cgdb 0.6.3 has been released some time ago. The 0.6.2 ebuild works almost without any changes, only the ${P}-parallel-make.patch needs to be removed.
Created attachment 100270 [details] ebuild for cgdb-0.6.3 I see nobody modified an ebuild for the new version, so i'm uploading this one, which works on my linux machine.
I suppose the dev-tools people are capable to implement my "...patch needs to be removed" one-liner themselves ;)
(In reply to comment #0) > cgdb 0.6.3 has been released some time ago. The 0.6.2 ebuild works almost > without any changes, only the ${P}-parallel-make.patch needs to be removed. > I try it and doesn't work (more lower). problem is that eautoreconf is called recursively to gdmi dir and in this dir isn't any config dir. I post patch which works. >>> Emerging (1 of 1) dev-util/cgdb-0.6.3 to / * cgdb-0.6.3.tar.gz MD5 ;-) ... [ ok ] * cgdb-0.6.3.tar.gz RMD160 ;-) ... [ ok ] * cgdb-0.6.3.tar.gz SHA1 ;-) ... [ ok ] * cgdb-0.6.3.tar.gz SHA256 ;-) ... [ ok ] * cgdb-0.6.3.tar.gz size ;-) ... [ ok ] * checking ebuild checksums ;-) ... [ ok ] * checking auxfile checksums ;-) ... [ ok ] * checking miscfile checksums ;-) ... [ ok ] * checking cgdb-0.6.3.tar.gz ;-) ... [ ok ] >>> Unpacking source... >>> Unpacking cgdb-0.6.3.tar.gz to /var/tmp/portage/dev-util/cgdb-0.6.3/work * Applying cgdb-fbsd.patch ... [ ok ] * Running eautoreconf in '/var/tmp/portage/dev-util/cgdb-0.6.3/work/cgdb-0.6.3/lib/gdbmi' ... * eaclocal: 'config' does not exist * Requested autoconf 2.5 * Using autoconf (GNU Autoconf) 2.61 * Using autoheader (GNU Autoconf) 2.61 * Requested automake latest: 1.10 * Using automake (GNU automake) 1.10 * Using aclocal (GNU automake) 1.10 * Running aclocal -I config ... [ !! ] * Failed Running aclocal ! * * Include in your bugreport the contents of: * * /var/tmp/portage/dev-util/cgdb-0.6.3/temp/aclocal-5325.out !!! ERROR: dev-util/cgdb-0.6.3 failed. Call stack: ebuild.sh, line 1593: Called dyn_unpack ebuild.sh, line 731: Called src_unpack cgdb-0.6.3.ebuild, line 34: Called eautoreconf autotools.eclass, line 76: Called eautoreconf autotools.eclass, line 86: Called eaclocal autotools.eclass, line 124: Called autotools_run_tool 'aclocal' '-I' 'config' autotools.eclass, line 248: Called die !!! Failed Running aclocal !
Created attachment 105773 [details, diff] patch from 0.6.2 to 0.6.3 removed line with patch (more previous comments) and change am4 dir to absolute patch.
This is due to a modification flameeyes made to cgdb-0.6.2.ebuild three days ago [1], it _did_ work before ;-) Thanks for your fix. Works fine here, too. BTW: ka0ttic, do you intend to include cgdb-0.6.3 some day? [1]http://viewcvs.gentoo.org/viewcvs.py/gentoo-x86/dev-util/cgdb/cgdb-0.6.2.ebuild?r1=1.1&r2=1.2
This one too...
Modified 0.6.3 ebuild is now in portage. It's a bit finicky about automake, but it builds and runs fine on amd64 with the WANT_AUTO*=1.9. Please report any new problems.
The build process breaks sometimes if parallel make (-j2 or more) is being used (makefile race conditions). This is already fixed in CVS; are you interested in a patch? A workaround would of course be to ignore MAKEOPTS (ie. use make or emake -j1 instead of emake). [...] if i686-pc-linux-gnu-gcc -DHAVE_CONFIG_H -I. -I. -I.. -I../src/mi_oc_parser -march=athlon-xp -O2 -pipe -MT gdbmi_oc.o -MD -MP -MF ".deps/gdbmi_oc.Tpo" -c -o gdbmi_oc.o gdbmi_oc.c; \ then mv -f ".deps/gdbmi_oc.Tpo" ".deps/gdbmi_oc.Po"; else rm -f ".deps/gdbmi_oc.Tpo"; exit 1; fi if i686-pc-linux-gnu-gcc -DHAVE_CONFIG_H -I. -I. -I.. -I../src/mi_oc_parser -march=athlon-xp -O2 -pipe -MT gdbmi_parser.o -MD -MP -MF ".deps/gdbmi_parser.Tpo" -c -o gdbmi_parser.o gdbmi_parser.c; \ then mv -f ".deps/gdbmi_parser.Tpo" ".deps/gdbmi_parser.Po"; else rm -f ".deps/gdbmi_parser.Tpo"; exit 1; fi if i686-pc-linux-gnu-gcc -DHAVE_CONFIG_H -I. -I. -I.. -I../src/mi_oc_parser -march=athlon-xp -O2 -pipe -MT gdbmi_driver.o -MD -MP -MF ".deps/gdbmi_driver.Tpo" -c -o gdbmi_driver.o gdbmi_driver.c; \ then mv -f ".deps/gdbmi_driver.Tpo" ".deps/gdbmi_driver.Po"; else rm -f ".deps/gdbmi_driver.Tpo"; exit 1; fi i686-pc-linux-gnu-gcc -I../src/mi_oc_parser -march=athlon-xp -O2 -pipe -o gdbmi_driver -L. gdbmi_driver.o -lgdbmi /usr/lib/gcc/i686-pc-linux-gnu/4.1.1/../../../../i686-pc-linux-gnu/bin/ld: cannot find -lgdbmi collect2: ld returned 1 exit status make[5]: *** [gdbmi_driver] Error 1 make[5]: *** Waiting for unfinished jobs....
Sure, a patch would definitely be preferred over ignoring -jN.
Created attachment 113766 [details, diff] cgdb-0.6.3-makefile-race.patch This patch fixes the aforementioned Makefile race conditions. In a similar form it is in cgdb CVS, which is not publicly accessible at the moment though (the sf.net repository is not in use anymore).
Thanks; patched and updated.
