[gs-bugs] [Bug 691427] 'make so' broken again with icc/lcms merge using special LCMS_CC, etc.

bugzilla-daemon at ghostscript.com bugzilla-daemon at ghostscript.com
Fri Jul 2 00:37:12 UTC 2010


http://bugs.ghostscript.com/show_bug.cgi?id=691427

Hin-Tak Leung <hintak at ghostscript.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |
         AssignedTo|hintak at ghostscript.com      |michael.vrhel at artifex.com

--- Comment #4 from Hin-Tak Leung <hintak at ghostscript.com> 2010-07-02 00:37:10 UTC ---
reopening and assign to Michael to review the revert r11451  which requires
r11460 and r11461 in ghostpdl to work.

I have a 2nd thought and think that if I had been more verbose with
base/bcwin32.mak, base/msvclib.mak, base/watclib.mak, base/watcw32.mak,
psi/msvc32.mak, and do:

+!ifndef LCMSPLATFORM 
+LCMSPLATFORM=win32
+!endif

instead of rolling it inside LCMSSRCDIR, 

 !ifndef LCMSSRCDIR
 LCMSSRCDIR=lcms
+LCMSPLATFORM=win32
 !endif

r11461 may not be needed, but I haven't tested this idea yet.

In any case, the change is to fix the "quick fix" of using "GLLCMSCC=$(CC) ..."
(instead of ".. $(CC_)...") to allow for microsoft language extensions. On many
kind of unix systems (and x86_64 linux in particular, but also Solaris etc for
which there are a few more "make so not working" bugs), object files intended
for linking into the final shared libraries needs to have a central CFLAGS
(-fPIC for gcc but -KPIC for Solaris) and other machineries involving LDFLAGS,
so it is best not to override the  central compiler/linker flags for specific
files; if it is necessary, make it explicitly platform specific as done here,
and provide generic ("no override") targets for other platforms.

The per-module *CC's ideally should only add includes/libraries, rather than
dropping/enabling language/platform specific extensions. I hope this is clear
about the intention of the patch.

-- 
Configure bugmail: http://bugs.ghostscript.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.


More information about the gs-bugs mailing list