Re: [Emacs-diffs] trunk r115596: Minor fixes for recent openp changes.

[Jarek Czekalski]

I have problems with understanding the new code. Especially the 
TYPE_MINIMUM part. Paul, maybe you could add a comment to explain why 
such a value must be put here? Could this be initialized to 0 or to 
invalid_timespec? It would be easier to understand.

When the save_mtime is invalid (nsec < 0), the result of its comparison 
could be undefined. Maybe this comparison never happens, however 0 value 
would be safer.

I have these 2 doubts. Maybe the problem comes from lack of knowledge on 
my side. I don't know :) Just sharing thoughts.

Thanks
Jarek

-  struct timespec save_mtime;
-  int save_fd = 0;
+  struct timespec save_mtime = make_timespec (TYPE_MINIMUM (time_t), -1);
+  int save_fd = -1;

-                      if (!save_fd || timespec_cmp (save_mtime, mtime) < 0)
+              if (timespec_cmp (mtime, save_mtime) <= 0)