Revision history - refs/heads/dcn-page-profiler-more-verbose-types - origin: https://github.com/JuliaLang/julia

visit type:

Newer
Older

Revision	Author	Date	Message	Commit Date
3f0f911	d-netto	09 February 2024, 04:32:01 UTC	better printing of types on page profiler	09 February 2024, 23:50:05 UTC
9523361	Sukera	09 February 2024, 15:18:03 UTC	Document & expose printing of custom `AbstractTestSet` (#53215) This was previously alluded to in the docs as being possible ("just record if we're not the top-level parent"), but trying to do that didn't actually do anything because the printing in the `DefaultTestSet` ignored anything that wasn't a `Test.Result` or `Test.DefaultTestSet`. This is particularly problematic if any of the child-testsets had failures, because those failures would never be shown to a user. Co-authored-by: Sukera <Seelengrab@users.noreply.github.com> Co-authored-by: Ian Butterworth <i.r.butterworth@gmail.com>	09 February 2024, 15:18:03 UTC
65d4ce2	Ben Baumgold	09 February 2024, 15:14:44 UTC	implement convert from NamedTuple to Tuple (#49879) See the discussion from Discourse: https://discourse.julialang.org/t/construct-tuple-containing-namedtuple-from-tuple	09 February 2024, 15:14:44 UTC
daac6a6	Simon May	09 February 2024, 15:13:33 UTC	doc: Rephrase some text referring to `DArray` (#53244) Some parts of `distributed-computing.md` were somewhat unclear/confusing after `DArray` was moved to `DistributedArrays.jl`.	09 February 2024, 15:13:33 UTC
36b7d3b	Harmen Stoppels	09 February 2024, 14:37:01 UTC	Add PGO+LTO Makefile (#45641) Adds a convenient way to enable PGO+LTO on Julia and LLVM together: 1. `cd contrib/pgo-lto` 2. `make -j$(nproc) stage1` 3. `make clean-profiles` 4. `./stage1.build/julia -O3 -e 'using Pkg; Pkg.add("LoopVectorization"); Pkg.test("LoopVectorization")'` 5. `make -j$(nproc) stage2` <details> <summary>* Output looks roughly like as follows</summary> ```c++ $ make -C contrib/pgo-lto top make: Entering directory '/dev/shm/julia/contrib/pgo-lto' llvm-profdata show --topn=50 /dev/shm/julia/contrib/pgo-lto/profiles/merged.prof \| c++filt Instrumentation level: IR entry_first = 0 Total functions: 85943 Maximum function count: 7867557260 Maximum internal block count: 3468437590 Top 50 functions with the largest internal block counts: llvm::BitVector::operator\|=(llvm::BitVector const&), max count = 7867557260 LateLowerGCFrame::ComputeLiveness(State&), max count = 3468437590 llvm::hashing::detail::hash_combine_recursive_helper::hash_combine_recursive_helper(), max count = 1742259834 llvm::SUnit::addPred(llvm::SDep const&, bool), max count = 511396575 llvm::LiveRange::overlaps(llvm::LiveRange const&, llvm::CoalescerPair const&, llvm::SlotIndexes const&) const, max count = 508061762 llvm::StringMapImpl::LookupBucketFor(llvm::StringRef), max count = 505682177 std::map<llvm::BasicBlock, BBState, std::less<llvm::BasicBlock>, std::allocator<std::pair<llvm::BasicBlock* const, BBState> > >::operator[](llvm::BasicBlock* const&), max count = 395628888 llvm::LiveRange::advanceTo(llvm::LiveRange::Segment const, llvm::SlotIndex) const, max count = 384642728 llvm::LiveRange::isLiveAtIndexes(llvm::ArrayRef<llvm::SlotIndex>) const, max