Here's the updated ebuild for the new GHC 5.04. There is one noteworthy change (the only one actually, except for the version number): I have removed the ability to build the documentation, because it now depends on another external program and I think it would be pointless to require that, just to build a doc tree you can get in the exact same form from the GHC website. This means that only the opengl USE var remains recognized. Attached you will find the ebuild, along with the new digest file and patches. Greetings, Sven Moritz
Created attachment 2421 [details] GHC 5.04 ebuild script
Created attachment 2422 [details] distfile digests
Created attachment 2423 [details, diff] Patch to the GHC 5.04 build system This patch is the 5.04 equivalent of ghc-5.02.3-gentoo.patch.
Hi Sven. Thanks for the submission! And sorry for taking this time to get to it (we are doing a major portage clean-up and new submissions/non-crytical stuff is on lower priority for now). 1. I am checking ebuild out and I have one question: why in all ./configure invocations "--host=${CHOST}" was replaced with "--build=${CHOST}"? Or was it this way in the version you started from? 2. (just a note). The patches are getting quite big, the last one >800k, quite a piece! (thanks for taking your time on this). I bzipped the patches and put bzcat where appropriate. George
Hi Sven. Thanks for clarifications. I am testing the ebuild, however I have trouble compiling it. this is the tail of what I get with both --build=${CHOST} and --host=...: =============part1=================================== ===fptools== Finished making `boot' in cbits ... PWD = /var/tmp/portage/ghc-5.04/work/stage1-build/ghc/lib/std ------------------------------------------------------------------------ ------------------------------------------------------------------------ ===fptools== Finished making `boot' in std ... PWD = /var/tmp/portage/ghc-5.04/work/stage1-build/ghc/lib ------------------------------------------------------------------------ ------------------------------------------------------------------------ ==fptools== make boot - --no-print-directory -r --jobserver-fds=3,4 - --jobserver-fds=3,4 -j; in /var/tmp/portage/ghc-5.04/work/stage1-build/ghc/compiler ------------------------------------------------------------------------ ../../glafp-utils/mkdependC/mkdependC -f .depend -D__GLASGOW_HASKELL__=408 -- -Iparser -I. -I../includes -O -- parser/hschooks.c ------------------------------------------------------------------------ ===fptools== Finished making `boot' in utils driver includes rts docs lib compiler ... PWD = /var/tmp/portage/ghc-5.04/work/stage1-build/ghc ------------------------------------------------------------------------ ------------------------------------------------------------------------ make: *** Waiting for unfinished jobs.... ===fptools== Finished making `boot' in glafp-utils hslibs ghc ... make: *** Waiting for unfinished jobs.... PWD = /var/tmp/portage/ghc-5.04/work/stage1-build make: *** Waiting for unfinished jobs.... ------------------------------------------------------------------------ make: *** Waiting for unfinished jobs.... ============end of part 1===================================== also I notice this scrolling by during the build (but it does not stop at that point): ===============part2============================= ==fptools== make all - --no-print-directory -r --jobserver-fds=3,4 - --jobserver-fds=3,4 - --jobserver-fds=3,4 - --jobserver-fds=3,4 -j; in /var/tmp/portage/ghc-5.04/work/stage1-build/hslibs/lang/cbits ------------------------------------------------------------------------ PWD = /var/tmp/portage/ghc-5.04/work/stage1-build/ghc/utils ------------------------------------------------------------------------ ------------------------------------------------------------------------ ==fptools== make boot - --no-print-directory -r --jobserver-fds=3,4 - --jobserver-fds=3,4 - --jobserver-fds=3,4 - --jobserver-fds=3,4 -j; in /var/tmp/portage/ghc-5.04/work/stage1-build/ghc/utils/hp2ps ------------------------------------------------------------------------ ../../../ghc/driver/ghc-inplace -I../../../ghc/includes -I../../../ghc/lib/std/cbits -static -c PackedString.c -o PackedString.o make[3]: ../../../ghc/driver/ghc-inplace: Command not found make[3]: *** [PackedString.o] Error 127 make[2]: *** [all] Error 1 make[1]: *** [all] Error 1 make: *** [all] Error 1 make: *** Waiting for unfinished jobs.... ../../../glafp-utils/mkdependC/mkdependC -f .depend -- -O -- AreaBelow.c AuxFile.c Axes.c Curves.c Deviation.c Dimensions.c Error.c HpFile.c Key.c Main.c Marks.c PsFile.c Reorder.c Scale.c Shade.c TopTwenty.c TraceElement.c Utilities.c ------------------------------------------------------------------------ ==fptools== make boot - --no-print-directory -r --jobserver-fds=3,4 - --jobserver-fds=3,4 - --jobserver-fds=3,4 - --jobserver-fds=3,4 -j; in /var/tmp/portage/ghc-5.04/work/stage1-build/ghc/utils/hscpp and it continue on... All this happens within 1st minute of build. Could you please take a look at what's going on? Can it be related to CFLAGS? I have them set quite strongly (partially to catch some of CFLAGS related bugs). What are your settings? George
Hi Sven. Thanks for finding the problem! (since you did not reply to the bug but rather directly via email, I am leaving the note here that the problem was due to parallel build, needed to change all emake's to make's). Updated ebuild and committed. Please test, I'll unmask this version upon successful report. George
Thanks for testing Sven! unmasked, closing this bug. George
