Hi,
For your information:
The recent sbase patches have been reverted.
For me it was easier to git push -f before the patches and reverts happened.
Some of the issues:
- head: performance regression (simple test showed it was 3x slower because of
less efficient getchar buffering and locking), it did not improve the code in
significant way.
Theres room for optimization and simplification, but it should be done in a
different way. Maybe just a simple read(), write() loop.
- sort: regression in sorting, for example when the input data had spaces and
sort -k -u 1,1 was used (reported on IRC, thanks).
- random.c logic: it is better to use libc for this.
- sbase didn't compile anymore (stdint.h was needed for random.c).
- cron: was already partly reverted due to regressions.
As always contributions are welcome. Make sure to test them properly though,
mistakes can happen, but sloppiness should be avoided.
And a reminder: writing clear commit messages or descriptions really helps in
reviewing for others.
--
Kind regards,
Hiltjo
Received on Sun Dec 22 2024 - 12:08:09 CET