/usr/share/aclocal/pstoedit.m4 no longer contains correct definition of AM_PATH_PSTOEDIT macro. That harms other builds, like media-gfx/autotrace. Reproducible: Always Steps to Reproduce: 1. Emerge media-gfx/pstoedit-3.44 2. Try to emerge media-gfx/autotrace with USE="pdf"; there may be other cases. Actual Results: Emerge fails with autoconf complaining of illegal AM_PATH_PSTOEDIT macro. Expected Results: The package should emerge. Between 3.42 and 3.44 versions of pstoedit, the underquoted definition problem was apparently dealt with upstream not by changing AC_DEFUN(AM_PATH_PSTOEDIT, to AC_DEFUN([AM_PATH_PSTOEDIT], but by removing the start of the macro definition. Or the pstoedit.m4 file got corrupted or whatever. The definition must be restored. For the time being it means to change the patch used in the pstoedit-3.42 ebuild slightly, from correcting a wrong start of the macro definition to inserting its correct start.
Created attachment 117070 [details, diff] pstoedit-3.44.ebuild modification to correct pstoedit.m4
Created attachment 117072 [details, diff] pstoedit-3.44-m4.patch to correct pstoedit.m4
hmm did you try to contact pstoedit upstream about this ? Perhaps they just dropped AM_PATH_PSTOEDIT for some reason I dont know. One thing is sure is that I dont like patching m4's files in a way that will make it differ from upstream, that will either lure our users that the function still exists or cause problems with packages made to work against the original one.
(In reply to comment #3) > hmm did you try to contact pstoedit upstream about this ? > Perhaps they just dropped AM_PATH_PSTOEDIT for some reason I dont know. The installed version of pstoedit.m4 doesn't even make sense syntactically. If removal of the 'AC_DEFUN(AM_PATH_PSTOEDIT' was deliberate, they still left the terminating ')' at the end of the file. I'll send an email to the author of pstoedit and ask him about it.
(In reply to comment #4) > I'll send an email to the author of pstoedit and ask him about it. I got this reply today from the author of pstoedit, Wolfgang Glunz: > Hi Ian, > > Yes - indeed the current version seems to be broken and your patch seems to > fix it. > Thanks - I will take this into the next release > > BR > > WOlfgang > >> -----Original Message----- >> From: Ian Abbott [mailto:abbotti@mev.co.uk] >> Sent: Wednesday, March 26, 2008 4:39 PM >> To: Wolfgang Glunz >> Subject: pstoedit.m4 in pstoedit-3.45 broken? >> >> >> Dear Wolfgang, >> >> The version of the pstoedit.m4 file installed by >> pstoedit-3.44 and pstoedit-3.45 appears to be broken. Could >> you confirm that the attached patch from the Gentoo Linux >> bugzilla is correct? >> >> pstoedit-3.44-m4.patch >> <http://bugs.gentoo.org/attachment.cgi?id=117072> >> Gentoo bug #175679 <http://bugs.gentoo.org/show_bug.cgi?id=175679> >> >> Best regards, >> Ian Abbott.
Note that this bug also applies to pstoedit-3.45. Should I create a separate bug report for 3.45? It's still a blocker for bugs reported in other programs, such as media-gfx/autotrace (bug #175414).
I verified the m4 in 3.50 incorporates this patch. Fixed in 3.44-r1, forced straight to stable as this is causing breakage in other stable packages (and probably should have been fixed a long time ago).