I found out today that some of my boxes running setiathome-4.18 and boinc-4* (because of some incompatibility with boinc-5* for seti) had stopped calculating for some time while some boxes with up to date boinc and the client auto downloaded by boinc (which is setiathome-5.12) keeped on working. So, I am asking for a new snapshot, because some of my boxes can't get client from boinc because they don't have precompiled clients for those arch.
I am going to try and get this done over the weekend. I have boinc and setiathome on my list, but they are not very well packaged upstream so it does take quite some time to get it all sorted (or it did last time). If anyone fancies helping me out then please feel free :)
I know I tried to make an ebuild out of it months ago, but feel it was too difficult for me at that time. Thanks for your help :)
fuuu, I just spent my morning on that and came up with nothing. This is really more difficult than what I though. Anyway, if you need some testers, I'll be happy to test anything you have.
I guess upstream deserves kicking their asses because the sources simply do not work ... there are such stupid things like _autoconfig stopping at version check of pkg-config complaining that it needs at least 0.15 while I have 0.20 ... and the worse thing is that there is no (usable) documentation! I downloaded binaries for AMD64 from http://www.lb.shuttle.de/apastron/boincDown.shtml and it seems working ... I just replaced the boinc_client binary (I emerged boinc-5.5.6 previously) and copied app_info.xml & setiathome binary btw, for the purposes of creating usable ebuilds, I would suggest using custom sources and the process described at http://lunatics.at/index.php?op=Articles;article=2 (did not check myself but to me it seems promising ...) p.s. does anybody know, what else boinc-based application can be run on AMD64? - all of the projects I participate in report unknown platform and I cannot find sources (or at least they were not available when I checked for the last time - more than year for some of them :-)
Created attachment 102001 [details] setiathome-enhanced ebuild patches will follow This setiathome-enhanced ebuild is based of the nightly tarballs which date up to Oct 31. The seti_boinc tree has been in the works still after the nightly tarballs have ceased production. At this moment I'm trying to find out a way to automate the upgrade process though cvs. I'm having difficultly downloading some files required because they do not show up existing in cvs status -v command but are present though the cvs web interface. With this ebuild, there are some patches that affect various boinc code and has been tested up to the suggested latest development build. If you want to try those out you will need to hack the ebuild to enable some of these patches. Boinc versions >5.4.11 seem to have problems downloading work units. I didn't have enough time to fully test out a work unit so consider this very experimental.
Created attachment 102002 [details, diff] setiathome-4.18-gcc41.patch for sah-enhanced ebuild
Created attachment 102003 [details, diff] sah-5.17-sysmacros-fix.patch for sah-enhanced ebuild bugfix
Created attachment 102004 [details, diff] sah-5.17-linux.patch for sah-enhanced ebuild bugfix
Created attachment 102005 [details, diff] sah-5.17-graphics-fix.patch for sah-enhanced ebuild bugfix
Created attachment 102006 [details, diff] sah-5.17-graphics-boinc-5.5.6-compatibility.patch for sah-enhanced ebuild bugfix
Created attachment 102007 [details, diff] sah-5.17-x86-asm-fix.patch for sah-enhanced ebuild bugfix
Created attachment 102008 [details, diff] sah-5.17-vector-ptrfuncts-fix.patch for sah-enhanced ebuild bugfix
Created attachment 102009 [details, diff] sah-5.17-db-makefile-exceptions-fix.patch for sah-enhanced ebuild bugfix
Created attachment 102010 [details, diff] sah-5.17-graphics-makefile-fix.patch for sah-enhanced ebuild bugfix
Created attachment 102011 [details, diff] sah-5.17-x86_86-prescott-sse3-cflags-fix-and-fpmath-to-funsafe-math-opt-fix.patch for sah-enhanced ebuild bugfix
Created attachment 102012 [details, diff] sah-5.17-boinc-5.4.11-boinc_app_mouse_button-fix.patch for sah-enhanced ebuild bugfix
