| Age | Commit message (Collapse) | Author |
|---|
|
the list of built files and switch the md5.h header to the libiberty
one. This change is needed by gcc4 and it is a major library crank for
libiberty.
tested by many people on many architectures and finally ok'd by deraadt@
|
|
when all archs are fully tested and working.
- major bump libstdc++
|
|
handling thread safe and enables thread safety for libstdc++
- include -D_REENTRANT when -pthread cpp arg is present (needed so
boost can detect -pthread on the command line).
- bump libstdc++ major
See faq/current.html for instructions on how to correctly build
userland for this change.
okay kettenis@
|
|
- fix g++ regress so it can find libstdc++ in the objdir
okay espie@ kettenis@
|
|
Slipped in when fixing libstdc++ for gtk2mm, spotted the hard way when
porting nemiver.
ok espie@
|
|
#include <math.h> and expect the C99 macros in that file to be available in
the global namespace. This include was removed upstream some time after
the libstdc++ we use was released.
ok naddy@
|
|
more default stuff.
as discussed with kettenis, landry, and others.
Allow some gnome stuff (gtk2mm) to compile, and does not break source
compatibility with the standard (well, not more than gcc 4.2 does)
|
|
|
|
functions: acos, asin, atan, atan2, cos, cosh, exp, fabs, floor,
log, log10, sqrt, sin, sinh, tan, tanh
fixes quite some ports
tested by brad@. no objections millert@
|
|
|
|
can use these functions in C++.
ok millert@. also looked over by espie@, fgsch@.
|
|
|
|
|
|
as has been done for the __sentinel__ attribute, since all toolchains support
it.
|
|
last update.
|
|
ok miod@
|
|
as the demangler provides new interfaces.
|
|
- pexecute implementation replaced with the older gcc 2.95 implementation,
as gcc 2.95's collect2 relies upon an implementation behaviour which has
been lost in further libiberty updates.
- keep the old hash table interfaces, which are used by gcc 2.95.
- keep the OpenBSD replacement for the md5 interface.
- keep the OpenBSD strl{cat,cpy} and snprinft usage.
Bump libiberty minor version due to the addition of a few new symbols.
|
|
identifies itself as the gcc 3 version...)
|
|
The sh architecture required some different linkage than most platforms,
This fixes those other requirements.
|
|
(GNU people high on LSD... ? ??? )
|
|
|
|
- disable thread priorities in libobjc to avoid thread starvation issues
- major bump lib due to new link dependency on pthread
- link objc regress with -pthread
Tested on i386/amd64/powerpc/landisk/vax/sparc64/hppa. okay espie@
|
|
pointed out by deanna@
|
|
Requested by Mr T.
|
|
various people.
|
|
warning otherwise.
ok herrb@
|
|
Approved deraadt@, kettenis@
|
|
|
|
ok beck@
|
|
reviewed by millert, otto, kevlo, naddy, kettenis...
libc+libstdc++ bump
|
|
okay and comments from otto@ (static cast) and kettenis@.
Major bump.
|
|
ok espie@
|
|
I did check on i386 that the compiled code was identical.
However, nm shows different symbols in old libobjc compiled with gcc 2.95
versus symbols in new libobjc compiled with gcc 3.3.5.
Since that makes Theo uneasy, and since I didn't remember to ask miod
to verify nm before/after the libobjc change on all architectures,
let's crank the major number.
This is probably 99.9% unnecessary, but the crank takes care of the
0.1% chance that I fucked up and that this change will break, say,
the Objective C ABI on mvme88k.
Fucked up process on my part. Theo's right to remind me.
|
|
gcc 2.95.
tests and okay miod@, marc@.
|
|
again...
|
|
bring some code from more recent gcc, in order to have size_t,
and to simplify things a little.
okay otto@ and friends.
|
|
API addition -> minor bump.
(part of the strlcpy change okayed by otto@ and friends, that's the
part that requires the bump)
|
|
|
|
ok pvalchev@, fine deraadt@
|
|
- activate partial wchar_t support;
- kill old compatibility function that happens to use strcpy, poor guy.
|
|
_GLIBCPP_USE_TYPE_WCHAR_T
|
|
|
|
|
|
create an actual openbsd configuration that we can tweak.
okay kettenis@
|
|
ok deraadt@, pvalchev@
|
|
ok deraadt@
|
|
ok deraadt@
|
|
|
|
the machine architecture. We now output amd64 instead
of x86_64 as it should be.
ok deraadt@ pvalchev@
|
