freebsd-nq

History

Jason Evans aa33517e94 Implement pthread_attr_[gs]etguardsize(). Non-default-size stacks used to be malloc()ed, but they are now allocated using mmap(), just as the default-size stacks are. A separate cache of stacks is kept for non-default-size stacks. Collaboration with: deischen		2001-07-20 04:23:11 +00:00
..
guard_b.c	Implement pthread_attr_[gs]etguardsize(). Non-default-size stacks used to	2001-07-20 04:23:11 +00:00
guard_b.exp	Implement pthread_attr_[gs]etguardsize(). Non-default-size stacks used to	2001-07-20 04:23:11 +00:00
guard_s.pl	Implement pthread_attr_[gs]etguardsize(). Non-default-size stacks used to	2001-07-20 04:23:11 +00:00
hello_b.c
hello_d.c
hello_d.exp
hello_s.c
join_leak_d.c	Add a test for PR 24345.	2001-05-20 23:12:13 +00:00
join_leak_d.exp	Add a test for PR 24345.	2001-05-20 23:12:13 +00:00
Makefile	Implement pthread_attr_[gs]etguardsize(). Non-default-size stacks used to	2001-07-20 04:23:11 +00:00
mutex_d.c
mutex_d.exp
propagate_s.pl	Add test to detect propagation of cancellation points within libc_r.	2000-04-26 23:25:58 +00:00
README
sem_d.c	Don't define _REENTRANT, since the Makefile does so.	2001-05-20 23:11:09 +00:00
sem_d.exp
sigsuspend_d.c
sigsuspend_d.exp
sigwait_d.c	Fix a typo.	2001-05-20 23:10:30 +00:00
sigwait_d.exp
verify	Update the verify script.	2001-05-20 23:11:54 +00:00

README

$FreeBSD$

This test suite is meant to test general functionality of pthreads, as well as
provide a simple framework for regression tests.  In general, this test suite
can be used with any pthreads library, but in reality there are a number of
libc_r-specific aspects to this test suite which would require some effort to
get around if testing another pthreads library.

This test suite assumes that libc_r is installed.

There are two forms of test that the 'verify' script understands.  The simpler
form is the diff format, where the output of the test program is diff'ed with
the correspondingly named .exp file.  If there is diff output, the test fails.
The sequence test format is somewhat more complex, and is documented in the
command line usage output for verify.  The advantage of this format is that it
allows multiple tests to pass/fail within one program.

There is no driving need for test naming consistency, but the existing tests
generally follow these conventions:

<name>_d.c <name>_d.exp     : Diff mode C test and expected output file.
<name>_s.c                  : Sequence mode C test.
<name>_b*.c                 : Back end C program used by perl tests.
<name>_d.pl <name>_d.pl.exp : Diff mode perl test and expected output file.
<name>_s.pl                 : Sequence mode perl test.

<name> is something descriptive, such as "pr14685" in the case of a PR-related
regression test, or "mutex" in the case of a test of mutexes.