Replies: 3 comments
-
in the 1-2 months before the release, the main branch is unstable as we are consolidating all runners on all platforms, updating all dependencies (which usually does a lot of damage). Furthermore, most users are on python, java, .NET and we need to build the artefacts for the code to be tested. The current plan is to do a 9.13 soon without a release candidate (next week, or the week after, no promise). Then quickly do a 10.0 rc1 which will break compatibility with the routing library. I expect that 10.0 real release is months away, and multiple release candidates away. |
Beta Was this translation helpful? Give feedback.
-
My 2 cents:
|
Beta Was this translation helpful? Give feedback.
-
So v99bugfix branch is pretty close to release. Please hammer away at it :-) |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
Hello,
I'm thinking whether it is possible to improve release management of ortools stable versions. In last versions, especially for cp-sat solver are shortly after release of stable version found and often quickly fixed important issues for me on main branch but not in stable branch.
Is it possible for example one month before a release to annouce public testing of or-tools to find errors that influence usability of future stable version?
Or should be main branch considered as stable branch? What version is used internally in google?
In case someone even outside google should maintain stable version (it seems that ortools are used heavily also outside google) with added fixes from main branch it should be nice to mark important fixes for stable version (label should be enough) and don't mix various changes in one commit like 5e62dde this week when it seems that fix of time limit is mixed with changes in routing cuts.
To what extent do users prefer working with the stable version of OR-Tools?
Thanks,
Jan
Beta Was this translation helpful? Give feedback.
All reactions