Closed LinqLover closed 3 years ago
@dalehenrich Please see the questions above, CI fails probably because of point 2 :)
CI passed on my fork, still running for this repository. @dalehenrich Sorry for paging, but it would be great if you could give me a short answer on whether you would like to see additional changes on this PR (see above) or whether this can be merged! :-)
@dalehenrich Hey, how do you think about this PR? Should we add other convenience buttons or can I just press the Merge button? :-)
sorry to have missed your earlier ping, do I read this correctly that you are changing the behavior such that the default committish is no longer 'master'? Perhaps not specifying the comittish in the url results in 'master' being defaulted elsewhere? If so then I'm comfortable with this change ... as it reduces the number of places where the default committish is hard coded ...
I have a sorta rule that behavior cannot be changed, because you never know who might depend upon that behavior and there's nothing worse than updating a project only to have things stop working ....
I think adding a test is a good idea, to "keep things honest" moving forward ...
So thumbs up and I appreciate your contributions
you are changing the behavior such that the default committish is no longer 'master'? Perhaps not specifying the comittish in the url results in 'master' being defaulted elsewhere?
Exactly, the decision about the default commitish is now totally up to GitHub. See also MC[Fetch]GitHubRepository>>projectZipUrlFor:versionString:
and the GitHub API docs about the zipball
/tarball
endpoints.
I have a sorta rule that behavior cannot be changed, because you never know who might depend upon that behavior and there's nothing worse than updating a project only to have things stop working ....
I totally agree with this strategy in general, but if we took this literally in this example, it hypothetically might be possible that someone set up a GitHub repository with both a master
and a main
branch and configured the main
branch as the default in the repository settings. In this case, if they did not explicitly specify the commitish, in the past, it would have fallen back to master
but now to main
... But personally, I think this would be a too speculative apprehension, what do you think?
I think adding a test is a good idea, to "keep things honest" moving forward ...
On it
I totally agree with this strategy in general, but if we took this literally in this example, it hypothetically might be possible that someone set up a GitHub repository with both a master and a main branch and configured the main branch as the default in the repository settings. In this case, if they did not explicitly specify the commitish, in the past, it would have fallen back to master but now to main ... But personally, I think this would be a too speculative apprehension, what do you think?
Haha, when the underlying system changes, you have to adapt, so I am in favor of your changes ...
Alright, then we should be done! New smoke tests exist and I did not find any regressions in the CI, so I'm going to merge this now. Thanks for your support!
you are changing the behavior such that the default committish is no longer 'master'? Perhaps not specifying the comittish in the url results in 'master' being defaulted elsewhere?
Exactly, the decision about the default commitish is now totally up to GitHub. See also
MC[Fetch]GitHubRepository>>projectZipUrlFor:versionString:
and the GitHub API docs about thezipball
/tarball
endpoints.
Argh, I need to correct myself. 'master'
is still hard-coded somewhere else so currently the default branch name will still be replaced by master
. This is not a regression but nevertheless not so convenient ...
Usage:
Open questions:
path
to'repository'
or isn't this convention?id
tag in themonticello.meta/version
file below manually?