count = 380291040 llvm::PassRegistry::enumerateWith(llvm::PassRegistrationListener), max count = 352313953 ijl_method_instance_add_backedge, max count = 349608221 llvm::SUnit::ComputeHeight(), max count = 336604330 llvm::LiveRange::advanceTo(llvm::LiveRange::Segment, llvm::SlotIndex), max count = 331030109 llvm::SmallPtrSetImplBase::insert_imp(void const), max count = 272966545 llvm::LiveIntervals::checkRegMaskInterference(llvm::LiveInterval&, llvm::BitVector&), max count = 257449540 LateLowerGCFrame::ComputeLiveSets(State&), max count = 252096274 /dev/shm/julia/src/jltypes.c:has_free_typevars, max count = 230879464 ijl_get_pgcstack, max count = 216953592 LateLowerGCFrame::RefineLiveSet(llvm::BitVector&, State&, std::vector<int, std::allocator<int> > const&), max count = 188013152 /dev/shm/julia/src/flisp/flisp.c:apply_cl, max count = 174863813 /dev/shm/julia/src/flisp/builtins.c:fl_memq, max count = 168621603 ``` </details> This results quite often in spectacular speedups for time to first X as it reduces the time spent in LLVM optimization passes by 25 or even 30%. Example 1: ```julia using LoopVectorization function f!(a, b) @turbo for i in eachindex(a) a[i] *= b[i] end return a end f!(rand(1), rand(1)) ``` ```console $ time ./julia -O3 lv.jl ``` Without PGO+LTO: 14.801s With PGO+LTO: 11.978s (-19%) Example 2: ```console $ time ./julia -e 'using Pkg; Pkg.test("Unitful");' ``` Without PGO+LTO: 1m47.688s With PGO+LTO: 1m35.704s (-11%) Example 3 (taken from issue #45395, which is almost only LLVM): ```console $ JULIA_LLVM_ARGS=-time-passes ./julia script-45395.jl ``` Without PGO+LTO: ``` ===-------------------------------------------------------------------------=== ... Pass execution timing report ... ===-------------------------------------------------------------------------=== Total Execution Time: 101.0130 seconds (98.6253 wall clock) ---User Time--- --System Time-- --User+System-- ---Wall Time--- --- Name --- 53.6961 ( 54.7%) 0.1050 ( 3.8%) 53.8012 ( 53.3%) 53.8045 ( 54.6%) Unroll loops 25.5423 ( 26.0%) 0.0072 ( 0.3%) 25.5495 ( 25.3%) 25.5444 ( 25.9%) Global Value Numbering 7.1995 ( 7.3%) 0.0526 ( 1.9%) 7.2521 ( 7.2%) 7.2517 ( 7.4%) Induction Variable Simplification 6.0541 ( 5.1%) 0.0098 ( 0.3%) 5.0639 ( 5.0%) 5.0561 ( 5.1%) Combine redundant instructions #2 ``` With PGO+LTO: ``` ===-------------------------------------------------------------------------=== ... Pass execution timing report ... ===-------------------------------------------------------------------------=== Total Execution Time: 72.6507 seconds (70.1337 wall clock) ---User Time--- --System Time-- --User+System-- ---Wall Time--- --- Name --- 36.0894 ( 51.7%) 0.0825 ( 2.9%) 36.1719 ( 49.8%) 36.1738 ( 51.6%) Unroll loops 16.5713 ( 23.7%) 0.0129 ( 0.5%) 16.5843 ( 22.8%) 16.5794 ( 23.6%) Global Value Numbering 5.9047 ( 8.5%) 0.0395 ( 1.4%) 5.9442 ( 8.2%) 5.9438 ( 8.5%) Induction Variable Simplification 4.7566 ( 6.8%) 0.0078 ( 0.3%) 4.7645 ( 6.6%) 4.7575 ( 6.8%) Combine redundant instructions #2 ``` Or -28% time spent in LLVM. `perf` reports show this is mostly fewer instructions and reduction in icache misses. --- Finally there's a significant reduction