkfloyd.chass.utoronto.ca> Message-ID: <20120226192330.GB8597@quince.home.blizinski.pl> Ben Walton wrote: > Excerpts from Ben Walton's message of Sun Feb 26 12:32:13 -0500 2012: > > > If anyone is putting real effort into making a nice package browser > > based on the buildfarm db though, let me know as I'd rather spend > > effort on that than polishing this turd if it's going to be > > replaced. :) > > Just to clarify, I'm referring to our limited legacy db schema and > single arch/rel view of the package db here. My code isn't perfect by > any stretch but I don't think it's a turd... *g* Sure. As far as the package browser is concerned, I wrote this AJAX proof of concept: http://quinoa.blizinski.pl/~maciej/opencsw/test.html It is a static HTML file, and yet it displays current information from the buildfarm database. It dynamically fetches data from the buildfarm and fills the fields on the page. Arguably it adds about 1s to the page latency, but frees the package browser from the problems with local database schema. Maciej -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 198 bytes Desc: Digital signature URL: From bwalton at opencsw.org Sun Feb 26 20:31:18 2012 From: bwalton at opencsw.org (Ben Walton) Date: Sun, 26 Feb 2012 14:31:18 -0500 Subject: [csw-maintainers] Inconsistency on package database In-Reply-To: <20120226192330.GB8597@quince.home.blizinski.pl> References: <7C3BAD6A-C8EF-4CC3-9B99-0BAFE73CE6B8@opencsw.org> <1330091219-sup-8511@pinkfloyd.chass.utoronto.ca> <996596CC-33B7-4705-832A-92000B9DEFF0@opencsw.org> <1330274181-sup-310@pinkfloyd.chass.utoronto.ca> <1330280059-sup-7355@pinkfloyd.chass.utoronto.ca> <20120226192330.GB8597@quince.home.blizinski.pl> Message-ID: <1330284511-sup-3223@pinkfloyd.chass.utoronto.ca> Excerpts from Maciej Blizi?ski's message of Sun Feb 26 14:23:30 -0500 2012: > http://quinoa.blizinski.pl/~maciej/opencsw/test.html > > It is a static HTML file, and yet it displays current information > from the buildfarm database. It dynamically fetches data from the > buildfarm and fills the fields on the page. > > Arguably it adds about 1s to the page latency, but frees the package > browser from the problems with local database schema. I'd think that replicating the db directly to the webserver and updating the package browser accordingly would be the best of both worlds though. It would skip the latency but get the same always accurate, full featured information. It could be ajax still, but wouldn't really need to be... Thoughts? Thanks -Ben -- Ben Walton Systems Programmer - CHASS University of Toronto C:416.407.5610 | W:416.978.4302 From maciej at opencsw.org Sun Feb 26 21:45:55 2012 From: maciej at opencsw.org (Maciej =?utf-8?Q?Blizi=C5=84ski?=) Date: Sun, 26 Feb 2012 20:45:55 +0000 Subject: [csw-maintainers] Inconsistency on package database In-Reply-To: <1330284511-sup-3223@pinkfloyd.chass.utoronto.ca> References: <7C3BAD6A-C8EF-4CC3-9B99-0BAFE73CE6B8@opencsw.org> <1330091219-sup-8511@pinkfloyd.chass.utoronto.ca> <996596CC-33B7-4705-832A-92000B9DEFF0@opencsw.org> <1330274181-sup-310@pinkfloyd.chass.utoronto.ca> <1330280059-sup-7355@pinkfloyd.chass.utoronto.ca> <20120226192330.GB8597@quince.home.blizinski.pl> <1330284511-sup-3223@pinkfloyd.chass.utoronto.ca> Message-ID: <20120226204555.GA10341@quince.home.blizinski.pl> Ben Walton wrote: > Excerpts from Maciej Blizi?ski's message of Sun Feb 26 14:23:30 -0500 2012: > > > http://quinoa.blizinski.pl/~maciej/opencsw/test.html > > > > It is a static HTML file, and yet it displays current information > > from the buildfarm database. It dynamically fetches data from the > > buildfarm and fills the fields on the page. > > > > Arguably it adds about 1s to the page latency, but frees the package > > browser from the problems with local database schema. > > I'd think that replicating the db directly to the webserver and > updating the package browser accordingly would be the best of both > worlds though. It would skip the latency but get the same always > accurate, full featured information. > > It could be ajax still, but wouldn't really need to be... > > Thoughts? We talked a little bit about it on IRC, and the rough conclusion was that the web host could have an own db with a minimal schema. This schema would be optimized for the web page serving, with potential denormalizations allowing to achieve better website performance. The integration from the buildfarm db to the web db could be done over REST, allowing easy schema changes to the buildfarm db, or even a change to a different database engine in the future. Maciej -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 198 bytes Desc: Digital signature URL: From bwalton at opencsw.org Sun Feb 26 21:56:37 2012 From: bwalton at opencsw.org (Ben Walton) Date: Sun, 26 Feb 2012 15:56:37 -0500 Subject: [csw-maintainers] Inconsistency on package database In-Reply-To: <20120226204555.GA10341@quince.home.blizinski.pl> References: <7C3BAD6A-C8EF-4CC3-9B99-0BAFE73CE6B8@opencsw.org> <1330091219-sup-8511@pinkfloyd.chass.utoronto.ca> <996596CC-33B7-4705-832A-92000B9DEFF0@opencsw.org> <1330274181-sup-310@pinkfloyd.chass.utoronto.ca> <1330280059-sup-7355@pinkfloyd.chass.utoronto.ca> <20120226192330.GB8597@quince.home.blizinski.pl> <1330284511-sup-3223@pinkfloyd.chass.utoronto.ca> <20120226204555.GA10341@quince.home.blizinski.pl> Message-ID: <1330289684-sup-3878@pinkfloyd.chass.utoronto.ca> Excerpts from Maciej Blizi?ski's message of Sun Feb 26 15:45:55 -0500 2012: > We talked a little bit about it on IRC, and the rough conclusion was > that the web host could have an own db with a minimal schema. This > schema would be optimized for the web page serving, with potential > denormalizations allowing to achieve better website performance. > The integration from the buildfarm db to the web db could be done > over REST, allowing easy schema changes to the buildfarm db, or even > a change to a different database engine in the future. I like this very much. We can define a new schema to meet our modern needs. It needs to allow for the ideas of catalogs and bundles. Keeping it only loosely coupled to the buildfarm is a net win, I think. Thanks -Ben -- Ben Walton Systems Programmer - CHASS University of Toronto C:416.407.5610 | W:416.978.4302 From bwalton at opencsw.org Sun Feb 26 22:20:04 2012 From: bwalton at opencsw.org (Ben Walton) Date: Sun, 26 Feb 2012 16:20:04 -0500 Subject: [csw-maintainers] ruby 1.9 madness In-Reply-To: References: <1329747521-sup-9372@pinkfloyd.chass.utoronto.ca> <1329871232-sup-6199@pinkfloyd.chass.utoronto.ca> Message-ID: <1330291165-sup-3947@pinkfloyd.chass.utoronto.ca> Excerpts from Jonathan Craig's message of Wed Feb 22 10:10:30 -0500 2012: > than OS specific packaging? Regardless, I agree with you in that we > should follow the lead set by Linux and use the API version rather > than the Ruby version number. Its awkward to have these decoupled > but I suppose ruby enthusiasts should be used to it, or get used to > it. Done. Test packages available in unstable. I'm just about done with the meta package now too. Thanks -Ben -- Ben Walton Systems Programmer - CHASS University of Toronto C:416.407.5610 | W:416.978.4302 From maciej at opencsw.org Sun Feb 26 22:29:07 2012 From: maciej at opencsw.org (Maciej =?utf-8?Q?Blizi=C5=84ski?=) Date: Sun, 26 Feb 2012 21:29:07 +0000 Subject: [csw-maintainers] Inconsistency on package database In-Reply-To: <1330289684-sup-3878@pinkfloyd.chass.utoronto.ca> References: <7C3BAD6A-C8EF-4CC3-9B99-0BAFE73CE6B8@opencsw.org> <1330091219-sup-8511@pinkfloyd.chass.utoronto.ca> <996596CC-33B7-4705-832A-92000B9DEFF0@opencsw.org> <1330274181-sup-310@pinkfloyd.chass.utoronto.ca> <1330280059-sup-7355@pinkfloyd.chass.utoronto.ca> <20120226192330.GB8597@quince.home.blizinski.pl> <1330284511-sup-3223@pinkfloyd.chass.utoronto.ca> <20120226204555.GA10341@quince.home.blizinski.pl> <1330289684-sup-3878@pinkfloyd.chass.utoronto.ca> Message-ID: <20120226212907.GB10341@quince.home.blizinski.pl> With relation to our discussion about how to integrate the web database with with buildfarm database, I'd like to raise the question of a REST interface. Trygve's ideas were ahead of his times. We're slowly catching up with the suggestions he presented in the summercamp in Oslo. My ad-hoc REST endpoints seem to serve us well. Current endpoints are linked from the buildfarm database page[2]. They can be summarized as follows: - all svr4 files in a catalog (catalogname, arch, os-release) has a quick (30s) version and a slow (180s) version - filename to pkgname in a catalog - catalogname to svr4 file in a catalog - pkgname (pkginst) to svr4 file in a catalog - maintainer details by id - basic svr4 file detail (fast) - fill svr4 file detail (slower) - list of files in a svr4 file What do people think about these endpoints? Can we keep on using them or do we need to redesign them before we build too much dependencies around them? Maciej [1] http://lists.opencsw.org/pipermail/maintainers/2012-February/016113.html [2] http://buildfarm.opencsw.org/pkgdb/ From maciej at opencsw.org Sun Feb 26 22:40:45 2012 From: maciej at opencsw.org (Maciej =?utf-8?Q?Blizi=C5=84ski?