I think I'm speaking for the majority of the curl user community when I say
that there's nothing to feel sorry for, quite to the contrary.
You've been doing a marvellous job in developing curl into one of the most
useful net-oriented tools in existence, and your responsiveness to feature
requests or bug reports has never left anything to be desired.
The latest "flood of memory bugs" is exactly what you called it, a temporary
inconvenience. And concerning the memory leaks discovered in and around
urlglob the fault lies with me, anyway. Your memory debugging initiative as
a whole is certainly appreciated because it will ultimately lead to a more
So quit rambling and get back to coding :-)).
> -----Original Message-----
> From: Daniel Stenberg [mailto:daniel_at_haxx.se]
> Sent: Tuesday, December 12, 2000 10:41 PM
> To: Curl Mailinglist
> Subject: Let me explain...
> ... the recent flood of memory bugs in src/main.c
> I feel I owe you this, as this explosion is a bit unusual for curl.
> Anyway, I made my memory debug system output any case where there is a
> malloc() without a corresponding free().
> Strictly speaking, those errors are only really harmful when
> found in the
> library, as the client code will exit anyway and everything
> will be freed.
> But, for completeness, I included the memory debug system all over.
> When I wrote my test suite, I've made it check the memory
> debug results after
> each test and this way I could fight all leaks that occurred.
> It then also
> forced me to clean up the memory usage in the client code as well.
> Unfortunately, I didn't fully realize how the urlglobbing
> worked together
> with the rest of the code (as that piece of code has been
> conitrbuted by
> Jörn) and now we've had to fight these bugs as a result of that.
> I'm sorry for the inconviniances and I hope you help me get over these
> Thanks for your patience.
> Daniel Stenberg -- curl project maintainer -- http://curl.haxx.se/
Received on 2000-12-13