Hi,
On Sun, May 22, 2016 at 11:00:22AM -0400, Selva Nair wrote:
> > On Sun, May 22, 2016 at 12:46:27AM -0400, Selva Nair wrote:
> > > {
> > > size += 10;
> > > - handles = realloc (handles, size * sizeof (HANDLE));
> > > - if (handles == NULL)
> > > - return ERROR_OUTOFMEMORY;
> > > + LPHANDLE tmp = realloc (handles, size * sizeof (HANDLE));
> > > + if (tmp == NULL)
> >
> > ... oh, and this might actually fail MSVC compilation, due to
> > "variable declared after other code parts" (size += 10)...
>
> Isn't it time to deprecate support for visual studio older than 2013? MSVC
> in VS-2015 supports C99 constructs and VS2013 may also do.
>
> If not, I'll submit a version 2.
Dunno... the only time I compiled on windows, I was using mingw+msys... but
since I get yelled at if we break MSVC compatibility, I started to look
out for such constructs.
Lev, what are you compiling with?
[..]
> > So we do not exit anymore when this fails, except if WAIT_FAILED?
>
> That is correct. I tested it and works nicely -- i.e., on realloc error,
> all previous connections continue while new connections get an error
> message.
This is very nice indeed!
(Not that a realloc() failure for +40 bytes would ever happen on its
own - if memory is that tight, other stuff will explode as well :-) - but
I really really like such defensive code that just works with what it
has, instead of "oh, panic, just crash")
Mmmh, anyway, I have two options now :-) - ask you for a v2 that is
MSVC 2010 compatible, and has the initial value reverted to "10", or
just apply...
gert
--
USENET is *not* the non-clickable part of WWW!
//www.muc.de/~gert/
Gert Doering - Munich, Germany g...@greenie.muc.de
fax: +49-89-35655025 g...@net.informatik.tu-muenchen.de
signature.asc
Description: PGP signature