=) Date: Sun, 26 Feb 2012 21:40:45 +0000 Subject: [csw-maintainers] Backing up the sudo binary Message-ID: <20120226214045.GA11411@quince.home.blizinski.pl> The current sudo package allows to deinstall it and lose the binary. If a pkgutil upgrade fails after deinstallation, and before the installation of sudo, you end up with a system with no sude. Here's an idea: - sudo preremove backs up the sudo binary somewhere on the filesystem - sudo postinstall checks if the new binary is actually installed, and if so, removes the backup (the old binary might have a security risk) This idea would help in my scenario, but has a drawback: when the administrator wants to get rid of the sudo binary from the filesystem. You run pkgrm, and the sudo binary remains somewhere in the system. Is it enough to display a postrm message: "The sudo binary has been backed up at $location, remove it manually if you really want to get rid of it." Thoughts? Maciej From bonivart at opencsw.org Mon Feb 27 00:04:47 2012 From: bonivart at opencsw.org (Peter Bonivart) Date: Mon, 27 Feb 2012 00:04:47 +0100 Subject: [csw-maintainers] Backing up the sudo binary In-Reply-To: <20120226214045.GA11411@quince.home.blizinski.pl> References: <20120226214045.GA11411@quince.home.blizinski.pl> Message-ID: 2012/2/26 Maciej Blizi?ski : > The current sudo package allows to deinstall it and lose the binary. If > a pkgutil upgrade fails after deinstallation, and before the > installation of sudo, you end up with a system with no sude. Here's an > idea: > > - sudo preremove backs up the sudo binary somewhere on the filesystem > - sudo postinstall checks if the new binary is actually installed, and > ?if so, removes the backup (the old binary might have a security risk) > > This idea would help in my scenario, but has a drawback: when the > administrator wants to get rid of the sudo binary from the filesystem. > You run pkgrm, and the sudo binary remains somewhere in the system. > > Is it enough to display a postrm message: "The sudo binary has been > backed up at $location, remove it manually if you really want to get rid > of it." > > Thoughts? I like the idea and you need a message anyway to make others aware of that safety net being available. /peter From jcraig at opencsw.org Mon Feb 27 14:09:41 2012 From: jcraig at opencsw.org (Jonathan Craig) Date: Mon, 27 Feb 2012 08:09:41 -0500 Subject: [csw-maintainers] C++ libraries built with GCC In-Reply-To: <20120226191929.GA8597@quince.home.blizinski.pl> References: <1330278296-sup-7303@pinkfloyd.chass.utoronto.ca> <20120226191929.GA8597@quince.home.blizinski.pl> Message-ID: On Sun, Feb 26, 2012 at 2:19 PM, Maciej Blizi?ski wrote: > Ben Walton wrote: >> Excerpts from Maciej (Matchek) Blizi?ski's message of Sun Feb 26 09:18:12 -0500 2012: >> >> Hi Maciej, >> >> The idea is good and we should definitely accommodate this nicely. ?I >> don't like /opt/csw/gcc, but I haven't got a better name that doesn't >> also suffer from: >> >> 1. Top level directory proliferation. >> 2. Confusion with /opt/csw/gcc[34]. > > Maybe /opt/csw/lib/gcc would be a better location? I've experimented > with this in the past, with mixed results. Software packages generally > like their layout being modified by --prefix, and not by --libdir. > As I understand it the issue is with ABI interoperability between compiled c++ code from different compilers. As such should we use /opt/csw/lib/gxx rather than gcc? This would be more indicative of gnu c++ libraries rather than gnu libraries in general. It would be nice to avoid a new top level directory such as /opt/csw/gcc as this feels cluttered . From ihsan at opencsw.org Mon Feb 27 15:59:29 2012 From: ihsan at opencsw.org (=?UTF-8?B?xLBoc2FuIERvxJ9hbg==?=) Date: Mon, 27 Feb 2012 15:59:29 +0100 Subject: [csw-maintainers] catalog / package name issues Message-ID: <4F4B9A51.3020101@opencsw.org> Hi, While packaging Email-MIME-Attachment-Stripper, I've run into several issues. The package name suggested by Gar is CSWpm-email-mime-attachment-stripper. Because the name is too long, pkgmk refuses to create a package. If I shorten the package name to CSWpmemailmimeattachmentstripper, checkpkg complains about CHECKPKG_OVERRIDES_CSWpmemailmimeattachmentstripper += pkginfo-bad-catalogname|pm_email_mime-attachment_stripper CHECKPKG_OVERRIDES_CSWpmemailmimeattachmentstripper += catalogname-does-not-match-pkgname|pkgname=CSWpmemailmimeattachmentstripper|catalogname=pm_email_mime-attachment_stripper|expected-catalogname=pmemailmimeattachmentstripper CHECKPKG_OVERRIDES_CSWpmemailmimeattachmentstripper += catalogname-too-long CHECKPKG_OVERRIDES_CSWpmemailmimeattachmentstripper += pkginfo-opencsw-repository-uncommitted What would be recommended name for this package? Ihsan -- ihsan at dogan.ch http://blog.dogan.ch/ From dam at opencsw.org Mon Feb 27 16:07:13 2012 From: dam at opencsw.org (Dagobert Michelsen) Date: Mon, 27 Feb 2012 16:07:13 +0100 Subject: [csw-maintainers] catalog / package name issues In-Reply-To: <4F4B9A51.3020101@opencsw.org> References: <4F4B9A51.3020101@opencsw.org> Message-ID: Hi Ihsan, Am 27.02.2012 um 15:59 schrieb ?hsan Do?an: > While packaging Email-MIME-Attachment-Stripper, I've run into several > issues. The package name suggested by Gar is > CSWpm-email-mime-attachment-stripper. Because the name is too long, > pkgmk refuses to create a package. > > If I shorten the package name to CSWpmemailmimeattachmentstripper, > checkpkg complains about > > CHECKPKG_OVERRIDES_CSWpmemailmimeattachmentstripper += > pkginfo-bad-catalogname|pm_email_mime-attachment_stripper > > CHECKPKG_OVERRIDES_CSWpmemailmimeattachmentstripper += > catalogname-does-not-match-pkgname|pkgname=CSWpmemailmimeattachmentstripper|catalogname=pm_email_mime-attachment_stripper|expected-catalogname=pmemailmimeattachmentstripper > > CHECKPKG_OVERRIDES_CSWpmemailmimeattachmentstripper += catalogname-too-long > > CHECKPKG_OVERRIDES_CSWpmemailmimeattachmentstripper += > pkginfo-opencsw-repository-uncommitted > > What would be recommended name for this package? Please always keep the syntax clean, that means leave the hyphens in place. If necessary, shorten the package name hard at the end (which GAR could have done). That means CSWpm-email-mime-attachment-stri The catalog name should include the full module name. Checkpkg should then be smart enough to not complain (I think it will complain, but that should be fixed then). Best regards -- Dago -- "You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process." - xkcd #896 From ihsan at opencsw.org Tue Feb 28 08:47:18 2012 From: ihsan at opencsw.org (=?UTF-8?B?xLBoc2FuIERvxJ9hbg==?=) Date: Tue, 28 Feb 2012 08:47:18 +0100 Subject: [csw-maintainers] catalog / package name issues In-Reply-To: References: <4F4B9A51.3020101@opencsw.org> Message-ID: <4F4C8686.7000402@opencsw.org> Hi Dago, On 02/27/2012 04:07 PM, Dagobert Michelsen wrote: >> While packaging Email-MIME-Attachment-Stripper, I've run into several >> issues. The package name suggested by Gar is >> CSWpm-email-mime-attachment-stripper. Because the name is too long, >> pkgmk refuses to create a package. >> >> If I shorten the package name to CSWpmemailmimeattachmentstripper, >> checkpkg complains about >> >> CHECKPKG_OVERRIDES_CSWpmemailmimeattachmentstripper += >> pkginfo-bad-catalogname|pm_email_mime-attachment_stripper >> >> CHECKPKG_OVERRIDES_CSWpmemailmimeattachmentstripper += >> catalogname-does-not-match-pkgname|pkgname=CSWpmemailmimeattachmentstripper|catalogname=pm_email_mime-attachment_stripper|expected-catalogname=pmemailmimeattachmentstripper >> >> CHECKPKG_OVERRIDES_CSWpmemailmimeattachmentstripper += catalogname-too-long >> >> CHECKPKG_OVERRIDES_CSWpmemailmimeattachmentstripper += >> pkginfo-opencsw-repository-uncommitted >> >> What would be recommended name for this package? > > Please always keep the syntax clean, that means leave the hyphens in place. > If necessary, shorten the package name hard at the end (which GAR could have > done). That means > CSWpm-email-mime-attachment-stri > > The catalog name should include the full module name. Thanks for the hint, Dago. > Checkpkg should then be smart enough to not complain (I think it will > complain, but that should be fixed then). Checkpkg still complains: CHECKPKG_OVERRIDES_CSWpm-email-mime-attachment-stri += pkginfo-bad-catalogname|pm_email_mime-attachment_stripper CHECKPKG_OVERRIDES_CSWpm-email-mime-attachment-stri += catalogname-does-not-match-pkgname|pkgname=CSWpm-email-mime-attachment-stri|catalogname=pm_email_mime-attachment_stripper|expected-catalogname=pm_email_mime_attachment_stri CHECKPKG_OVERRIDES_CSWpm-email-mime-attachment-stri += catalogname-too-long I guess, I can override these? Ihsan -- ihsan at dogan.ch http://blog.dogan.ch/ From bonivart at opencsw.org Tue Feb 28 09:05:41 2012 From: bonivart at opencsw.org (Peter Bonivart) Date: Tue, 28 Feb 2012 09:05:41 +0100 Subject: [csw-maintainers] catalog / package name issues In-Reply-To: <4F4C8686.7000402@opencsw.org> References: <4F4B9A51.3020101@opencsw.org> <4F4C8686.7000402@opencsw.org> Message-ID: 2012/2/28 ?hsan Do?an : > CHECKPKG_OVERRIDES_CSWpm-email-mime-attachment-stri += > catalogname-does-not-match-pkgname|pkgname=CSWpm-email-mime-attachment-stri|catalogname=pm_email_mime-attachment_stripper|expected-catalogname=pm_email_mime_attachment_stri You still have a dash in the catalog name "mime-attachment". /peter From pfelecan at opencsw.org Tue Feb 