/tg/Station 13

Is there a proper way to fix map conflicts when another PR that modifies a map file gets merged before mine?

So far I've been doing it the caveman way where I copy+paste my changes onto the version found on master. There has got to be a more efficient way.

If you're careful you can edit the .dmm file in a text editor and solve the conflict there, but it's kind of tricky. Your current method is not so bad. In your experience, do the conflicts appear spurious or are you editing the same locations as others are?

I'm working on automatic conflict resolver scripts for DMM and DMI files, but it'll be some time until I'm confident in their success rate.

SpaceManiac wrote:In your experience, do the conflicts appear spurious or are you editing the same locations as others are?

So far, I mostly have issues with conflicts when I make changes that can't easily be copypasted. For example, adding 20 light switches that are spread all over Pubbystation.

Update your master branch. Rebase changes onto master. Git reset to point of master update. Save maps in mapmerge.git reset to last commit. Mapmerge.
This way all your changes will be properly merged on new version of map.

Selea wrote:Update your master branch. Rebase changes onto master. Git reset to point of master update. Save maps in mapmerge.git reset to last commit. Mapmerge.
This way all your changes will be properly merged on new version of map.

Thank you for that, I'll test it on my branch tomorrow.

SpaceManiac wrote:If you're careful you can edit the .dmm file in a text editor and solve the conflict there, but it's kind of tricky. Your current method is not so bad. In your experience, do the conflicts appear spurious or are you editing the same locations as others are?

I'm working on automatic conflict resolver scripts for DMM and DMI files, but it'll be some time until I'm confident in their success rate.

When the old python merger was removed, the script for resolving map conflicts was also removed (although the improved merger didn't include a script like it). You can look at it, the results are as good as they can be, provided there are no bugs.