While compiling app-text/sigil I've noticed the following: /usr/lib64/distcc/bin/x86_64-pc-linux-gnu-g++ -DXERCES_STATIC_LIBRARY -march=core2 -O2 -pipe -floop-interchange -floop-strip-mine -floop-block -O3 -DNDEBUG ... -Wall (full build.log attached) ...which, I think, enforces -O3 optimization (which I consider being bad) and -Wall warning level (which is just a bit untidy). Maybe there's something more, but I'm not that proficient in C++ compiling to notice. So, -O3 *should* be stripped and -Wall *may* be stripped, while we're at it anyways :)
Oh blast, I can't attach full build.log, it's too huge. Is it really needed?
-Wall is not a problem...About -O3 we need to know if drop it may break something. In that case use custom-cflags && sed "s:-O3::" $makefile
my only question: does 0.6.2 the same? I want to remove anything older than 0.6.0 in a few days...
already removed. feel free to reopen, if that happens with 0.6.2.
> my only question: does 0.6.2 the same? Excuse me, but did you compile it prior to adding to the tree? That's a strange question for maintainer. Yes, there's nothing new there: cd /var/tmp/portage/app-text/sigil-0.6.2/work/sigil-0.6.2_build/src/pcre && /usr/lib64/distcc/bin/x86_64-pc-linux-gnu-gcc -DHAVE_CONFIG_H -march=core2 -O2 -pipe -floop-interchange -floop-strip-mine -floop-block -O3 -DNDEBUG -I/var/tmp/portage/app-text/sigil-0.6.2/work/sigil-0.6.2_build/src/pcre -I/var/tmp/portage/app-text/sigil-0.6.2/work/src/pcre -o CMakeFiles/pcre.dir/pcre16_exec.c.o -c /var/tmp/portage/app-text/sigil-0.6.2/work/src/pcre/pcre16_exec.c Reopening.
of course! I compile all my ebuilds more than once before I commit them. But your bugreport came AFTER I commited all these ebuilds. So before I recompile everything just for you, you can just answer my questions, if you know the answer.
I would like to add sigil-0.7.1 to portage, but it seems to need Qt5, which is not in portage yet.
0.6.2 can be compiled. And yes, I see a -O3 there. BUT: ~# fgrep "O3" . -r ./src/zlib/crc32.c: * factor of two increase in speed on a Power PC G4 (PPC7455) using gcc -O3. ./src/zlib/ChangeLog:- make zlib warning-free with "gcc -O3 -Wall -Wwrite-strings -Wpointer-arith ./src/zlib/ChangeLog:- use STDC instead of __GO32__ to avoid redeclaring exit, calloc, etc... Übereinstimmungen in Binärdatei ./src/Sigil/Resource_Files/main/format-case-capitalize_22px.png. Übereinstimmungen in Binärdatei ./src/Sigil/Resource_Files/icon/epub.icns. Übereinstimmungen in Binärdatei ./src/Sigil/Resource_Files/icon/app_icon_512.png. Übereinstimmungen in Binärdatei ./src/Sigil/Resource_Files/icon/Sigil.icns. Übereinstimmungen in Binärdatei ./installer/msvc_crt/x86/msvcr100.dll. Übereinstimmungen in Binärdatei ./installer/msvc_crt/x86/msvcp100.dll. Übereinstimmungen in Binärdatei ./installer/msvc_crt/x64/msvcr100.dll. This -O3 must come from elsewhere... it's at least not in the sources.
(In reply to comment #6) > But your bugreport came AFTER I commited all these ebuilds I reported a bug on Nov 1, and 0.6.2 was added on Jan 25. Anyways, > This -O3 must come from elsewhere... it's at least not in the sources Kind of (`ag` is sys-apps/the_silver_searcher): % ag O3 src/zlib/crc32.c 9: * factor of two increase in speed on a Power PC G4 (PPC7455) using gcc -O3. src/zlib/ChangeLog 1242:- make zlib warning-free with "gcc -O3 -Wall -Wwrite-strings -Wpointer-arith 1267:- use STDC instead of __GO32__ to avoid redeclaring exit, calloc, etc... % cmake . ... % ag O3 src/zlib/crc32.c 9: * factor of two increase in speed on a Power PC G4 (PPC7455) using gcc -O3. src/zlib/ChangeLog 1242:- make zlib warning-free with "gcc -O3 -Wall -Wwrite-strings -Wpointer-arith 1267:- use STDC instead of __GO32__ to avoid redeclaring exit, calloc, etc... CMakeCache.txt 105:CMAKE_CXX_FLAGS_RELEASE:STRING=-O3 -DNDEBUG 125:CMAKE_C_FLAGS_RELEASE:STRING=-O3 -DNDEBUG
(In reply to Agostino Sarubbo from comment #2) > -Wall is not a problem...About -O3 we need to know if drop it may break > something. > > In that case use custom-cflags && sed "s:-O3::" $makefile We don't need to know it beforehand. Just expose it.
whoa there ... never knew i had this bug open. anyway, in progress now. finally.
https://github.com/gentoo/gentoo/pull/130
took a while, but it's done.
