В сообщении от Среда 05 мая 2010 20:47:36 автор Rafael Diniz написал:
> And how about x86_64, is uclibc/nptl running fine?

With all fixes I've submitted recently there are still some problems left as 
per nptl tests:

tst-cancel7
tst-cond8
tst-mutex8
tst-once3
tst-once4

For this group it seems that cancellations need more love on x86_64.

tst-cputimer*, although that work same as glibc versions apart from cputimer2 
which fails with "thread CPU clock timer support missing, skipping test" here.

tst-mqueue3
tst-mqueue5
tst-mqueue6
tst-mqueue8

Something is broken in mqueues.

All of above is considered non-critical for our project, but still it'd be 
nice to fix those if anyone cares and/or interested.

In generic code I have tst-tls13 and tst-tls16 failing, but that seems to be 
because of LAZY relocations, tst_wcsftime, didn't look into that closely and 
floating-point tests have some errors (that are also present when compiled for 
glibc and thus considered non-harmful) and differences in corner case handling 
(like pow(1, NaN) and such).

-- 
 http://roman.khimov.ru
mailto: [email protected]
gpg --keyserver hkp://subkeys.pgp.net --recv-keys 0xE5E055C3

Attachment: signature.asc
Description: This is a digitally signed message part.

_______________________________________________
uClibc mailing list
[email protected]
http://lists.busybox.net/mailman/listinfo/uclibc

Reply via email to