diff options
author | Dan Fandrich <danf@mageia.org> | 2024-03-14 18:50:26 -0700 |
---|---|---|
committer | Dan Fandrich <danf@mageia.org> | 2024-03-14 19:21:14 -0700 |
commit | c9bb45ef0018eddbd0172f38db46a96e961a7bd1 (patch) | |
tree | 57051ac16ceb59d6b5e238e6e531c715aa6387ab /tmpl/report.txt | |
parent | c9c5444f4f27f0a7375088a1cc0b4594742d8fe8 (diff) | |
download | mgaadvisories-c9bb45ef0018eddbd0172f38db46a96e961a7bd1.tar mgaadvisories-c9bb45ef0018eddbd0172f38db46a96e961a7bd1.tar.gz mgaadvisories-c9bb45ef0018eddbd0172f38db46a96e961a7bd1.tar.bz2 mgaadvisories-c9bb45ef0018eddbd0172f38db46a96e961a7bd1.tar.xz mgaadvisories-c9bb45ef0018eddbd0172f38db46a96e961a7bd1.zip |
Use multiprocessing in mksite for improved speeduser/danf/osv
This generates templated files using some parallelism, reducing the
total mksite time to less than half in my tests. Increasing parallelism
even further is possible, but would make the code harder to understand.
The obvious technique of generating each templated file in its own
process is actually far slower because the overhead of process creation
dwarfs the time spent processing the template, which is on average very
small and quick.
Diffstat (limited to 'tmpl/report.txt')
0 files changed, 0 insertions, 0 deletions