| commit | 29afaa5b8c81e09a86654c8b37fdf083f25d564f | |
| author | Kyle J. McKay <mackyle@gmail.com> | |
| Wed, 8 Jul 2020 03:55:38 +0000 (7 20:55 -0700) | ||
| committer | Kyle J. McKay <mackyle@gmail.com> | |
| Wed, 8 Jul 2020 03:55:38 +0000 (7 20:55 -0700) | ||
| tree | e1601141fa41457f7f931eb352534f334a32a7c3 | treesnapshot (tar.gz zip) |
| parent | 9d20b60fafb364d369c1633315c6e75639d86dba | commitdiff |
gitweb/genindex.sh: vastly improve speed of genindex
Replace the hodgepodge version of genindex.sh with something that
avoids spawning multiple git processes, slow shell while loops and
forking lots of subshells.
Since perl ultimately runs to update the owner name hash, just
have perl handle it all for an order of magnitude speed up.
While in there, also make sure the atomic move-into-place operation
takes place on the same device as the final file destinations --
otherwise the moves might not actually be atomic.
Signed-off-by: Kyle J. McKay <mackyle@gmail.com>
Replace the hodgepodge version of genindex.sh with something that
avoids spawning multiple git processes, slow shell while loops and
forking lots of subshells.
Since perl ultimately runs to update the owner name hash, just
have perl handle it all for an order of magnitude speed up.
While in there, also make sure the atomic move-into-place operation
takes place on the same device as the final file destinations --
otherwise the moves might not actually be atomic.
Signed-off-by: Kyle J. McKay <mackyle@gmail.com>
| gitweb/genindex.sh | diffblobblamehistory |