Problem with github PR on Julia/METADATA



First things first: I’m not great with github. The PR in question is here.

So, I was attempting to tag and update a new version of my package DependentBootstrap that fixes all deprecation warnings for v0.6 (and hopefully most from v0.7). This includes changing REQUIRE from julia v0.5 to v0.6.

To do this, I used PkgDev.tag() and PkgDev.publish(). However, my first time round, I did a “patch” increase (v"0.0.1" to v"0.0.2"). The PR on METADATA failed the travis-ci with a warning that I should have used a “minor” increase (v"0.0.1" to v"0.1.0"), since I changed REQUIRE from v0.5 to v0.6.

So I used PkgDev.tag() and PkgDev.publish() a second time to get the PR linked above (and closed the first PR). This one is passing the checks. However, when I look at the commits, it would appear that the commits from both PR’s are present.

I guess this makes sense, since I don’t think I deleted the first tag (v"0.0.2") (I’m not 100% sure how to do this), so the second PR is attempting to push both tags.

So, what is the best way for me to fix this? Should I close the second PR, attempt to delete the first tag (any clues on how?) and then use Pkg.publish() to open a third PR that hopefully contains only the commit from the second tag?

Any help would be greatly appreciated.




I would highly recommend just deleting the tag and setting up attobot.


Thanks for responding.

Okay, I just read about attobot and it sounds great. It also looks like it has an integrated tool for deleting tags, so should I reverse the order of your recommendations? ie set up attobot, then delete the tag using attobot?

Also, I’m assuming you also think I should close the PR that is currently open? (I really don’t know much about github…)


Well, there is a way to fix it, but I think that if you’re not comfortable with git then it’s much easier to just close the PR.

I don’t think it matters. Attobot has stuff for closing PRs that it made when you delete tags, but since it didn’t make this PR I am not sure it would do anything here.


Thanks again. I’ll give it a go tonight.




Wow. That was really easy. Awesome tip.


Seems that PkgDev is still the suggested method in the Julia manual?


We should get that changed.