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 v2 2/4] manual: Add documentation for pthread_tryjoin_np and pthread_timedjoin_np


On Thursday 26 September 2019 at 14:26:00 -0700, Adhemerval Zanella wrote:
> On 18/09/2019 05:30, Mike Crowe wrote:
> > This documentation isn't perfect, but it's better than nothing and can
> > hopefully act as a basis for future improvement. In particular, although
> > I'm certain that the functions are MT-Safe, I'm not sure about the rest of
> > the @safety line.
> > 
> > Yann Droneaud pointed out that pthread_timedjoin_np would wait forever if
> > the timeout parameter is passed as NULL. Since it's now too late to change
> > that, I've documented it.
> 
> To which semantic do you think it would be better to change? Since this
> symbol are not yet in POSIX there still the possibility to change it
> by using a new plus a compat symbol.

The alternative would be to mark the timeout parameter as __nonnull and not
permit it to be NULL, as is the case for the other pthread timed*/clock*
functions. I imagine that if the functions were ever standardised then this
would be the case - but that doesn't look like it's on the horizon.

Although we wouldn't break existing binaries if we change the behaviour and
add a compat symbol, we could break people compiling existing source. I
don't think it's a serious enough problem for it to be worth even doing
that.

However, if others feel strongly about it then I'm prepared to make that
change.

Thanks.

Mike.


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