pgRouting / admin

Cross-project tasks and organizational items
1 stars 0 forks source link

Move code to master on alpha? #8

Open cvvergara opened 6 years ago

cvvergara commented 6 years ago

@dkastl Up to today, when releasing -alpha -beta RC, the code was not moved to master. Should it stay that way? What is best? @robe2 you have some experience on this topic?

robe2 commented 6 years ago

@cvvergara I don't think it matters too much. I'd lean more toward changing it because you want people to test it before you release and they probably won't if it's not master and they are used to pulling from master.

In the case of both PostGIS / PostgreSQL master (the default branch) always has the bleeding edge stuff not the stable stuff. So pgRouting is different already in that major respect that what we do can't be applied to what you do.

On PostGIS we don't define a new branch until we officially release but that's really a matter of laziness than anything. On PostgreSQL I think they create a new stable branch as soon as they are done with their codefests with the reasoning they can start development on new stuff right away without being hindered by release. Then their master becomes the next not yet released major version.

woodbri commented 6 years ago

On pgRouting 2.0, we used the branching model where the bleeding edge was "develop" branch and all releases were merged into master and tagged.

On 8/21/2017 9:40 PM, Regina Obe wrote:

@cvvergara https://github.com/cvvergara I don't think it matters too much. I'd lean more toward changing it because you want people to test it before you release and they probably won't if it's not master and they are used to pulling from master.

In the case of both PostGIS / PostgreSQL master (the default branch) always has the bleeding edge stuff not the stable stuff. So pgRouting is different already in that major respect that what we do can't be applied to what you do.

On PostGIS we don't define a new branch until we officially release but that's really a matter of laziness than anything. On PostgreSQL I think they create a new stable branch as soon as they are done with their codefests with the reasoning they can start development on new stuff right away without being hindered by release. Then their master becomes the next not yet released major version.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/pgRouting/admin/issues/8#issuecomment-323895553, or mute the thread https://github.com/notifications/unsubscribe-auth/AAxJguaMiP5LqQ38sbYVS67MOPq65tLnks5sajGHgaJpZM4O9_pn.


This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus

cvvergara commented 6 years ago

@woodbri Git branching model: http://nvie.com/posts/a-successful-git-branching-model/ So, if I am going to tag 2.5.0-alpha it has to move to master.

cvvergara commented 6 years ago

And put this in the readme:

cvvergara commented 6 years ago

Maybe they move because people tend to do PR to master branch, and actually, when the release is done, I create the next micro release when code changes on master, so kind of its the bleeding edge for the current minor.

dkastl commented 6 years ago

I don't have a strong opinion on this, but I like the following:

So if our pre-releases are seen as stable, then why not move them to master. I just want to avoid, that users clone master and compile an unstable version.

cvvergara commented 6 years ago

Yes, alpha is quite stable. develop is not