28 09:26:59 2012 From: pfelecan at opencsw.org (Peter FELECAN) Date: Tue, 28 Feb 2012 09:26:59 +0100 Subject: [csw-maintainers] C++ libraries built with GCC In-Reply-To: (Jonathan Craig's message of "Mon, 27 Feb 2012 08:09:41 -0500") References: <1330278296-sup-7303@pinkfloyd.chass.utoronto.ca> <20120226191929.GA8597@quince.home.blizinski.pl> Message-ID: Jonathan Craig writes: > On Sun, Feb 26, 2012 at 2:19 PM, Maciej Blizi?ski wrote: >> Ben Walton wrote: >>> Excerpts from Maciej (Matchek) Blizi?ski's message of Sun Feb 26 09:18:12 -0500 2012: >>> >>> Hi Maciej, >>> >>> The idea is good and we should definitely accommodate this nicely. ?I >>> don't like /opt/csw/gcc, but I haven't got a better name that doesn't >>> also suffer from: >>> >>> 1. Top level directory proliferation. >>> 2. Confusion with /opt/csw/gcc[34]. >> >> Maybe /opt/csw/lib/gcc would be a better location? I've experimented >> with this in the past, with mixed results. Software packages generally >> like their layout being modified by --prefix, and not by --libdir. >> > > As I understand it the issue is with ABI interoperability between > compiled c++ code from different compilers. As such should we use > /opt/csw/lib/gxx rather than gcc? This would be more indicative of > gnu c++ libraries rather than gnu libraries in general. It would be > nice to avoid a new top level directory such as /opt/csw/gcc as this > feels cluttered . +1 for gxx /opt/csw/gxx is also a good prefix; what's the point with cluttered top level? -- Peter From ihsan at opencsw.org Tue Feb 28 09:43:57 2012 From: ihsan at opencsw.org (=?UTF-8?B?xLBoc2FuIERvxJ9hbg==?=) Date: Tue, 28 Feb 2012 09:43:57 +0100 Subject: [csw-maintainers] catalog / package name issues In-Reply-To: References: <4F4B9A51.3020101@opencsw.org> <4F4C8686.7000402@opencsw.org> Message-ID: <4F4C93CD.3030303@opencsw.org> On 02/28/2012 09:05 AM, Peter Bonivart wrote: >> CHECKPKG_OVERRIDES_CSWpm-email-mime-attachment-stri += >> catalogname-does-not-match-pkgname|pkgname=CSWpm-email-mime-attachment-stri|catalogname=pm_email_mime-attachment_stripper|expected-catalogname=pm_email_mime_attachment_stri > > You still have a dash in the catalog name "mime-attachment". Argh. Thanks. Working before having coffee doesn't seem to be a good idea. :-) Still got the complaint, that the package name is too long. Ihsan -- ihsan at dogan.ch http://blog.dogan.ch/ From dam at opencsw.org Tue Feb 28 10:27:16 2012 From: dam at opencsw.org (Dagobert Michelsen) Date: Tue, 28 Feb 2012 10:27:16 +0100 Subject: [csw-maintainers] catalog / package name issues In-Reply-To: <4F4C8686.7000402@opencsw.org> References: <4F4B9A51.3020101@opencsw.org> <4F4C8686.7000402@opencsw.org> Message-ID: <4CD93121-7DDD-4B01-91A9-803A8BE269BB@opencsw.org> Hi Ihsan, Am 28.02.2012 um 08:47 schrieb ?hsan Do?an: > On 02/27/2012 04:07 PM, Dagobert Michelsen wrote: >> Please always keep the syntax clean, that means leave the hyphens in place. >> If necessary, shorten the package name hard at the end (which GAR could have >> done). That means >> CSWpm-email-mime-attachment-stri >> >> The catalog name should include the full module name. > > Thanks for the hint, Dago. > >> Checkpkg should then be smart enough to not complain (I think it will >> complain, but that should be fixed then). > > Checkpkg still complains: > > CHECKPKG_OVERRIDES_CSWpm-email-mime-attachment-stri += > pkginfo-bad-catalogname|pm_email_mime-attachment_stripper I tend to relax the constraints on catalog name length. Thoughts? > CHECKPKG_OVERRIDES_CSWpm-email-mime-attachment-stri += > catalogname-does-not-match-pkgname|pkgname=CSWpm-email-mime-attachment-stri|catalogname=pm_email_mime-attachment_stripper|expected-catalogname=pm_email_mime_attachment_stri Maciej: I think this check should only take the max number of chars of the package name into account. > CHECKPKG_OVERRIDES_CSWpm-email-mime-attachment-stri += catalogname-too-long Same as above. > I guess, I can override these? Probably we need to talk about enhancing the catalog name length. Best regards -- Dago -- "You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process." - xkcd #896 From maciej at opencsw.org Tue Feb 28 10:48:17 2012 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Tue, 28 Feb 2012 09:48:17 +0000 Subject: [csw-maintainers] catalog / package name issues In-Reply-To: <4CD93121-7DDD-4B01-91A9-803A8BE269BB@opencsw.org> References: <4F4B9A51.3020101@opencsw.org> <4F4C8686.7000402@opencsw.org> <4CD93121-7DDD-4B01-91A9-803A8BE269BB@opencsw.org> Message-ID: 2012/2/28 Dagobert Michelsen : > Maciej: I think this check should only take the max number of chars of > the package name