This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH 09/10] Fix tst-mutexpi8


> On Thu, Jan 10, 2013 at 01:16:16PM -0800, Roland McGrath wrote:
> > This does not depend on anything else and should be separate.
> 
> The lock elision patchkit depends on it: you cannot successfully
> run the test suite without it, since it triggers this bug.
> 
> I prefer to have all patches needed in one patchkit.

The point is that with nit cleanup this one will go in right away.
The rest of the changes will be mulled over for a long time and
there's certainly no guarantee they will go in at all.  In general,
test changes can and should go in before changes to the code being
tested.

> > The log entry is inadequate.  If you write a log entry that
> > follows our standards, then this change will be fine.
> 
> Not fully sure what is missing? AFAIK it describes the situation.
> Since it's a simple situation only two sentences are needed.

ChangeLog entries must always say explicitly what they do with
concrete function, variable, and macro names in the standard format,
not vague high-level descriptions.  A correct entry for this
change is:

	* tst-mutexpi8.c (TYPE): New macro, set to PTHREAD_MUTEX_TIMED_NP.
	* tst-mutex8.c (do_test) [ENABLE_PI]: Set PTHREAD_PRIO_INHERIT 
	in attributes of mutexes created.

The error messages should also follow the existing examples in the
same file instead of being something that looks like a function
called "pthread_mutexattr_setprotocol3" exists.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]