in binary sizes. For libLLVM.so: ``` 79M usr/lib/libLLVM-13jl.so (before) 67M usr/lib/libLLVM-13jl.so (after) ``` And it can be reduced by another 2MB with `--icf=safe` when using LLD as a linker anyways. - [x] Two out-of-source builds would be better than a single in-source build, so that it's easier to find good profile data --------- Co-authored-by: Oscar Smith <oscardssmith@gmail.com> Co-authored-by: Lilith Orion Hafner <lilithhafner@gmail.com>	09 February 2024, 14:37:01 UTC
27b31d1	Jishnu Bhattacharya	09 February 2024, 10:56:32 UTC	Reroute algebraic functions for `Symmetric`/`Hermitian` through triangular (#52942) This ensures that only the triangular indices are accessed for strided parent matrices. Fix #52895 ```julia julia> M = Matrix{Complex{BigFloat}}(undef, 2, 2); julia> M[1,1] = M[2,2] = M[1,2] = 2; julia> H = Hermitian(M) 2×2 Hermitian{Complex{BigFloat}, Matrix{Complex{BigFloat}}}: 2.0+0.0im 2.0+0.0im 2.0-0.0im 2.0+0.0im julia> H + H # works after this 2×2 Hermitian{Complex{BigFloat}, Matrix{Complex{BigFloat}}}: 4.0+0.0im 4.0+0.0im 4.0-0.0im 4.0+0.0im ``` This also provides a speed-up in several common cases (allocations mentioned only when they differ): ```julia julia> H = Hermitian(rand(ComplexF64,1000,1000)); julia> H2 = Hermitian(rand(ComplexF64,1000,1000),:L); ``` \| Operation \| master \| PR \| \| ---- \| ---- \| ---- \| \|`-H` \|2.247 ms \| 1.384 ms \| \| `real(H)` \|1.544 ms \|1.175 ms \| \|`H + H` \|2.288 ms \|1.978 ms \| \|`H + H2` \|5.139 ms \|3.287 ms \| \| `isdiag(H)` \|23.042 ns (1 allocation: 16 bytes) \|16.778 ns (0 allocations: 0 bytes) \| I'm not entirely certain why `isdiag(H)` allocates on master, as union splitting should handle this automatically, but manually splitting the union appears to help.	09 February 2024, 10:56:32 UTC
63e95d4	Ian Butterworth	09 February 2024, 07:34:34 UTC	Add nested precomp highlight debug (#53111)	09 February 2024, 07:34:34 UTC
e507785	Simon Etter	08 February 2024, 22:30:27 UTC	Support escape expressions in @kwdef (#53230) `@kwdef` currently does not handle escaped type names and default values correctly. This makes it impossible to write a (correctly escaped) macro that internally uses `@kwdef`: ```julia julia> module M macro define_struct(name) quote @kwdef struct $(esc(name)) field::Int end end end end; julia> @macroexpand M.@define_struct(Foo) ERROR: Invalid usage of @kwdef Stacktrace: [1] error(s::String) @ Base ./error.jl:35 [2] var"@kwdef"(__source__::LineNumberNode, __module__::Module, expr::Any) @ Base ./util.jl:603 [3] #macroexpand#66 @ Base ./expr.jl:122 [inlined] [4] top-level scope @ REPL[5]:1 ``` This PR patches `@kwdef` so various types of escaping are handled correctly. Specifically, all of the following now works (see new test case [here](https://github.com/ettersi/julia/blob/adf7e20378ae7d0c3cf24a2cc3c815e560913b78/test/misc.jl#L1290-L1332)): ```julia module KwdefWithEsc const Int1 = Int const val1 = 42 macro define_struct() quote @kwdef struct $(esc(:Struct)) a b = val1 c::Int1 d::Int1 = val1 $(esc(quote e f = val2 g::Int2 h::Int2 = val2 end)) $(esc(:(i = val2))) $(esc(:(j::Int2))) $(esc(:(k::Int2 = val2))) l::$(esc(:Int2)) m::$(esc(:Int2)) = val1 n = $(esc(:val2)) o::Int1 = $(esc(:val2)) $(esc(:p)) $(esc(:q)) = val1 $(esc(:s))::Int1 $(esc(:t))::Int1 = val1 end end end end module KwdefWithEsc_TestModule using ..KwdefWithEsc const Int2 = Int const val2 = 42 KwdefWithEsc.@define_struct() end ``` --------- Co-authored-by: inky <git@wo-class.cn>	08 February 2024, 22:30:27 UTC