into account. Which of the many package name concepts do you mean? From dam at opencsw.org Tue Feb 28 10:52:01 2012 From: dam at opencsw.org (Dagobert Michelsen) Date: Tue, 28 Feb 2012 10:52:01 +0100 Subject: [csw-maintainers] catalog / package name issues In-Reply-To: References: <4F4B9A51.3020101@opencsw.org> <4F4C8686.7000402@opencsw.org> <4CD93121-7DDD-4B01-91A9-803A8BE269BB@opencsw.org> Message-ID: <19C07446-D8B8-4492-BFCF-31FA0E5A1A41@opencsw.org> Hi Maciej, Am 28.02.2012 um 10:48 schrieb Maciej (Matchek) Blizi?ski: > 2012/2/28 Dagobert Michelsen : >> Maciej: I think this check should only take the max number of chars of >> the package name into account. > > Which of the many package name concepts do you mean? IIRC we currently have 32 chars max. for package names and 29 chars max. for the catalog name where usually the package name should equal the catalog name with the CSW prefix stripped and "-" replaced with "_". I tend to relax the length of catalog names to whatever is necessary to give Perl modules a 1:1 naming. The package name would be stripped to 32 chars, though. Best regards -- Dago -- "You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process." - xkcd #896 From maciej at opencsw.org Tue Feb 28 10:51:52 2012 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Tue, 28 Feb 2012 09:51:52 +0000 Subject: [csw-maintainers] Running git send-email on the buildfarm Message-ID: I want to send my patches upstream. I'm running git send-email, but it's taking an awful lot of time, or maybe it's stuck forever (hard to tell, I waited a few minutes). maciej at login [login]:~/src/vlc > git send-email --in-reply-to="CALtRa-5qU9QZD=Vp_2-LceLy4wrjV2yvCvj20ip3d9eaa3BcsQ at mail.gmail.com" 0001-Test-for-posix_memalign-in-.-configure.patch (...) Send this email? ([y]es|[n]o|[q]uit|[a]ll): y And it's stuck there, the prompt does not return. Am I doing it wrong? What are other people's experiences? From maciej at opencsw.org Tue Feb 28 11:08:56 2012 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Tue, 28 Feb 2012 10:08:56 +0000 Subject: [csw-maintainers] catalog / package name issues In-Reply-To: <19C07446-D8B8-4492-BFCF-31FA0E5A1A41@opencsw.org> References: <4F4B9A51.3020101@opencsw.org> <4F4C8686.7000402@opencsw.org> <4CD93121-7DDD-4B01-91A9-803A8BE269BB@opencsw.org> <19C07446-D8B8-4492-BFCF-31FA0E5A1A41@opencsw.org> Message-ID: Alright, maybe we need to settle the terminology. I tend to call them 'pkgnames', as opposed to package names, because the word 'package' is so overloaded it barely has any meaning for us anymore. If you mean what I call pkgname, feel free to adjust settings in checkpkg and announce on the list. Maciej -------------- next part -------------- An HTML attachment was scrubbed... URL: From maciej at opencsw.org Tue Feb 28 13:28:44 2012 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Tue, 28 Feb 2012 12:28:44 +0000 Subject: [csw-maintainers] What to do when upstream declares Solaris unsupported Message-ID: I've been poking VLC in the last days just because it was fun to fiddle with autotools. I was working around missing posix_memalign, and when someone on the vlc-devel list sent me a patch that redefines this function. I suggested testing for posix_memalign and saw this reply: "It is not optional. Why would we test for something that is always there except on broken systems that we do not support?" http://mailman.videolan.org/pipermail/vlc-devel/2012-February/086834.html I didn't see much about Solaris on the VLC website, although they say it works on Unix. What can we reply to the "unsupported" remark? Maciej From dam at opencsw.org Tue Feb 28 14:29:19 2012 From: dam at opencsw.org (Dagobert Michelsen) Date: Tue, 28 Feb 2012 14:29:19 +0100 Subject: [csw-maintainers] What to do when upstream declares Solaris unsupported In-Reply-To: References: Message-ID: Hi Maciej, Am 28.02.2012 um 13:28 schrieb Maciej (Matchek) Blizi?ski: > I've been poking VLC in the last days just because it was fun to > fiddle with autotools. I was working around missing posix_memalign, > and when someone on the vlc-devel list sent me a patch that redefines > this function. I suggested testing for posix_memalign and saw this > reply: > > "It is not optional. Why would we test for something that is always there > except on broken systems that we do not support?" > > http://mailman.videolan.org/pipermail/vlc-devel/2012-February/086834.html > > I didn't see much about Solaris on the VLC website, although they say > it works on Unix. What can we reply to the "unsupported" remark? "Well, the VLC webpage says that 'VLC works on most platforms: Windows, Linux, Mac OS