-
-
Notifications
You must be signed in to change notification settings - Fork 10.7k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Make spyder.rb user python 3.4 #8130
Conversation
Apologies, I should have waited for feedback before merging. Elision of Of course, this is only applicable in the absence of a casked Spyder-Py3 package. According to the Spyder download page, the Python 3–based package not only exists, but is also more popular. With the intention of offering both packages, what is our best course of action? We would normally be inclined to move the less popular package to Versions. However, considering that both Python2 and Python3 are available through the main repository, we may want to be include both here. (Addendum: switching a payload without renaming the Cask is undesirable, and will be even more problematic once we implement our |
No worries. I left this purposefully unmerged for discussion on its desirability.
Agreed.
Isn’t the goal of storing metadata to mitigate those? |
As to this Cask, I reckon that the As to the larger point, we have done a fair job reducing the conflicting constraints on the token. However, there are two that remain: following the app bundle is good, because it gives at least some protection against collisions, and is straightforward to derive. But rename events are bad, because even after storing metadata, we don't track renames at all. So There are a few different ways to improve this situation, such as declaring a freeze on all existing tokens, or starting to track renames with a |
Hi guys, Spyder maintainer here. I released our dmgs for Python 2 and 3 with different names, so that your users can install them side by side without problems. I didn't know about your project until last week but it seems pretty cool, so I'd like to support it better. Please let me know if you need anything else from us. |
Thank your for the offer, @ccordoba12. @rolandwalker I certainly see your point. However, how do we add I also had the idea tracking renames was one of the features of storing metadata (even if not yet implemented). Was that idea dropped, or was it never there, and I just though it was? |
Hi @ccordoba12 ! Welcome! @vitorgalvao I have no strong opinion here; I was just trying to look at past practices and seek consistency. It looks like the way we handled most variants so far was to put one under the defined/default token in the main repo, and the other under a more freeform token in homebrew-versions. Like There are some other cases where both variants live in the main repo such as I never got too worried about oddball tokens in the versions repo because hopefully we will inline all that info into the primary Casks later using a As to tracking renames: no we don't have that feature, but I probably gave an impression there was an implementation just because I have raised the point often whenever we talked about upgrade/metadata. In fact there is no way to track renames currently, which was why stable tokens were a prerequisite for the upgrade verb. We could add a way to track rename events, but nobody has worked on it. |
I'd just want to let you know that we depend on our app name internally to check for several things, so if you rename |
@ccordoba12 No worries there; our discussion is only about the app’s token: meaning the name you type when telling homebrew-cask to interact with it, it’ll have no effect on the app bundle’s name.
Because they’re not different (numeric) versions of the same app, but different apps (editions) altogether, I’d wager; same reason there are a ton of Eclipse variants present. We can likely find examples that contradict such a “rule”, though, and I’d also like to see them all consolidated. For such cases, though, choosing a variant would have to be mandatory, since no single one could be assumed as the default (and I have no problem with that notion). As you, @rolandwalker, I have no strong opinion on this. It was actually due to forgetting that the rule to remove framework substrings exists that I submitted the change. It was due to @ndr-qef’s pertinent comment on the If you all agree with this solution, I’ll proceed to make the change. Else, we shall find a different one. |
Agreeable here. I also noticed a similar case |
Done in Homebrew/homebrew-cask-versions@a1ec501 and this very issue (updated). |
Make spyder.rb user python 3.4
Refs #8127
A change in the app bundle’s name implies a change in the token.
2
is included since it refers to the version of python, not the softeware itself.