Created attachment 102013 [details, diff] boinc-5.5.6-funroll-propagation-fix.patch for sah-enhanced ebuild bugfix
Created attachment 102014 [details, diff] boinc-5.5.6-server-exceptions-fix.patch boinc-5.5.6-server-exceptions-fix.patch
Created attachment 102015 [details, diff] boinc-5.4.11-gui_rpc_server-casting-fix.patch sah-enhanced ebuild bugfix
Created attachment 102016 [details, diff] boinc-5.2.14-x86_64-ptr-compat-fix.patch sah-enhanced ebuild bugfix
Created attachment 102017 [details, diff] boinc-5.2.14-gcc-4.1.patch sah-enhanced ebuild bugfix
Created attachment 102018 [details, diff] boinc-5.6.4-clientgui-wxgtk-libs-fix.patch sah-enhanced ebuild bugfix
Created attachment 102019 [details, diff] boinc-5.7.2.20061107-clientgui-header-fix.patch for sah-enhanced ebuild bugfix
Created attachment 102020 [details, diff] boinc-5.7.2.20061107-makefile-headers-fix.patch for sah-enhanced ebuild bugfix
Created attachment 102021 [details, diff] boinc-4.72.20050813-gcc4-fix.patch boinc-4.72.20050813-gcc4-fix.patch
Comment on attachment 102011 [details, diff] sah-5.17-x86_86-prescott-sse3-cflags-fix-and-fpmath-to-funsafe-math-opt-fix.patch on this one... -march=prescott outputs compiler failure and something like that your x86_64 machine is not 64 bit compatible if you enable sse3 use flag for this ebuild. mfpmath=sse , 386 is experimental so you may want to edit it out when you apply this patch.
(In reply to comment #4) > I guess upstream deserves kicking their asses because the sources simply do not > work ... there are such stupid things like _autoconfig stopping at version > check of pkg-config complaining that it needs at least 0.15 while I have 0.20 > ... and the worse thing is that there is no (usable) documentation! > > I downloaded binaries for AMD64 from > http://www.lb.shuttle.de/apastron/boincDown.shtml and it seems working ... I > just replaced the boinc_client binary (I emerged boinc-5.5.6 previously) and > copied app_info.xml & setiathome binary > > btw, for the purposes of creating usable ebuilds, I would suggest using custom > sources and the process described at > http://lunatics.at/index.php?op=Articles;article=2 (did not check myself but to > me it seems promising ...) > > p.s. does anybody know, what else boinc-based application can be run on AMD64? > - all of the projects I participate in report unknown platform and I cannot > find sources (or at least they were not available when I checked for the last > time - more than year for some of them :-) > ...As for the autoconfig problems. This affects boinc 5.4.11 also, which doesn't compile here on stock ebuild. You need to run the following command or edit the ebuild to run aclocal -I m4;autoheader;automake;autoconf ...For your unknown platform problems, you may need to edit your app_info.xml found at /var/lib/boinc/projects folders for anonymous platform support. See http://boinc.berkeley.edu/anonymous_platform.php for details.
Comment on attachment 102001 [details] setiathome-enhanced ebuild patches will follow This ebuild is intended to replace the nonfunctional setiathome 4.x packages. Rumor... is it that the 4.x no longer processes new workunits The enhanced version has vectorization enhancedments (sse/mmx = parallel processing), it able to utilize IPP (Intel Processing Primatives) for multi-core-support, vectorization (sse/mmx) support, etc. You will need to download the library from Intel if you want to use this special library. See the ebuild for more directions.
Created attachment 102027 [details, diff] app_info.xml to be processed by the ebuild app_info.xml to be processed by the ebuild
Comment on attachment 102001 [details] setiathome-enhanced ebuild patches will follow setiathome-enhanced-5.17.20061115 soon
Comment on attachment 102004 [details, diff] sah-5.17-linux.patch already upstream fixed
Comment on attachment 102007 [details, diff] sah-5.17-x86-asm-fix.patch already upstream fixed
Comment on attachment 102002 [details, diff] setiathome-4.18-gcc41.patch already upstream fixed
Comment on attachment 102003 [details, diff] sah-5.17-sysmacros-fix.patch already upstream fixed
Created attachment 102044 [details] setiathome-enhanced-5.17.20061115.ebuild I was checking the nightly tarballs at seti's website and they updated today. So far the compile went well.
