Re: g_io_unix_read/g_io_unix_write

[Sebastian Wilhelmi <wilhelmi ira uka de>]

Hi David,

> Although I think it is guaranteed, I cannot find substantiating
> docs :(  The whole point of EINTR is to interrupt blocking system
> calls b/c something needs immediate attention.  Hence if you aren't
> going to block...

As far as I got it, it is a problem of some OSes, that they can't restart a
read/write or something, after a signal occured. But I might be wrong. Stevens
"Unix Network Programming" suggests that in saying: 
"Some kernels automatically restart some interrupted system calls. For
portability when we write a program that catches signals , we must be prepared
for system calls to return EINTR.

> > if (result == G_IO_ERROR_AGAIN ||
> >     result == G_IO_ERROR_UNKNOWN && errno == EINTR)
> >   /* do it again */
> 
> Obviously no one should be forced to write that.

I think, everyone should.
 
> I don't really have a problem with the patch, I just think
> it is completely unnecessary, b/c all it does is make
> broken code get G_IO_ERROR_AGAIN instead of
> G_IO_ERROR_UNKNOWN.

I dont think, code that gets EINTR is broken.

Bye,
Sebastian
-- 
Sebastian Wilhelmi                   |            här ovanför alla molnen
mailto:wilhelmi@ira.uka.de           |     är himmlen så förunderligt blå
http://goethe.ira.uka.de/~wilhelmi   |