| commit | e2a865854cf4729b3944ae7ee5179d939c140abe | |
| author | Kyle J. McKay <mackyle@gmail.com> | |
| Sun, 19 Apr 2015 01:12:05 +0000 (18 18:12 -0700) | ||
| committer | Kyle J. McKay <mackyle@gmail.com> | |
| Sun, 19 Apr 2015 01:12:05 +0000 (18 18:12 -0700) | ||
| tree | 8285973c2dd7303c295453721ba032feb70e5d12 | treesnapshot (tar.gz zip) |
| parent | 95d7756af9baa058186f784e3f6ab8f6b961fa4f | commitdiff |
gitweb: use length 32768 instead of 16384 when inspecting content
Since we copy raw data from the input to the output 32768 bytes
at a time, we will end up lining up better with the underlying
buffers if we also read the initial 32768 bytes for inspection
instead of only 16384.
We only look at it anyway if the type appears to be text/html
and it does not contain valid UTF-8 byte sequences.
Signed-off-by: Kyle J. McKay <mackyle@gmail.com>
Since we copy raw data from the input to the output 32768 bytes
at a time, we will end up lining up better with the underlying
buffers if we also read the initial 32768 bytes for inspection
instead of only 16384.
We only look at it anyway if the type appears to be text/html
and it does not contain valid UTF-8 byte sequences.
Signed-off-by: Kyle J. McKay <mackyle@gmail.com>
| gitweb/gitweb.perl | diffblobblamehistory |