changelog
- Mon, 04 Apr 2011 20:34:12 +0200
- by Michael Schloh von Bennewitz <michael@schloh.com> [Mon, 04 Apr 2011 20:34:12 +0200] rev 339
- Clean duplicate fortran binary leaving only canonical gfortran(1).
- Sun, 03 Apr 2011 15:43:53 +0200
- by Michael Schloh von Bennewitz <michael@schloh.com> [Sun, 03 Apr 2011 15:43:53 +0200] rev 338
- Remove duplicate java and unused include paths, and correct jar path.
- Sun, 03 Apr 2011 13:34:55 +0200
- by Michael Schloh von Bennewitz <michael@schloh.com> [Sun, 03 Apr 2011 13:34:55 +0200] rev 337
- Correct and improve several packaging aspects including...
Correct datadir path for python modules, correct jar(1) path for
building libgcj classes, strip libexecdir path of version numbers,
improve name of oblbld build path, clean whitespace from as(1) and
ld(1) GNU detection, remove seemingly discarded '--with-local-prefix'
configure argument, and correct hardcoded lto plugin libtool archive
dependency information.
Most importantly, correct IA32 architecture detection logic in
config.gcc to correctly emit SSE2 instructions conditionally, leading
to the removal of all '-march' bootstrap options and replacement with
unconditional (for IA32/AMD64) '-mtune=native' options. Comments and
buildtime warnings are corrected appropriately. In theory these changes
cause a more portable, orthoganal, and optimal bootstrap to be built.
- Wed, 30 Mar 2011 19:37:58 +0200
- by Michael Schloh von Bennewitz <michael@schloh.com> [Wed, 30 Mar 2011 19:37:58 +0200] rev 336
- Renumber patch hunks to avoid patch(1) warnings, this time successfully.
- Wed, 30 Mar 2011 19:35:16 +0200
- by Michael Schloh von Bennewitz <michael@schloh.com> [Wed, 30 Mar 2011 19:35:16 +0200] rev 335
- Finally correct packaging logic enough to allow for successful building:
Force Java dependency to C++ and POSIX threads to satisfy buildtime
requirements of GCJ, add conditional java dependency to infozip as
required by libjava/configure, remove outdated fortran conditions,
correct buildconf causing jcf-dump target to fail, add and improve
warnings regarding profile feedback, build duration, and parallel
make, correct make environment when building C++ dependent languages,
and introduce new helpful comments.
- Tue, 29 Mar 2011 20:04:34 +0200
- by Michael Schloh von Bennewitz <michael@schloh.com> [Tue, 29 Mar 2011 20:04:34 +0200] rev 334
- Rework package yet again, correcting and introducing new buildconf logic:
Conditionally disable bootstrap stage comparison correctly, correct
english grammar, better find system as(1) and ld(1), indotruce detailed
optimization option messages, more completely guess cpu types, allow
profiled bootstrapping without a preinstalled GCC because many other
compilers have long since implemented 64-bit arithmetic, instruct make
to build sequentially (not in sparallel) when building a profiled
bootstrap as GCC online documents recommend, and generally improve
comment blocks.
The single most important correction in this changeset relates to the
GCC changed optimization policy since at least GCC 4.5, in which -march
is always passed and not always correctly guessed. In the case of this
package, allowing GCC to guess the architecture leads to wild build
errors at various subcomponents (zlib, libgcc, libiberty...) and
bootstrap stages. It seems quite platform specific, and the safest
approach to correcting this seems to be explicitly always specifying the
-march argument when bootstrapping GCC. Because the best choice 'native'
is not available when bootstrapping using a foreign (non GCC) compiler,
a guess is made according to rpmmacros l_platform in that case.
It is questionable as to whether these recent optimization changes
on the part of GCC or this package are compatible with each other,
or if either are complete or correct at all. At least applying these
corrections allows this package to build again in most cases test.
- Tue, 29 Mar 2011 19:46:35 +0200
- by Michael Schloh von Bennewitz <michael@schloh.com> [Tue, 29 Mar 2011 19:46:35 +0200] rev 333
- Correct and introduce slightly needed logic, leading to better reliability:
Update bash(1) patch logic, correct several buildconf make location
errors, correct failed bash(1) configure invocation, enable perl(1) to
build with unpathed make(1), patch rpm(1) to correctly link with
internal libdb(3), and lastly unsuscessfully try to reorganize rpm patch
hunks.
- Mon, 28 Mar 2011 19:41:02 +0200
- by Michael Schloh von Bennewitz <michael@schloh.com> [Mon, 28 Mar 2011 19:41:02 +0200] rev 332
- Completely rework package including...
Simpify 64-bit -fPIC logic, instruct make to build in parallel,
force error condition on failed profile dependency, build the
standard 'bootstrap' target as suggested, correct grammar, wording,
and punctuation in general, upgrade to latest upstream vendor version,
rearrange package dependencies mpc, mpfr, and gmp, correct buildconf
thanks to new lto-plugin forced, avoid failed platform specific stage
comparison, adjust patch code to reflect new version update, and most
importantly adjust optimization flags to platform. Please note that
block 'correct hardcoded paths' is likely no yet complete.
- Sat, 26 Mar 2011 17:31:59 +0100
- by Michael Schloh von Bennewitz <michael@schloh.com> [Sat, 26 Mar 2011 17:31:59 +0100] rev 331
- Upgrade to current upstream vendor version and correct autoconf logic.
- Sat, 26 Mar 2011 17:30:52 +0100
- by Michael Schloh von Bennewitz <michael@schloh.com> [Sat, 26 Mar 2011 17:30:52 +0100] rev 330
- Solve function declaration signature mismatch by redefining locally.
