| 1 |
Hi, |
| 2 |
|
| 3 |
Just a quick note: thanks to hard work of radhermit, the newest release |
| 4 |
of pkgcheck features internal parallelization of checks. While it's not |
| 5 |
perfect and there's still room for improvement (I mean, it could run |
| 6 |
even faster!), it's a major step forward. |
| 7 |
|
| 8 |
This means that I've finally retired that ugly hack that split |
| 9 |
categories into multiple groups and ran them in parallel. This means |
| 10 |
that CI is no longer bound by the longest category (hello, media-libs/) |
| 11 |
and can utilize all 32 cores instead of ~14. |
| 12 |
|
| 13 |
This also means that if you're planning to run tree-wide scans with |
| 14 |
pkgcheck, you no longer have to employ huge hacks to make them faster. |
| 15 |
You just run it normally! In your face, repoman! |
| 16 |
|
| 17 |
Once again, thanks to radhermit. His work on pkgcheck is astonishing, |
| 18 |
and it's becoming a great tool for developers. |
| 19 |
|
| 20 |
-- |
| 21 |
Best regards, |
| 22 |
Michał Górny |
Archives