On Thu, May 04, 2017 at 06:42:58AM -0700, Bart Van Assche wrote:
> On 05/04/17 00:18, Magnus Fromreide wrote:
> > I just noticed that your alternate version of vasnprintf calls realloc on
> > the *strp argument but when I look at my system's version it seems to call
> > malloc and just assign the r
On Thu, May 04, 2017 at 06:40:14AM -0700, Bart Van Assche wrote:
> On 05/04/17 00:25, Magnus Fromreide wrote:
> > I looks like the change 9641b472ef01208f92631016e91b0a4a518163f0 (openssl
> > 1.1.0: Fix several build errors) broke the build on machines where openssl
> > is missing, I suppose this w
On 05/04/17 00:18, Magnus Fromreide wrote:
> I just noticed that your alternate version of vasnprintf calls realloc on
> the *strp argument but when I look at my system's version it seems to call
> malloc and just assign the result and given that one could imagine people
> calling it using an unini
On 05/04/17 00:25, Magnus Fromreide wrote:
> I looks like the change 9641b472ef01208f92631016e91b0a4a518163f0 (openssl
> 1.1.0: Fix several build errors) broke the build on machines where openssl
> is missing, I suppose this wasn't intentional?
Hello Magnus,
Indeed ... the goal was to fix build e
Hi!
I looks like the change 9641b472ef01208f92631016e91b0a4a518163f0 (openssl
1.1.0: Fix several build errors) broke the build on machines where openssl
is missing, I suppose this wasn't intentional?
/MF
--
Check out the
Hi!
I just noticed that your alternate version of vasnprintf calls realloc on
the *strp argument but when I look at my system's version it seems to call
malloc and just assign the result and given that one could imagine people
calling it using an uninitilaized pointer then that behaviour sounds mo