Ehm, doesn't build here: make[1]: Entering directory `/var/tmp/portage/sci-misc/setiathome-enhanced-5.17.20061115/work/seti_boinc' Making all in client make[2]: Entering directory `/var/tmp/portage/sci-misc/setiathome-enhanced-5.17.20061115/work/seti_boinc/client' if i686-pc-linux-gnu-g++ --include ../config.h -DHAVE_CONFIG_H -I. -I. -I.. -DHAVE_CONFIG_H -DTEXT_UI -DNDEBUG -DCLIENT -I../db -I/var/tmp/portage/sci-misc/setiathome-enhanced-5.17.20061115/work/boinc/api -I/var/tmp/portage/sci-misc/setiathome-enhanced-5.17.20061115/work/boinc/lib -pthread -O2 -Wall -MT seti_boinc-main.o -MD -MP -MF ".deps/seti_boinc-main.Tpo" -c -o seti_boinc-main.o main.cpp; then mv -f ".deps/seti_boinc-main.Tpo" ".deps/seti_boinc-main.Po"; else rm -f ".deps/seti_boinc-main.Tpo"; exit 1; fi if i686-pc-linux-gnu-g++ -DHAVE_CONFIG_H -I. -I. -I.. -m3dnow -mmmx -msse -mfpmath=sse,387 -O3 -march=athlon-xp -msse -mfpmath=sse -ffast-math -ftracer -pipe -fomit-frame-pointer -I/usr/include -I/usr/include/openssl -fPIC -DPIC -DHAVE_CONFIG_H -m3dnow -mmmx -msse -mfpmath=sse,387 -O3 -march=athlon-xp -msse -mfpmath=sse -ffast-math -ftracer -pipe -fomit-frame-pointer -I/usr/include -I/usr/include/openssl -DHAVE_CONFIG_H -DTEXT_UI -DNDEBUG -DCLIENT -I../db -I/var/tmp/portage/sci-misc/setiathome-enhanced-5.17.20061115/work/boinc/api -I/var/tmp/portage/sci-misc/setiathome-enhanced-5.17.20061115/work/boinc/lib -pthread -I/usr/ -pthread -pthread -pthread -O3 -march=athlon-xp -msse -mfpmath=sse -ffast-math -ftracer -pipe -fomit-frame-pointer -fvisibility-inlines-hidden -ffriend-injection -Wno-deprecated -I/usr/include -I/usr/include/openssl -MT seti_boinc-analyzeFuncs_vector.o -MD -MP -MF ".deps/seti_boinc-analyzeFuncs_vector.Tpo" -c -o seti_boinc-analyzeFuncs_vector.o `test -f 'vector/analyzeFuncs_vector.cpp' || echo './'`vector/analyzeFuncs_vector.cpp; \ then mv -f ".deps/seti_boinc-analyzeFuncs_vector.Tpo" ".deps/seti_boinc-analyzeFuncs_vector.Po"; else rm -f ".deps/seti_boinc-analyzeFuncs_vector.Tpo"; exit 1; fi if i686-pc-linux-gnu-g++ -DHAVE_CONFIG_H -I. -I. -I.. -m3dnow -mmmx -msse -mfpmath=sse,387 -O3 -march=athlon-xp -msse -mfpmath=sse -ffast-math -ftracer -pipe -fomit-frame-pointer -I/usr/include -I/usr/include/openssl -fPIC -DPIC -DHAVE_CONFIG_H \ -m3dnow -mmmx -msse -mfpmath=sse,387 -O3 -march=athlon-xp -msse -mfpmath=sse -ffast-math -ftracer -pipe -fomit-frame-pointer -I/usr/include -I/usr/include/openssl -DHAVE_CONFIG_H -DTEXT_UI -DNDEBUG -DCLIENT -I../db -I/var/tmp/portage/sci-misc/setiathome-enhanced-5.17.20061115/work/boinc/api -I/var/tmp/portage/sci-misc/setiathome-enhanced-5.17.20061115/work/boinc/lib -pthread -I/usr/ -pthread -pthread -pthread -O3 -march=athlon-xp -msse -mfpmath=sse -ffast-math -ftracer -pipe -fomit-frame-pointer -fvisibility-inlines-hidden -ffriend-injection -Wno-deprecated -I/usr/include -I/usr/include/openssl -DUSE_SSE \ -MT seti_boinc-analyzeFuncs_sse.o -MD -MP -MF ".deps/seti_boinc-analyzeFuncs_sse.Tpo" -msse \ -c -o seti_boinc-analyzeFuncs_sse.o ; then mv -f ".deps/seti_boinc-analyzeFuncs_sse.Tpo" ".deps/seti_boinc-analyzeFuncs_sse.Po"; \ else rm -f ".deps/seti_boinc-analyzeFuncs_sse.Tpo"; exit 1; fi i686-pc-linux-gnu-g++: no input files make[2]: *** [seti_boinc-analyzeFuncs_sse.o] Error 1 make[2]: *** Waiting for unfinished jobs.... make[2]: Leaving directory `/var/tmp/portage/sci-misc/setiathome-enhanced-5.17.20061115/work/seti_boinc/client' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/var/tmp/portage/sci-misc/setiathome-enhanced-5.17.20061115/work/seti_boinc' make: *** [all] Error 2 !!! ERROR: sci-misc/setiathome-enhanced-5.17.20061115 failed. Call stack: ebuild.sh, line 1568: Called dyn_compile ebuild.sh, line 937: Called src_compile setiathome-enhanced-5.17.20061115.ebuild, line 149: Called die !!! Error: failed to make setiathome !!! If you need support, post the topmost build error, and the call stack if relevant. emerge --info is here: http://kripton.kripserver.net/n0th1ng/einfo
(In reply to comment #27) > ...For your unknown platform problems, you may need to edit your app_info.xml > found at /var/lib/boinc/projects folders for anonymous platform support. See > http://boinc.berkeley.edu/anonymous_platform.php for details. that page does not tell me where to get the binaries for my platform (except for SAH) and without the binaries the app_info.xml is just useless waste of the disk space ...
tried to emerge, looks ok, but I was surprised by a few things: sse flags are filtered out while my processor supports it (except for sse3 for which it is the other way round) configure complains: checking for mysql_config... no MYSQL not found configure: WARNING: mysql not found. ============================================================================ WARNING: trying to build the seti_boinc server but mysql was not found. If you don't want to build the server you should use --disable-server. I am continuing now as if --disable-server had been specified. ============================================================================ ... well, who runs own server at home? guess --disable-server should be off by default (or follow the "server" use flag as with the boinc from portage) then the comments: * After attaching the client to Boinc, please perform * * /etc/init.d/restart * /etc/init.d/restart I do not understand the need to restart more than once, and the init script is "boinc" and not "restart" ;-) however, it is a good start, keep the good work ;-)
Comment on attachment 102008 [details, diff] sah-5.17-vector-ptrfuncts-fix.patch already solved upstream
Comment on attachment 102006 [details, diff] sah-5.17-graphics-boinc-5.5.6-compatibility.patch not needed patch... the ebuild doesn't make use of Makefile.graphics -> Makefile.in
Comment on attachment 102010 [details, diff] sah-5.17-graphics-makefile-fix.patch not needed patch... the ebuild doesn't make use of Makefile.in.graphics -> Makefile.in
Comment on attachment 102021 [details, diff] boinc-4.72.20050813-gcc4-fix.patch not needed to build for boinc 5.4.11
Comment on attachment 102016 [details, diff] boinc-5.2.14-x86_64-ptr-compat-fix.patch not needed for boinc 5.4.11
Comment on attachment 102017 [details, diff] boinc-5.2.14-gcc-4.1.patch not needed to build boinc 5.4.11
Comment on attachment 102019 [details, diff] boinc-5.7.2.20061107-clientgui-header-fix.patch boinc 5.7.2.20061107 is too experimental and doesn't compile so remove this patch
Comment on attachment 102020 [details, diff] boinc-5.7.2.20061107-makefile-headers-fix.patch boinc 5.7.2.20061107 patch removal...unnecessary
Created attachment 102111 [details] setiathome-enhanced-5.17.20061115.ebuild update Updated ebuild with removal of patches for boinc <5.4.11 and setiathome <5.17.20061115.
Comment on attachment 102111 [details] setiathome-enhanced-5.17.20061115.ebuild update wrong file sorry...
