Hyrum K Wright wrote: > On Sun, May 8, 2011 at 3:34 AM, <rhuij...@apache.org> wrote: > > - for (i = relpath_depth(local_relpath); i > op_depth; --i) > > + for (i = (int)relpath_depth(local_relpath); i > op_depth; --i) > > > > - for (i = relpath_depth(current_relpath); i > op_depth; i--) > > + for (i = (int)relpath_depth(current_relpath); i > op_depth; i--) > > > > - lock_depth = relpath_depth(local_relpath); > > + lock_depth = (int)relpath_depth(local_relpath); > > > > - levels += relpath_depth(lock_relpath); > > + levels += (int)relpath_depth(lock_relpath); > > > > - int depth = relpath_depth(local_relpath); > > + int depth = (int)relpath_depth(local_relpath); > > > > - lock_level = relpath_depth(local_relpath); > > + lock_level = (int)relpath_depth(local_relpath);
> Instead of all the casting, why don't we just change the types of the > various local variables? Alternatively... I'd like to change all our uses of apr_int64_t for (wc_id, repos_id, op_depth) to plain "int". I think that would be better because it would only occupy one syllable of thought-space rather than eight syllables. And I believe there's no functional or programmatic reason or benefit for having these three types be int64_t. To do that, I would need to introduce an additional type letter into the "bindf" format, because some parameters will still need to be int64_t (I can't remember what, but I looked at this before and there were some). Note that revnum_t is not relevant because it already has its own type letter "r". I was thinking like this, in sequential steps: 1. In "bindf" introduce letter "I" for int64_t (in parallel with the current "i" which is currently for int64_t). 2. Change bindf's callers to use "I" for any parameters that are still going to be using int64_t (i.e. not wc_id, repos_id, or op_depth). 3. Change the types of (wc_id, repos_id, op_depth) variables throughout the code to "int", and simultaneously change "bindf" so type code "i" means plain "int". Thoughts? - Julian
