Try bumping JuliaSyntax to 0.4.10 #56110
Conversation
|
REPLCompletions issue is: vs |
This adopts the semantics discussed in #1289, namely the `else` block is executed whenever the loop never runs. Unlike Python, `break` and `continue` are irrelevant. Multidimensional loops are not supported since there is some ambiguity whether e.g. ```julia for i in 1:3, j in 1:0 print(1) else print(2) end ``` should print 2 once, thrice or maybe not even at all. Currently only supported in the flisp parser, so requires `JULIA_USE_FLISP_PARSER=1`. I could use some guidance on the necessary steps to add this to JuliaSyntax as well - AFAIU this would also require #56110 first. closes #1289
|
@c42f could you take a look at the JuliaSyntax issue that's blocking this? Hopefully it should be fairly straightforward, but I'd really like to make sure that we can start bumping JuliaSyntax again. |
The last time we tried to bump JuliaSyntax (#53119), it didn't work out due to test failues. As a result, we haven't actually bumped JuliaSyntax in about a year and there are several syntax changes in JuliaSyntax that we haven't picked up yet (i.e. are waiting in PRs). I don't know what the current situation is, but I think we need to get back to a point where we can bump JuliaSyntax regualarly, so let's try to take the first step.
|
I think this broke Revise tests which didn't run on this PR for some reason. Maybe it needed a rebase to pick up on the new buildkite config? On https://buildkite.com/julialang/julia-master/builds/42947#0193b85a-4e0e-4101-b248-0e43faef0607 there is I'll see if I can fix it quickly on the Revise side. |
|
That error happens when Revise fails to parse base. one possible reason is the include issue I fixed last week. I'm on my phone so it's hard to see if that build is on master. If it's not, them the error is unsurprising because it needs my fixes from last week. Otherwise it is indeed possible that this commit broke it and will need to look at what suppressed error revise had while parsing Base. |
|
Maybe it is fine considering for example #56798 is all green which was made after this PR. I also couldn't immediately reproduce this locally. The PRs that fail probably have an old base commit. (I still think it is kinda crazy not to run CI on the merge commit). |
|
Yeah, old base commits will fail, because they're using Revise master, which needs a fix on julia master. |
The last time we tried to bump JuliaSyntax (#53119), it didn't work out due to test failues. As a result, we haven't actually bumped JuliaSyntax in about a year and there are several syntax changes in JuliaSyntax that we haven't picked up yet (i.e. are waiting in PRs). I don't know what the current situation is, but I think we need to get back to a point where we can bump JuliaSyntax regualarly, so let's try to take the first step.