Created attachment 102115 [details] setiathome-enhanced-5.17.20061115.ebuild
(In reply to comment #36) > Ehm, doesn't build here: > make[1]: Entering directory > `/var/tmp/portage/sci-misc/setiathome-enhanced-5.17.20061115/work/seti_boinc' > Making all in client > make[2]: Entering directory > `/var/tmp/portage/sci-misc/setiathome-enhanced-5.17.20061115/work/seti_boinc/client' > if i686-pc-linux-gnu-g++ --include ../config.h -DHAVE_CONFIG_H -I. -I. -I.. > -DHAVE_CONFIG_H -DTEXT_UI -DNDEBUG -DCLIENT -I../db > -I/var/tmp/portage/sci-misc/setiathome-enhanced-5.17.20061115/work/boinc/api > -I/var/tmp/portage/sci-misc/setiathome-enhanced-5.17.20061115/work/boinc/lib > -pthread -O2 -Wall -MT seti_boinc-main.o -MD -MP -MF > ".deps/seti_boinc-main.Tpo" -c -o seti_boinc-main.o main.cpp; then mv -f > ".deps/seti_boinc-main.Tpo" ".deps/seti_boinc-main.Po"; else rm -f > ".deps/seti_boinc-main.Tpo"; exit 1; fi > if i686-pc-linux-gnu-g++ -DHAVE_CONFIG_H -I. -I. -I.. -m3dnow -mmmx -msse > -mfpmath=sse,387 -O3 -march=athlon-xp -msse -mfpmath=sse -ffast-math -ftracer > -pipe -fomit-frame-pointer -I/usr/include -I/usr/include/openssl -fPIC -DPIC > -DHAVE_CONFIG_H -m3dnow -mmmx -msse -mfpmath=sse,387 -O3 -march=athlon-xp > -msse -mfpmath=sse -ffast-math -ftracer -pipe -fomit-frame-pointer > -I/usr/include -I/usr/include/openssl -DHAVE_CONFIG_H -DTEXT_UI -DNDEBUG > -DCLIENT -I../db > -I/var/tmp/portage/sci-misc/setiathome-enhanced-5.17.20061115/work/boinc/api > -I/var/tmp/portage/sci-misc/setiathome-enhanced-5.17.20061115/work/boinc/lib > -pthread -I/usr/ -pthread -pthread -pthread -O3 -march=athlon-xp -msse > -mfpmath=sse -ffast-math -ftracer -pipe -fomit-frame-pointer > -fvisibility-inlines-hidden -ffriend-injection -Wno-deprecated -I/usr/include > -I/usr/include/openssl -MT seti_boinc-analyzeFuncs_vector.o -MD -MP -MF > ".deps/seti_boinc-analyzeFuncs_vector.Tpo" -c -o > seti_boinc-analyzeFuncs_vector.o `test -f 'vector/analyzeFuncs_vector.cpp' || > echo './'`vector/analyzeFuncs_vector.cpp; \ > then mv -f ".deps/seti_boinc-analyzeFuncs_vector.Tpo" > ".deps/seti_boinc-analyzeFuncs_vector.Po"; else rm -f > ".deps/seti_boinc-analyzeFuncs_vector.Tpo"; exit 1; fi > if i686-pc-linux-gnu-g++ -DHAVE_CONFIG_H -I. -I. -I.. -m3dnow -mmmx -msse > -mfpmath=sse,387 -O3 -march=athlon-xp -msse -mfpmath=sse -ffast-math -ftracer > -pipe -fomit-frame-pointer -I/usr/include -I/usr/include/openssl -fPIC -DPIC > -DHAVE_CONFIG_H \ > -m3dnow -mmmx -msse -mfpmath=sse,387 -O3 -march=athlon-xp -msse > -mfpmath=sse -ffast-math -ftracer -pipe -fomit-frame-pointer -I/usr/include > -I/usr/include/openssl -DHAVE_CONFIG_H -DTEXT_UI -DNDEBUG -DCLIENT -I../db > -I/var/tmp/portage/sci-misc/setiathome-enhanced-5.17.20061115/work/boinc/api > -I/var/tmp/portage/sci-misc/setiathome-enhanced-5.17.20061115/work/boinc/lib > -pthread -I/usr/ -pthread -pthread -pthread -O3 -march=athlon-xp -msse > -mfpmath=sse -ffast-math -ftracer -pipe -fomit-frame-pointer > -fvisibility-inlines-hidden -ffriend-injection -Wno-deprecated -I/usr/include > -I/usr/include/openssl -DUSE_SSE \ > -MT seti_boinc-analyzeFuncs_sse.o -MD -MP -MF > ".deps/seti_boinc-analyzeFuncs_sse.Tpo" -msse \ > -c -o seti_boinc-analyzeFuncs_sse.o ; then mv -f > ".deps/seti_boinc-analyzeFuncs_sse.Tpo" ".deps/seti_boinc-analyzeFuncs_sse.Po"; > \ > else rm -f ".deps/seti_boinc-analyzeFuncs_sse.Tpo"; exit 1; fi > i686-pc-linux-gnu-g++: no input files > make[2]: *** [seti_boinc-analyzeFuncs_sse.o] Error 