Re: [PATCH] add pool pointer to svn_string_t

So Greg, you're arguing we toss svn_string_t and use "(const char *) +
pools" as a substitute wherever we can. It seems that your points
are:

        1. We're hardly using svn_string_t's buffering features

        2. Where we *do* use it, it would be cleaner, simpler code
            just to create a new (char *) by reallocing from a pool

        3. Where we need actual variable length "byte" strings
            (rather than C strings), use an svn_buffer_t.

These seem like reasonable ideas, but I'd like to ponder them. What
you are advocating is *not* a trivial thing -- it would take quite a
while to go through all the code and make these cleanups... definitely
a post-milestone-1 thing.
Received on Sat Oct 21 14:36:10 2006