This is the mail archive of the
cygwin-developers@sourceware.cygnus.com
mailing list for the Cygwin project.
Re: patch submission: fixes in find_unused_handle() and cygwin_select()
- To: Lincoln Myers <lincoln at netapp dot com>
- Subject: Re: patch submission: fixes in find_unused_handle() and cygwin_select()
- From: Chris Faylor <cgf at cygnus dot com>
- Date: Sun, 10 Oct 1999 14:09:18 -0400
- Cc: cygwin-developers at sourceware dot cygnus dot com
- References: <199910070036.RAA20891@herra.corp.netapp.com>
- Reply-To: cygwin-developers at sourceware dot cygnus dot com
On Wed, Oct 06, 1999 at 05:36:58PM -0700, Lincoln Myers wrote:
> - hinfo::find_unused_handle() now accepts a start one greater
> than any existing descriptor, something which make_pipe()
> sometimes needs for the write fd, if the read fd was the
> last allocated.
This was actually fixed a couple of snapshots below.
> - select_stuff::wait() no longer uses a separate copy
> of the fd_sets given as arguments. If I understood
> the comment above cygwin_select(), it used to do
> this because the fd_sets given to cygwin_select()
> might not be a complete sizeof(fd_set) in size, but
> then select_stuff::wait() used to do a structure
> copy of r,w,s on top of readfds,writefds,exceptfds,
> so cygwin_select() was not really handling smaller
> fd_sets.
Actually, the select() comment was woefully out of date and had no
bearing on reality. The copy was done to avoid setting anything until
the select succeeded.
>Consequently, I'm pretty sure my changes
> handle caller's arguments at least as gingerly, and
> it seems to be perl-ly correct, though I'm only
> hoping that implies it to be posix-ly correct...
Unfortunately, I don't think that your change duplicates the functionality
of the original code. I'll take a look at select and see if there is
some other way to handle this.
cgf