<html><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; "><font class="Apple-style-span" face="Courier">Hi Chad,</font><div><font class="Apple-style-span" face="Courier"><br></font><div><div><font class="Apple-style-span" face="Courier">Am 01.04.2009 um 19:15 schrieb Chad Harp:</font></div><blockquote type="cite"><div><font class="Apple-style-span" face="Courier">I think I need a better understanding of what I should be doing.</font></div></blockquote><blockquote type="cite"><div><font class="Apple-style-span" face="Courier">Several of my packages have path problems due to quoting. Should</font></div></blockquote><blockquote type="cite"><div><font class="Apple-style-span" face="Courier">I be looking for the correct combinations of variable to fix it,</font></div></blockquote><blockquote type="cite"><div><font class="Apple-style-span" face="Courier">or will GAR be fixed to deal with it?<br></font></div></blockquote><font class="Apple-style-span" face="Courier"><br></font></div><div><font class="Apple-style-span" face="Courier">Ok, now here is the rationale why things are as they are now:</font></div><div><font class="Apple-style-span" face="Courier"><br></font></div><div><font class="Apple-style-span" face="Courier">- /opt/csw/lib/$ISALIST is needed if and only if libraries are</font></div><div><font class="Apple-style-span" face="Courier"> linked that may contain optimizations. Otherwise it is best to</font></div><div><font class="Apple-style-span" face="Courier"> avoid $ISALIST completely. This can be done by setting</font></div><div><font class="Apple-style-span" face="Courier"> NOISALIST = 1</font></div><div><font class="Apple-style-span" face="Courier"><br></font></div><div><font class="Apple-style-span" face="Courier">- I put -R for runtime linker path in LDFLAGS because in the last</font></div><div><font class="Apple-style-span" face="Courier"> discussion LD_OPTIONS was considered bad, as it is always applied</font></div><div><font class="Apple-style-span" face="Courier"> first in linking regardless of other flags. Unfortunately these</font></div><div><font class="Apple-style-span" face="Courier"> flags are passwd through multiple commands, each with evaluation</font></div><div><font class="Apple-style-span" face="Courier"> of the argument. Now $ISALIST looks like a variable and must be</font></div><div><font class="Apple-style-span" face="Courier"> shielded from evaluation by quoting. The amount of quoting depends</font></div><div><font class="Apple-style-span" face="Courier"> on the number of evaluation, e. g. if libtool is used, a custom</font></div><div><font class="Apple-style-span" face="Courier"> Makefile, etc.</font></div><div><font class="Apple-style-span" face="Courier"> The number of quotings can be adjusted by the simple parameter</font></div><div><font class="Apple-style-span" face="Courier"> RUNPATHQUOTE = 1 for single evaluation</font></div><div><font class="Apple-style-span" face="Courier"> RUNPATHQUOTE = 2 for double evaluation</font></div><div><font class="Apple-style-span" face="Courier"> Unfortunately this obviously proved to be not sufficient as the</font></div><div><font class="Apple-style-span" face="Courier"> bug reports from James show.</font></div><div><font class="Apple-style-span" face="Courier"><br></font></div><div><font class="Apple-style-span" face="Courier"> (BTW, if you are interested: Making the literal '$' as prefix is</font></div><div><font class="Apple-style-span" face="Courier"><div> _Q = \\\$$\$$ for RUNPATHQUOTE = 1</div><div> and</div><div> _Q = \\\\\\\$$\$$ for RUNPATHQUOTE = 2</div><div> in gar.conf.mk).</div><div><br></div><div>- Moving -R to LD_OPTIONS would make things considerable easier,</div><div> because the number of evaluations is fixed as it is not processed</div><div> by multiple commands. It would just be needed to decide if</div><div> $ISALIST should be in or not. NOISALIST is already there for this.</div><div> This would conflict with packages already using LD_OPTIONS:</div><div><div> amarok</div><div> cups</div><div> daimonin</div><div> exim</div><div> gftp</div><div> gphoto2</div><div> hatari</div><div> hypermail</div><div> kdesvn</div><div> kile</div><div> libgphoto2</div><div> lyx</div><div> mysql-python</div><div> netsnmp</div><div> slrn</div><div> tsclient</div><div> w3m</div><div> To be consistent, EXTRA_LD_OPTIONS should be introduced and the</div><div> Makefiles adjusted accordingly when the GAR modifications have</div><div> been made.</div><div><br></div></div><div><br></div><div>So the GAR integration is possible, but needs some experimentation</div><div>in quoting and best use.</div><div><br></div><div><br></div><div>Best regards</div><div><br></div><div> -- Dago</div></font></div></div></body></html>