1 > make[2]: *** Waiting for unfinished jobs.... > make[2]: Leaving directory > `/var/tmp/portage/sci-misc/setiathome-enhanced-5.17.20061115/work/seti_boinc/client' > make[1]: *** [all-recursive] Error 1 > make[1]: Leaving directory > `/var/tmp/portage/sci-misc/setiathome-enhanced-5.17.20061115/work/seti_boinc' > make: *** [all] Error 2 > > !!! ERROR: sci-misc/setiathome-enhanced-5.17.20061115 failed. > Call stack: > ebuild.sh, line 1568: Called dyn_compile > ebuild.sh, line 937: Called src_compile > setiathome-enhanced-5.17.20061115.ebuild, line 149: Called die > > !!! Error: failed to make setiathome > !!! If you need support, post the topmost build error, and the call stack if > relevant. > > emerge --info is here: > http://kripton.kripserver.net/n0th1ng/einfo > option 1 I updated the ebuild again. You might wanna try this one since upstream applied a solution to this one. option 2 If the above fails, I need a more detailed view of the output Can you do the following for me please. Go to your ebuild ebuild setiathome-enhanced-5.17.20061115.ebuild clean ebuild setiathome-enhanced-5.17.20061115.ebuild unpack *Make sure you comment out your cflags cxxflags temporarly (by adding a # infront of them) you have -fvisibility-inlines-hidden cd /var/tmp/portage/seti*1115*/work/seti_boinc aclocal -I m4;autoheader;automake;autoconfig BOINCDIR=../boinc ./configure --disable-graphics make clean cd client make Send me any new info relating to compiling analyzeFuncs_sse It seems that there might be a problem related to codeblocks for x86 environment and analyzeFuncs_sse.c and analyzeFuncs* option3 You may also want to disable the sse use flag in your package.use for this package and then remerge it. nano /etc/portage/package.use #use your favorite text editor and edit package.use add or modify the line sci-astronomy/setiathome-enhanced -sse -3dnow -mmx -sse2 -sse3 -fast-math -altivec option4 Also try out the bitness compile option which selectively compiles for 32/64bit environment see ./configure --help for details
(In reply to comment #51) > (In reply to comment #36) > > Ehm, doesn't build here: > > [...] > > option 1 > I updated the ebuild again. You might wanna try this one since upstream applied > a solution to this one. > > option 2 > If the above fails, I need a more detailed view of the output > Can you do the following for me please. > Go to your ebuild > ebuild setiathome-enhanced-5.17.20061115.ebuild clean > ebuild setiathome-enhanced-5.17.20061115.ebuild unpack > *Make sure you comment out your cflags cxxflags temporarly (by adding a # > infront of them) you have -fvisibility-inlines-hidden > cd /var/tmp/portage/seti*1115*/work/seti_boinc > aclocal -I m4;autoheader;automake;autoconfig > BOINCDIR=../boinc ./configure --disable-graphics > make clean > cd client > make > > Send me any new info relating to compiling analyzeFuncs_sse > It seems that there might be a problem related to codeblocks for x86 > environment and analyzeFuncs_sse.c and analyzeFuncs* > > option3 > You may also want to disable the sse use flag in your package.use for this > package and then remerge it. > > nano /etc/portage/package.use #use your favorite text editor and edit > package.use > add or modify the line > sci-astronomy/setiathome-enhanced -sse -3dnow -mmx -sse2 -sse3 -fast-math > -altivec > > option4 > Also try out the bitness compile option which selectively compiles for 32/64bit > environment see ./configure --help for details > Tried option 1 to 3 and they all don't work, here is the output of my shell (I know it's quite huge, but might be something in there to help you): http://kripton.kripserver.net/n0th1ng/seti-fail
