[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

nptl 0.47

Hash: SHA1

The sources of the next revision are available


Changes since 0.46 include:

~ tons of cancellation handling changes.  Mostly little stuff.  Half
  of the changed code is in glibc proper, so you need the current CVS

~ lots of new tests, mainly cancellation.  I think the primary
  cancellation points are now all covered

~ hammer problem with missing asm clobber fixed.  This apparently
  only hit while debugging

On x86 with a gcc 3.3 with forced unwind and cleanup attribute support
all tests pass.  Hammer, the only other platform I tested so far, has
problems with the generation of the unwind info in libc.so.  I'll need
somebody more familiar with the linker look at that.

ia64 I compiled a few days ago and since it doesn't use exception based
unwind yet it should still be fine.

This reminds me.  Hopefully nobody installs a glibc with nptl based on
the current sources for anything but x86 and even for x86 glibc must be
compiled with a gcc which supports the cleanup attribute.  Not doing any
of this creates big problems.

As already mentioned, you need to have the current CVS glibc.  Something
like those at timestamp


As for requests for me to spell out every single details of the build
process.  Any other wishes?  Maybe I should read you the morning paper?
 I have better things to do.  I only test with our internal tools and
occasionally the official gcc.  That's it.  The build process hasn't
changed since I first explained it.  If you want more detailed
descriptions maybe it's time for you to step up and do some work.

- -- 
- --------------.                        ,-.            444 Castro Street
Ulrich Drepper \    ,-----------------'   \ Mountain View, CA 94041 USA
Red Hat         `--' drepper at redhat.com `---------------------------
Version: GnuPG v1.2.1 (GNU/Linux)


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]