[patch 4/4] wordexp: avoid strdup and cleanup memory handling

[Corinna Vinschen]

On Oct  9 13:54, Peter Rosin wrote:
> On 2012-10-09 13:00, Peter Rosin wrote:
> > On 2012-10-09 11:58, Corinna Vinschen wrote:
> >> Hi Peter,
> >>
> >> On Oct  9 10:06, Peter Rosin wrote:
> >>> [...]
> >>> @@ -142,11 +148,14 @@ wordexp(const char *words, wordexp_t *pw
> >>>  
> >>>        num_words = atoi(tmp);
> >>>  
> >>> -      wordv = (char **)realloc(pwordexp->we_wordv,
> >>> -                               (pwordexp->we_wordc + num_words + offs + 1) * sizeof(char *));
> >>> +      if (pwordexp->we_wordv)
> >>> +        wordv = we_wordv_to_ext_wordv(pwordexp->we_wordv);
> >>                    ^^^^^^^^^^^^^^^^^^^^^
> >>                    This is supposed to be uppercase, right?
> > 
> > Right. I'm not sure how that ended up in there, but I have tested
> > this standalone without rebuilding newlib, so I guess I must have
> > screwed up the code juggling somewhere. I had it lower case first,
> > as an inline, but made it a macro instead. The reason for not going
> > with the inline function escapes me at the moment.
> > 
> > But that mixup of course make me unsure if the rest is also as I
> > intended. I'll get back once I have retested it...
> 
> Ok, I have retested and rechecked, and that case cock-up is the
> only badness I can find. Please go ahead.

Done.  Thanks for the patches!


Corinna

-- 
Corinna Vinschen
Cygwin Project Co-Leader
Red Hat