did 2 and got : if g++ --include ../config.h -DHAVE_CONFIG_H -I. -I. -I.. -DHAVE_CONFIG_H -DTEXT_UI -DNDEBUG -DCLIENT -I../db -I/var/tmp/portage/sci-astronomy/setiathome-enhanced-5.17.20061115/work/boinc/api -I/var/tmp/portage/sci-astronomy/setiathome-enhanced-5.17.20061115/work/boinc/lib -pthread -O2 -Wall -MT seti_boinc-main.o -MD -MP -MF ".deps/seti_boinc-main.Tpo" -c -o seti_boinc-main.o main.cpp; then mv -f ".deps/seti_boinc-main.Tpo" ".deps/seti_boinc-main.Po"; else rm -f ".deps/seti_boinc-main.Tpo"; exit 1; fi if g++ -DHAVE_CONFIG_H -I. -I. -I.. -g -O2 -I/usr/include -I/usr/include/openssl -fPIC -DPIC -DHAVE_CONFIG_H -g -O2 -I/usr/include -I/usr/include/openssl -DHAVE_CONFIG_H -DTEXT_UI -DNDEBUG -DCLIENT -I../db -I/var/tmp/portage/sci-astronomy/setiathome-enhanced-5.17.20061115/work/boinc/api -I/var/tmp/portage/sci-astronomy/setiathome-enhanced-5.17.20061115/work/boinc/lib -pthread -I/usr/include -I/usr/include/openssl -MT seti_boinc-analyzeFuncs_vector.o -MD -MP -MF ".deps/seti_boinc-analyzeFuncs_vector.Tpo" -c -o seti_boinc-analyzeFuncs_vector.o `test -f 'vector/analyzeFuncs_vector.cpp' || echo './'`vector/analyzeFuncs_vector.cpp; \ then mv -f ".deps/seti_boinc-analyzeFuncs_vector.Tpo" ".deps/seti_boinc-analyzeFuncs_vector.Po"; else rm -f ".deps/seti_boinc-analyzeFuncs_vector.Tpo"; exit 1; fi if g++ -DHAVE_CONFIG_H -I. -I. -I.. -g -O2 -I/usr/include -I/usr/include/openssl -fPIC -DPIC -DHAVE_CONFIG_H \ -g -O2 -I/usr/include -I/usr/include/openssl -DHAVE_CONFIG_H -DTEXT_UI -DNDEBUG -DCLIENT -I../db -I/var/tmp/portage/sci-astronomy/setiathome-enhanced-5.17.20061115/work/boinc/api -I/var/tmp/portage/sci-astronomy/setiathome-enhanced-5.17.20061115/work/boinc/lib -pthread -I/usr/include -I/usr/include/openssl -DUSE_SSE \ -MT seti_boinc-analyzeFuncs_sse.o -MD -MP -MF ".deps/seti_boinc-analyzeFuncs_sse.Tpo" -msse \ -c -o seti_boinc-analyzeFuncs_sse.o ; then mv -f ".deps/seti_boinc-analyzeFuncs_sse.Tpo" ".deps/seti_boinc-analyzeFuncs_sse.Po"; \ else rm -f ".deps/seti_boinc-analyzeFuncs_sse.Tpo"; exit 1; fi g++: no input files make: *** [seti_boinc-analyzeFuncs_sse.o] Error 1 which is basically the same problem that jannis encountered. Portage 2.1.2_rc1-r5 (default-linux/x86/2006.1/desktop, gcc-4.1.1, glibc-2.5-r0, 2.6.17-suspend2-r4 i686) ================================================================= System uname: 2.6.17-suspend2-r4 i686 Pentium III (Katmai) Gentoo Base System version 1.12.6 CFLAGS="-O2 -march=pentium3 -fomit-frame-pointer -ftree-vectorize -pipe" CXXFLAGS="-O2 -march=pentium3 -fomit-frame-pointer -ftree-vectorize -pipe -fvisibility-inlines-hidden" LDFLAGS="-Wl,--as-needed" su # emerge -1av setiathome-enhanced [ebuild N ] sci-astronomy/setiathome-enhanced-5.17.20061115 USE="X mmx opengl sse ssl -3dnow (-altivec) -fast-math -intrinsics -ipp -sse2 -sse3" 0 kB [6]
Comment on attachment 102115 [details] setiathome-enhanced-5.17.20061115.ebuild superceeded for sah-5.17-makefile-i686-target-fix.patch inclusion
Created attachment 102244 [details, diff] sah-5.17-makefile-i686-target-fix.patch This patch fixes the Makefile to allow for seti_boinc-analyzeFuncs_sse.o to be built on i386 to i686 machines. Fixes the problem of the target trying to be built without any dependency.
