On Fri, Jan 02, 2009 at 09:48:54PM +0530, Balbir Singh wrote:
> * Dhaval Giani <[email protected]> [2009-01-02 17:55:14]:
> 
> > At a number of places I noticed a following combination of functions
> > 
> > build_path;
> > strcat;
> > 
> > Realizing that the way build_path is designed is not very conducive
> > to such a scenario, and also requires memory to have been pre-allocated
> > which caused a whole lot of other issues, such as using statically sized
> > arrays which could overflow, not giving us the opportunity to reallocate,
> > it looks like to have this sort of a function is a better option. We can
> > now have dynamically allocated strings which will help us move from static
> > arrays to dynamic arrays.
> > 
> > As of now, only build tested
> > 
> > Changes from v1
> > 1. Used vasprintf as pointed out by Dan Smith. Minimal testing shows the
> > usage is correct, but would require closer review
> > 2. Corrected some other bugs which were pointed out in testing
> 
> Bad changelog, please document them.
> 

This is exactly why I am supposed to document changes when I make them.
Let me try to recall these changes and repost.

> > 3. Minimally tested using the test cases Sudhir has submitted.
> > 
> > Signed-off-by: Dhaval Giani <[email protected]>
> 
> Both 4 and 5 look too ugly and big (in fact even 3) for review, I
> would like to review them *only* once they are tested and known to
> work.

I can split 4, introduce the API in one patch and use the API in another
one. I am not sure about 3 though. All the patches are tested with the
test cases we have and do not introduce new regressions as per the test
cases.

thanks,
-- 
regards,
Dhaval

------------------------------------------------------------------------------
_______________________________________________
Libcg-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/libcg-devel

Reply via email to