bug#6600: [PATCH] sort: add --threads option to parallelize internal sort.

2010-07-13 Thread Pádraig Brady
Here's a small cleanup I missed. Alternatively one could make heap() return NULL rather than aborting, but since it already used xmalloc, I'm tending to this.. commit cec33eb226df63f406f7eb70cd46d960ee02a060 Author: Pádraig Brady p...@draigbrady.com Date: Tue Jul 13 08:23:52 2010 +0100

bug#6600: [PATCH] sort: add --threads option to parallelize internal sort.

2010-07-13 Thread Jim Meyering
Pádraig Brady wrote: Here's a small cleanup I missed. Alternatively one could make heap() return NULL rather than aborting, but since it already used xmalloc, I'm tending to this.. commit cec33eb226df63f406f7eb70cd46d960ee02a060 Author: Pádraig Brady p...@draigbrady.com Date: Tue Jul 13

bug#6600: [PATCH] sort: add --threads option to parallelize internal sort.

2010-07-13 Thread Jim Meyering
Pádraig Brady wrote: On 13/07/10 15:35, Jim Meyering wrote: I noticed that heap_insert's reallocation was awkward and inefficient. Using x2nrealloc rather than xrealloc makes the code cleaner as well as more efficient in the face of a growing heap, and also handles integer overflow. diff

bug#6600: [PATCH] sort: add --threads option to parallelize internal sort.

2010-07-13 Thread Pádraig Brady
On 13/07/10 15:35, Jim Meyering wrote: I noticed that heap_insert's reallocation was awkward and inefficient. Using x2nrealloc rather than xrealloc makes the code cleaner as well as more efficient in the face of a growing heap, and also handles integer overflow. diff --git a/gl/lib/heap.c

bug#6600: [PATCH] sort: add --threads option to parallelize internal sort.

2010-07-13 Thread Chen Guo
Thanks a lot for all the hard work reviewing and revising this, Pádraig.