Created attachment 102245 [details] setiathome-enhanced-5.17.20061115.ebuild updated ebuild with inclusion of sah-5.17-makefile-i686-target-fix.patch
Created attachment 102264 [details, diff] setiathome-enhanced/files/sah-5.17-makefile-i686-target-fix.patch new version of proposed patch. The problem here was that custom rules in Makefile.am were not complete and removing them was just making ld fail later because some symbols that are defined in seti_boinc-analyzeFuncs_x86_64 are defined only when compiled wit -msse2
Created attachment 102356 [details, diff] sah-5.17-makefile-i686-target-fix.patch Better patch (since preceding one didn't apply well)
Created attachment 102432 [details, diff] sah-5.17-makefile-i686-target-fix.patch hopefully, latest try. Fixes the little typo that makes this patch useful again. Things I've noticed is that my i586 box can't run setiathome-enhanced for now. Portage 2.1.1-r1 (default-linux/x86/2006.1, gcc-3.4.6, glibc-2.4-r4, 2.6.17-gentoo-r8 i586) ================================================================= System uname: 2.6.17-gentoo-r8 i586 Pentium MMX CBUILD="i586-pc-linux-gnu" CFLAGS="-march=pentium-mmx -O2 -funroll-loops -fomit-frame-pointer" It says : 2006-11-20 12:33:46 [SETI@home] Unrecoverable error for result 16jn03ab.9360.7712.403412.3.174_2 (process got signal 4) which after some googling could mean that generated bin is not actually i586 compatible. I'll make more tests on PI mmx and PII.
Apperently, the source-file used in this ebuild is no longer available. I tested it with the oldest nightly (that is 26. Jan 07) and was not able to compile the thing. Flags used: ipp sse3 ssl It quits with: /var/tmp/portage/sci-astronomy/setiathome-enhanced-5.17.20061115/work/boinc/api/boinc_gl.h:40:20: Error: config.h: No such file or directory If I'll comment that out, the same error occurs at seti_boin/db/scheme_master.cpp:2:20 but also this: x86_64-pc-linux-gnu-gcc -fPIC -shared -o sahgui.so -nodefaultlibs sahgui_so-sah_gfx_base.o sahgui_so-sah_gfx.o sahgui_so-version.o -pthread -L/var/tmp/portage/sci-astronomy/setiathome-enhanced-5.17.20061115/work/boinc/api -L/var/tmp/portage/sci-astronomy/setiathome-enhanced-5.17.20061115/work/boinc/lib -lboinc_graphics_impl -lboinc -Bdynamic -lGL -Bdynamic -lm -Bdynamic -lGLU -Bdynamic -lGL -Bdynamic -lm /usr/lib64/libglut.a -Bdynamic -lGLU -Bdynamic -lGL -Bdynamic -lm /usr/lib64/libjpeg.a /usr/lib/gcc/x86_64-pc-linux-gnu/4.1.2/../../../../x86_64-pc-linux-gnu/bin/ld: /usr/lib64/libglut.a(libglut_la-freeglut_callbacks.o): relocation R_X86_64_32 against `a local symbol' can not be used when making a shared object; recompile with -fPIC /usr/lib64/libglut.a: could not read symbols: Bad value emerge --info attached
Created attachment 111293 [details] emerge --info by georg.wacker
sci-astronomty/setiathome is masked for removal in 30 days. Please comment on bug #257233 if you see any reason to keep it around. Thanks.
I commented but was ignored anyway...
Closing. not in the tree anymore and boinc is. PLease reopen if you disagree.