[PATCH] read-cache: avoid memcpy in expand_name_field in index v4

perf reports memcpy at the the 6th position [1] in "git status -uno" using index v4, and strbuf_remove() in expand_name_field() accounts for 25% of that. What we need here is a simple string cut and a cheaper strbuf_setlen() should be enough. After this change, memcpy drops down to the 13th position [2] and is dominated by read_index_from.

[1] before
+     15.74%   git  git                [.] blk_SHA1_Block
+     13.22%   git  [kernel.kallsyms]  [k] link_path_walk
+     10.91%   git  [kernel.kallsyms]  [k] __d_lookup
+      8.17%   git  [kernel.kallsyms]  [k] strncpy_from_user
+      4.75%   git  [kernel.kallsyms]  [k] memcmp
+      2.42%   git  libc-2.11.2.so     [.] memcpy

[2] after
+     16.30%   git  git                [.] blk_SHA1_Block
+     13.43%   git  [kernel.kallsyms]  [k] link_path_walk
+     11.45%   git  [kernel.kallsyms]  [k] __d_lookup
+      8.73%   git  [kernel.kallsyms]  [k] strncpy_from_user
+      5.14%   git  [kernel.kallsyms]  [k] memcmp
+      2.29%   git  [kernel.kallsyms]  [k] do_lookup
+      2.21%   git  libc-2.11.2.so     [.] 0x6daf6
+      1.98%   git  [kernel.kallsyms]  [k] _atomic_dec_and_lock
+      1.98%   git  [kernel.kallsyms]  [k] _raw_spin_lock
+      1.86%   git  [kernel.kallsyms]  [k] acl_permission_check
+      1.61%   git  [kernel.kallsyms]  [k] kmem_cache_free
+      1.59%   git  git                [.] unpack_trees
+      1.47%   git  libc-2.11.2.so     [.] memcpy

Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
---
 I was after something else when I noticed this. Seems like a simple
 and safe change.

 read-cache.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)