diff options
author | Jim Meyering <jim@meyering.net> | 1998-04-06 00:26:54 +0000 |
---|---|---|
committer | Jim Meyering <jim@meyering.net> | 1998-04-06 00:26:54 +0000 |
commit | 72193ba149210654a35592bcbad41461c78c9ffc (patch) | |
tree | 70308c500e2446ba679c6f796f393c5e84af1ce9 /lib/TODO | |
parent | 854b5197ad7122b76778b960fd0663b1dca3b884 (diff) | |
download | coreutils-72193ba149210654a35592bcbad41461c78c9ffc.tar.xz |
*** empty log message ***
Diffstat (limited to 'lib/TODO')
-rw-r--r-- | lib/TODO | 14 |
1 files changed, 14 insertions, 0 deletions
@@ -14,3 +14,17 @@ like stdc-print and think I sent you an early verison. I don't like the name `cursor'. I much prefer short names like `p' for index variables. I doubt I'll change all of them, but thought you should know why some will probably end up changing. + +#define USE_OBSTACK somewhere + +Fix this comment. Depending on system and application... +Mention fragmentation. + +#if USE_OBSTACK + + /* Whenever obstacks are used, it is possible to allocate all overflowed + + entries into a single stack, so they all can be freed in a single + + operation. It is not clear if the speedup is worth the trouble. */ + + struct obstack entry_stack; + +#endif + + +assert (bucket_limit - bucket == n_buckets) ? |