X, Unix...' and I am offering to push 'most' to one more of the Unix platforms. Is there something wrong in fixing support for a platform?" If they say "We don't want to clutter our code with obsoleted tests" than you are out of luck, though. :-) Best regards -- Dago -- "You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process." - xkcd #896 From Joerg.Schilling at fokus.fraunhofer.de Tue Feb 28 14:39:20 2012 From: Joerg.Schilling at fokus.fraunhofer.de (Joerg Schilling) Date: Tue, 28 Feb 2012 14:39:20 +0100 Subject: [csw-maintainers] What to do when upstream declares Solaris unsupported In-Reply-To: References: Message-ID: <4f4cd908.jpa0ddNNkBih8dEs%Joerg.Schilling@fokus.fraunhofer.de> Dagobert Michelsen wrote: > > I didn't see much about Solaris on the VLC website, although they say > > it works on Unix. What can we reply to the "unsupported" remark? > > "Well, the VLC webpage says that 'VLC works on most platforms: > Windows, Linux, Mac OS X, Unix...' and I am offering to push > 'most' to one more of the Unix platforms. Is there something > wrong in fixing support for a platform?" > > If they say "We don't want to clutter our code with obsoleted tests" > than you are out of luck, though. If they did a good job, then their code should of course work on a standard compliant Solaris. If not, they have a problem in their code. J?rg -- EMail:joerg at schily.isdn.cs.tu-berlin.de (home) J?rg Schilling D-13353 Berlin js at cs.tu-berlin.de (uni) joerg.schilling at fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily From bwalton at opencsw.org Tue Feb 28 15:37:00 2012 From: bwalton at opencsw.org (Ben Walton) Date: Tue, 28 Feb 2012 09:37:00 -0500 Subject: [csw-maintainers] Running git send-email on the buildfarm In-Reply-To: References: Message-ID: <1330439759-sup-9902@pinkfloyd.chass.utoronto.ca> Excerpts from Maciej (Matchek) Blizi?ski's message of Tue Feb 28 04:51:52 -0500 2012: > And it's stuck there, the prompt does not return. > > Am I doing it wrong? What are other people's experiences? I got a spew of cron output from rss/webdb/mantisdb updates last night. They all had problems pulling json data via rest calls to the buildfarm. I suspect some sort of memory exhaustion globally there. Dago? Thanks -Ben -- Ben Walton Systems Programmer - CHASS University of Toronto C:416.407.5610 | W:416.978.4302 From bwalton at opencsw.org Tue Feb 28 15:44:06 2012 From: bwalton at opencsw.org (Ben Walton) Date: Tue, 28 Feb 2012 09:44:06 -0500 Subject: [csw-maintainers] C++ libraries built with GCC In-Reply-To: References: <1330278296-sup-7303@pinkfloyd.chass.utoronto.ca> <20120226191929.GA8597@quince.home.blizinski.pl> Message-ID: <1330440134-sup-9283@pinkfloyd.chass.utoronto.ca> Excerpts from Peter FELECAN's message of Tue Feb 28 03:26:59 -0500 2012: Hi Peter, > +1 for gxx That works. > /opt/csw/gxx is also a good prefix; what's the point with cluttered top > level? For me it's two things, both of which I can live with but prefer to avoid if possible. 1. It can and has caused pain in the past when we add extra top level things...this is mostly with specific packages (apache2, php5, etc) and has less chance of negative long-term impact for a case like this. 2. Aesthetics. It just feels ugly to me. If it's the best option available though and it solves the problem, I'm ok with it. They're just things that I'd avoid 'if possible.' Thanks -Ben -- Ben Walton Systems Programmer - CHASS University of Toronto C:416.407.5610 | W:416.978.4302 From dam at opencsw.org Tue Feb 28 16:22:16 2012 From: dam at opencsw.org (Dagobert Michelsen) Date: Tue, 28 Feb 2012 16:22:16 +0100 Subject: [csw-maintainers] Running git send-email on the buildfarm In-Reply-To: <1330439759-sup-9902@pinkfloyd.chass.utoronto.ca> References: <1330439759-sup-9902@pinkfloyd.chass.utoronto.ca> Message-ID: <84334370-239E-41C3-B2F8-D0DFC4569218@opencsw.org> Hi Ben, Am 28.02.2012 um 15:37 schrieb Ben Walton: > Excerpts from Maciej (Matchek) Blizi?ski's message of Tue Feb 28 04:51:52 -0500 2012: >> And it's stuck there, the prompt does not return. >> >> Am I doing it wrong? What are other people's experiences? > > I got a spew of cron output from rss/webdb/mantisdb updates last > night. They all had problems pulling json data via rest calls to the > buildfarm. I suspect some sort of memory exhaustion globally > there. Do you have more specific data for me? The only errors I can see are: > [Tue Feb 28 04:39:34 2012] [error] [client 213.178.77.176] File does not exist: /opt/csw/apache2/share/htdocs/buildfarm/opencsw/unstable/sparc/5.10.2 > [Tue Feb 28 04:39:35 2012] [error] [client 213.178.77.176] File does not exist: /opt/csw/apache2/share/htdocs/buildfarm/opencsw/unstable/sparc/5.10.2 > [Tue Feb 28 04:39:38 2012] [error] [client 213.178.77.176] File does not exist: /opt/csw/apache2/share/htdocs/buildfarm/opencsw/unstable/i386/5.10.2 > [Tue Feb 28 04:39:38 2012] [error] [client 213.178.77.176] File does not exist: /opt/csw/apache2/share/htdocs/buildfarm/opencsw/unstable/i386/5.10.2 > [Tue Feb 28 04:39:41 2012] [error] [client 213.178.77.176] File does not exist: /opt/csw/apache2/share/htdocs/buildfarm/opencsw/testing/sparc/5.10.2 > [Tue Feb 28 04:39:41 2012] [error] [client 213.178.77.176] File does not exist: /opt/csw/apache2/share/htdocs/buildfarm/opencsw/testing/sparc/5.10.2 > [Tue Feb 28 04:39:43 2012] [error] [client 213.178.77.176] File does not exist: /opt/csw/apache2/share/htdocs/buildfarm/opencsw/testing/i386/5.10.2 > [Tue Feb 28 04:39:43 2012] [error] [client 213.178.77.176] File does not exist: /opt/csw/apache2/share/htdocs/buildfarm/opencsw/testing/i386/5.10.2 Best regards -- Dago -- "You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process." - xkcd #896 From jh at opencsw.org Wed Feb 29 10:00:09 2012 From: jh at opencsw.org (Jan Holzhueter) Date: Wed, 29 Feb 2012 10:00:09 +0100 Subject: [csw-maintainers] 389admin depends on ap2_worker In-Reply-To: <84334370-239E-41C3-B2F8-D0DFC4569218@opencsw.org> References: <1330439759-sup-9902@pinkfloyd.chass.utoronto.ca> <84334370-239E-41C3-B2F8-D0DFC4569218@opencsw.org> Message-ID: <4F4DE919.1020504@opencsw.org> Hi, I just noticed that 389admin depends on ap2_worker. Is there a special reason for this? Or would it work with the prefork mpm too? Just to keep it simple it would be good to depend on the normal apache2 package. If possible that is Greetings Jan From grzemba at contac-dt.de Mon Feb 20 09:21:00 2012 From: grzemba at contac-dt.de (Carsten Grzemba) Date: Mon, 20 Feb 2012 09:21:00 +0100 Subject: [csw-maintainers] Dublin release highlights In-Reply-To: <7490cbaf30d2.4f42010a@contac-dt.de> References: <7490cbaf30d2.4f42010a@contac-dt.de> Message-ID: <7460ee13cd7.4f42107c@contac-dt.de> The 389 Directory Server Packages in version 1.2.9.9 are complete. For usage there are some notes at wiki: http://wiki.opencsw.org/project-389directoryserver. Besides the packages NSS and NSPR are updated on 3.12.11 respectively 4.8.9 and the Mozilla LDAP SDK and Tools v6.0.7 are available. Am 16.02.12, schrieb "Maciej (Matchek) Blizinski" : > Hello maintainers, > > With the upcoming dublin release, we'd like to write an article > outlining its main highlights. I'd like to ask everyone to send in > their highlights of the dublin release. Something that was significant > and you'd like it mentioned. For example, I'd put forward the GCC > update: version upgrade and moving it from /opt/csw/gcc4 to /opt/csw. > We also had a collective effort to repackage shared libraries. I'm > sure there's more and that every maintainer has something significant > that they did in the last year. Please send them in, either on or off > the list, by the end of the next week, which is the 24th of February. > > Maciej > _______________________________________________ > maintainers mailing list > maintainers at lists.opencsw.org > https://lists.opencsw.org/mailman/listinfo/maintainers > .:: This mailing list's archive is public. ::. > > -- Carsten Grzemba Tel.:?? +49 3677 64740 Mobil: +49 171 9749479 Fax::?? +49 3677 6474111 Email: carsten.grzemba at contac-dt.de contac Datentechnik GmbH -------------- next part -------------- An HTML attachment was scrubbed... URL: From grzemba at contac-dt.de Wed Feb 29 10:21:55 2012 From: grzemba at contac-dt.de (Carsten Grzemba) Date: Wed, 29 Feb 2012 10:21:55 +0100 Subject: [csw-maintainers] 389admin depends on ap2_worker In-Reply-To: <73f0b87e2345.4f4dee10@contac-dt.de> References: <73f0b87e2345.4f4dee10@contac-dt.de> Message-ID: <73f0e6a36c49.4f4dfc43@contac-dt.de> It works on prefork, but it give a warning on start "authentication cache not working properly" or something, so the worker model is recommended. Greetings Carsten Am 29.02.12, schrieb Jan Holzhueter : > Hi, > I just noticed that 389admin depends on ap2_worker. > Is there a special reason for this? > Or would it work with the prefork mpm too? > Just to keep it simple it would be good to depend on the normal apache2 > package. If possible that is > > Greetings > Jan > _______________________________________________ > maintainers mailing list > maintainers at lists.opencsw.org > https://lists.opencsw.org/mailman/listinfo/maintainers > .:: This mailing list's archive is public. ::. > > -------------- next part -------------- An HTML attachment was scrubbed... URL: