-
-
Notifications
You must be signed in to change notification settings - Fork 117
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
fixes for nightly: Statistics stdlib #583
Conversation
@KristofferC, any advice on the
error that AppVeyor is showing on nightly? (I have to admit I find the name-resolution errors in the new package system rather frustrating.) |
The problem seems to be related to that the registry doesn't get downloaded at all Compare on Linux (where the registry get downloaded):
but on the mac builder
This could happen if there is an old, empty Even so, I will try debug it. |
Seems like something is wrong with how the current project is detected: #584 (comment). |
@@ -1716,7 +1747,7 @@ if VERSION < v"0.7.0-DEV.4585" | |||
end | |||
|
|||
if VERSION < v"0.7.0-DEV.4064" | |||
for f in (:mean, :cumsum, :cumprod, :sum, :prod, :maximum, :minimum, :all, :any, :median) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
If we remove these we'll break any packages which depend on this, right? I'm afraid we'll have to keep them, at least on Julia 0.6. That's what we have done for several other features which have changed over the 0.7 development period.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yes, I guess we'll need to replace this loop with a loop that imports them from Compat.Statistics
.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
While keeping them to avoid breaking packages that have already adapted to the keyword, we should erase them from the README though, to encourage use of Compat.Statistics
instead.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I'll just change the mentions in the README to Compat.Statistics.mean
etcetera.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Should be fixed now
I think this will break Compat on |
@fredrikekre, because of the |
Yes exactly. It will end up in the registry as a deps, like for StaticArrays: https://github.com/JuliaRegistries/Uncurated/blob/b4b9bbbd0a7a4aef7d5aa1ecb802f54694f5cf53/S/StaticArrays/Deps.toml#L15 So I think if this is merged and released, then Compat will only work on |
Ah, that is quite unfortunate. So we postpone this one until -rc1 is out and support for -beta may be considered less important? Or what's the best way forward? If we need to wait, pulling 9d264ec ( |
Yes, something like that. And then we can just as well avoid all the tricky version |
|
Let's rerun CI and see... |
Hmm, why does close and reopen not work? Rebase and push instead? |
Took the liberty and rebased this. |
Ok, CI is happy with this. Go? |
Thanks! |
* Bump required Julia version to 1.0 * Remove compatibility support code for: * `at-__MODULE__` (from #363) * `devnull`, `stdin`, `stdout`, and `stderr` from #499 * `at-nospecialize` (from #385 and #409) * `isabstracttype` and `isconcretetype` (from #477) * `invokelatest` from #424 * array-like access to `Cmd` from #379 * `Val(n)` and `ntuple`/`reshape` with `Val` from #381 and #399 * `logdet(::Any)` fallback from #382 * `chol(::UniformScaling)` from #382 * `pushfirst!`, `popfirst!` from #444 * `fieldcount` from #386 * `read(obj, ::Type{String})` from #385 and #580 * `InexactError`, `DomainError`, and `OverflowError` constructors from #393 * `corrected` kw arg to `cov` from #401 * `adjoint` from #401 * `partialsort` from #401 * `pairs` from #428 * `AbstractRange` from #400 * `rtoldefault` from #401 * `Dates.Period` rounding from #462 * `IterativeEigensolvers` from #435 * `occursin` from #520 * `Char` concatenation from #406 * `BitSet` from #407 * `diagm` and `spdiagm` with pairs from #408 * `Array` c'tors from `UniformScaling` from #412 and #438 * `IOContext` ctor taking pairs from #427 * `undef` from #417 and #514 * `get` on `ENV` from #430 * `ComplexF...` from #431 * `tr` from #614 * `textwidth` from #644 * `isnumeric` from #543 * `AbstractDict` from #435 * `axes` #435 and #442 * `Nothing` and `Cvoid` from #435 * `Compat.SuiteSparse` from #435 * `invpermute!` from #445 * `replace` with a pair from #445 * `copyto!` from #448 * `contains` from #452 * `CartesianIndices` and `LinearIndices` from #446, #455, and #524 * `findall` from #466 (and #467). * `argmin` and `argmax` from #470, #472, and #622 * `parentmodule` from #461 * `codeunits` from #474 * `nameof` from #471 * `GC` from #477 * `AbstractDisplay` from #482 * `bytesavailable` from #483 * `firstindex` and `lastindex` from #480 and #494 * `printstyled` from #481 * `hasmethod` from #486 * `objectid` from #486 * `Compat.find*` from #484 and #513 * `repr` and `showable` from #497 * `Compat.names` from #493 and #505 * `Compat.round` and friends #500, #530, and #537 * `IOBuffer` from #501 and #504 * `range` with kw args and `LinRange` from #511 * `cp` and `mv` from #512 * `indexin` from #515 * `isuppercase` and friends from #516 * `dims` and `init` kwargs from #518, #528, #590, #592, and #613 * `selectdim` from #522 and #531 * `repeat` from #625 * `fetch(::Task)` from #549 * `isletter` from #542 * `isbitstype` from #560 * `at-cfunction` from #553 and #566 * `codeunit` and `thisind` and friends from #573 * `something` from #562 * `permutedims` from #582 * `atan` from #574 * `split` and `rsplit` from #572 * `mapslices` from #588 * `floatmin` and `floatmax` from #607 * `dropdims` from #618 * required keyword arguments from #586 * `CartesianRange` in `at-compat` from #377 * `finalizer` from #416 * `readline`, `eachline`, and `readuntil` from #477, #541, and #575 * curried `isequal`, `==`, and `in` from #517 * `Some` from #435 and #563 * `at-warn` and friends from #458 * Remove old deprecations * Deprecate: * `Compat.Sockets` from #545 and #594 * `TypeUtils` from #304 * `macros_have_sourceloc` from #355 * `Compat.Sys` from #380, #433, and #552 * `Compat.MathConstants` from #401 * `Compat.Test`, `Compat.SharedArrays`, `Compat.Mmap`, and `Compat.DelimitedFiles` from #404 * `Compat.Dates` from #413 * `Compat.Libdl` from #465 (and #467) * `AbstractDateTime` from #443 * `Compat.Printf` from #435 * `Compat.LinearAlgebra` from #463 * `Compat.SparseArrays` from #459 * `Compat.Random` from #460, #601, and #647 * `Compat.Markdown` from #492 * `Compat.REPL` from #469 * `Compat.Serialization` from #473 * `Compat.Statistics` from #583 * `Fix2` from #517 * `Compat.Base64` from #418 * `Compat.Unicode` from #432 and #507 * `notnothing` from #435 and #563 * `Compat.IteratorSize` and `Compat.IteratorEltype` from #451 * `enable_debug(::Bool)` from #458 * `Compat.Distributed` from #477 * `Compat.Pkg` from #485 * `Compat.InteractiveUtils` from #485 * `Compat.LibGit2` from #487 * `Compat.UUIDs` from #490 * `Compat.qr` from #534 * `Compat.rmul!` from #546 * `Compat.norm` abd friends from #577 * Remove obsolete README entry, missed in #385 * Remove obsolete tests (e.g. missed in #372) * Remove obsolete `VERSION` conditionals and some minor clean-up
This fixes Compat to work with Julia nightly by adding
Compat.Statistics
(for JuliaLang/julia#27834) and theinit
keyword formapreduce
andreduce
(for JuliaLang/julia#27711). Hopefully tests should pass again now!I didn't add the
init
keyword support for*foldl
and*foldr
yet due to lack of time — they can always be added in a separate PR — but thereduce
andmapreduce
changes are urgent to work around JuliaLang/julia#27871