GNOME Bugzilla – Bug 426609
po/Makefile.in.in depends on gmake's shell feature
Last modified: 2009-02-12 00:21:33 UTC
Please describe the problem: Pan 0.126 requires gmake to process po/Makefile.in.in. It appears to be the same problem reported by Per Hedeland on 4 Feb 2007 and in Bug 404515. The 0.126 version of Makefile.in.in (as well as 0.122) depends on gmake's shell function which makes it non-portable to systems using other make tools. Steps to reproduce: 1. 2. 3. Actual results: Expected results: Does this happen every time? Yes Other information: Here's some further details: x2:tmp/pan 609>grep shell pan-0.12?/po/Makefile.in.in pan-0.122/po/Makefile.in.in:PO_LINGUAS=$(shell if test -r $(srcdir)/LINGUAS; then grep -v "^\#" $(srcdir)/LINGUAS; fi) pan-0.122/po/Makefile.in.in:POFILES=$(shell if test -n "$(PO_LINGUAS)"; then LINGUAS="$(PO_LINGUAS)"; else LINGUAS="$(ALL_LINGUAS)"; fi; for lang in $$LINGUAS; do printf "$$lang.po "; done) pan-0.122/po/Makefile.in.in:CATALOGS=$(shell if test -n "$(PO_LINGUAS)"; then LINGUAS="$(PO_LINGUAS)"; else LINGUAS="$(ALL_LINGUAS)"; fi; for lang in $$LINGUAS; do printf "$$lang.gmo "; done) pan-0.126/po/Makefile.in.in:PO_LINGUAS=$(shell if test -r $(srcdir)/LINGUAS; then grep -v "^\#" $(srcdir)/LINGUAS; fi) pan-0.126/po/Makefile.in.in:POFILES=$(shell if test -n "$(PO_LINGUAS)"; then LINGUAS="$(PO_LINGUAS)"; else LINGUAS="$(ALL_LINGUAS)"; fi; for lang in $$LINGUAS; do printf "$$lang.po "; done) pan-0.126/po/Makefile.in.in:CATALOGS=$(shell if test -n "$(PO_LINGUAS)"; then LINGUAS="$(PO_LINGUAS)"; else LINGUAS="$(ALL_LINGUAS)"; fi; for lang in $$LINGUAS; do printf "$$lang.gmo "; done) x2:tmp/pan 610> x2:tmp/pan 610>diff -wbu pan-0.12[26]/po/Makefile.in.in|wc 0 0 0 x2:tmp/pan 611>ls -l pan-0.12[26]/po/Makefile.in.in|less x2:tmp/pan 612>ls -l pan-0.12[26]/po/Makefile.in.in -rw-r--r-- 1 rd wsrc 6578 Feb 3 14:02 pan-0.122/po/Makefile.in.in -rw-r--r-- 1 rd wsrc 6578 Mar 28 12:33 pan-0.126/po/Makefile.in.in x2:tmp/pan 613>ls -l pan-0.12?/po/Makefile.in.in -rw-r--r-- 1 rd wsrc 6578 Feb 3 14:02 pan-0.122/po/Makefile.in.in -rwxr-xr-x 1 rd wsrc 7722 Feb 6 09:27 pan-0.123/po/Makefile.in.in -rwxr-xr-x 1 rd wsrc 7722 Feb 12 15:36 pan-0.124/po/Makefile.in.in -rwxr-xr-x 1 rd wsrc 7722 Feb 12 15:36 pan-0.125/po/Makefile.in.in -rw-r--r-- 1 rd wsrc 6578 Mar 28 12:33 pan-0.126/po/Makefile.in.in x2:tmp/pan 614>md5 pan-0.12?/po/Makefile.in.in MD5 (pan-0.122/po/Makefile.in.in) = 559e59b813648b9b8c6f0c6ee444c795 MD5 (pan-0.123/po/Makefile.in.in) = f7b4f9ee98c0238ad33db2cb4814e02c MD5 (pan-0.124/po/Makefile.in.in) = f7b4f9ee98c0238ad33db2cb4814e02c MD5 (pan-0.125/po/Makefile.in.in) = f7b4f9ee98c0238ad33db2cb4814e02c MD5 (pan-0.126/po/Makefile.in.in) = 559e59b813648b9b8c6f0c6ee444c795
Back when this was reported in Bug 404515 I assumed it was because I was on a hacked-up box, but the Pan 0.126 tarball was built on a stock RHEL5 box, so my next guess is that it's something wrong with the gnome-common or intltool. Bumping upstream for a second opinion.
*** Bug 430781 has been marked as a duplicate of this bug. ***
Makefile.in.in comes from intltool. -> back to intltool
Hello all, great job on pan, I thank you. :) My bug has been marked as a duplicate of this one, so I will provide what information I have that seems pertinent. Bug 430781 describes some symptoms when "make"ing on MacOS X using (an old) fink unstable that worked to build 0.125. As you can see below, I am using GNU Make, so if my bug is truly a duplicate of this one, it may be appropriate to expand the scope of your considerations. If I can help more (any?) please let me know, I'm on the CC list for this bug (Thanks Charles). -- Jerry ************* Hardware and software **************** MacOS X 10.4.9 (8P135) on powerbook hardware as follows Machine Name: PowerBook G4 17" Machine Model: PowerBook5,5 CPU Type: PowerPC G4 (1.1) Number Of CPUs: 1 CPU Speed: 1.5 GHz L2 Cache (per CPU): 512 KB Memory: 1.5 GB Bus Speed: 167 MHz Boot ROM Version: 4.8.4f1 Pan 0.128 (Source from http://pan.rebelbase.com/download/releases/0.128/source/) GTK+ 2.6.10-1002 (fink unstable) Gmime 2.2.2-1 (fink unstable) GTKSpell 2.0.11-1002 (fink unstable) PCRE 6.4 (Manually added to fink. I don't remember where I got it, sorry) GNU Make 3.80 (From Apple) ************* END Hardware and software **************** ************* Bash output ************* jjlaptop:~/Applications/pan-0.128 jj$ which make /usr/bin/make jjlaptop:~/Applications/pan-0.128 jj$ /usr/bin/make -v GNU Make 3.80 <snip> jjlaptop:~/Applications/pan-0.128/po jj$ md5 Makefile.in.in MD5 (Makefile.in.in) = 559e59b813648b9b8c6f0c6ee444c795 ************* END Bash output ************* ************* MD5 hashes review ************* for ease in comparison, here are the md5 hashes from before along with the one I am adding for 0.128. MD5 (pan-0.122/po/Makefile.in.in) = 559e59b813648b9b8c6f0c6ee444c795 MD5 (pan-0.123/po/Makefile.in.in) = f7b4f9ee98c0238ad33db2cb4814e02c MD5 (pan-0.124/po/Makefile.in.in) = f7b4f9ee98c0238ad33db2cb4814e02c MD5 (pan-0.125/po/Makefile.in.in) = f7b4f9ee98c0238ad33db2cb4814e02c MD5 (pan-0.126/po/Makefile.in.in) = 559e59b813648b9b8c6f0c6ee444c795 MD5 (pan-0.128/po/Makefile.in.in) = 559e59b813648b9b8c6f0c6ee444c795
I don't see any way to do what we need in Makefile.in.in and not use the $(shell) execution feature. All of the @INTLTOOL_FOO_RULE@ rules also depend on the % variable replace feature of gmake as well. If anyone can point me to features that let me achieve what we are doing here, I would be glad to implement them, but I can't do anything without more info.
Closing this bug report as no further information has been provided. Please feel free to reopen this bug if you can provide the information asked for. Thanks!