aba8afc	tecosaur	08 February 2024, 20:35:32 UTC	Improve truecolor terminal capability detection (#53235) Terminal color detection is a fickle beast, but it should behave as expected more often with these changes. Note to self: don't go "investigating" informal terminal standards again if I value my time, or sanity.	08 February 2024, 20:35:32 UTC
dd83530	inky	08 February 2024, 18:47:15 UTC	deps/csl: Use platform-dependent path separator (#52687) When building julia from source in MSYS2, you need to use `;` as a path separator, otherwise you won't be able to find the required DLLs! ## pr changes - On win, use `;` as path separator - Always search for DLLs in the `bin/` directory ### test steps Clean csl first: `make -C deps/ USE_SYSTEM_CSL=1 clean-csl` Before pr: ```sh $ make -C deps/ USE_SYSTEM_CSL=1 install-csl make: Entering directory '/d/jl/julia/deps' make: [/d/jl/julia/deps/csl.mk:59: /d/jl/julia/usr/bin/libgfortran-3.dll] Error 1 (ignored) make: [/d/jl/julia/deps/csl.mk:60: /d/jl/julia/usr/bin/libgfortran-4.dll] Error 1 (ignored) make: [/d/jl/julia/deps/csl.mk:61: /d/jl/julia/usr/bin/libgfortran-5.dll] Error 1 (ignored) make: [/d/jl/julia/deps/csl.mk:64: /d/jl/julia/usr/bin/libquadmath-0.dll] Error 1 (ignored) make: [/d/jl/julia/deps/csl.mk:65: /d/jl/julia/usr/bin/libstdc++-6.dll] Error 1 (ignored) make: [/d/jl/julia/deps/csl.mk:66: /d/jl/julia/usr/bin/libssp-0.dll] Error 1 (ignored) make: [/d/jl/julia/deps/csl.mk:67: /d/jl/julia/usr/bin/libatomic-1.dll] Error 1 (ignored) make: [/d/jl/julia/deps/csl.mk:68: /d/jl/julia/usr/bin/libgomp-1.dll] Error 1 (ignored) make: [/d/jl/julia/deps/csl.mk:75: /d/jl/julia/usr/bin/libgcc_s_seh-1.dll] Error 1 (ignored) make: [/d/jl/julia/deps/csl.mk:86: /d/jl/julia/usr/bin/libwinpthread-1.dll] Error 1 (ignored) make: Leaving directory '/d/jl/julia/deps' ``` After pr: ```sh $ make -C deps/ USE_SYSTEM_CSL=1 install-csl make: Entering directory '/d/jl/julia/deps' make: [/d/jl/julia/deps/csl.mk:59: /d/jl/julia/usr/bin/libgfortran-3.dll] Error 1 (ignored) make: [/d/jl/julia/deps/csl.mk:60: /d/jl/julia/usr/bin/libgfortran-4.dll] Error 1 (ignored) make: [/d/jl/julia/deps/csl.mk:66: /d/jl/julia/usr/bin/libssp-0.dll] Error 1 (ignored) make: Leaving directory '/d/jl/julia/deps' ``` ## Env - Win 11 - MSYS2 `STD_LIB_PATH` ``` $ uname -a MINGW64_NT-10.0-22631 A309-Y9000P 3.4.10.x86_64 2023-11-30 06:09 UTC x86_64 Msys $ make -C deps/ print-PATHSEP print-STD_LIB_PATH make: Entering directory '/d/jl/julia/deps' PATHSEP=; STD_LIB_PATH=D:/env/msys64/mingw64/bin/../bin/gcc/x86_64-w64-mingw32/13.2.0/;D:/env/msys64/mingw64/bin/../bin/gcc/;D:/env/msys64/mingw64/bin/../bin/gcc/x86_64-w64-mingw32/13.2.0/../../../../x86_64-w64-mingw32/bin/x86_64-w64-mingw32/13.2.0/;D:/env/msys64/mingw64/bin/../bin/gcc/x86_64-w64-mingw32/13.2.0/../../../../x86_64-w64-mingw32/bin/ ;D:/env/msys64/mingw64/bin/../bin/gcc/x86_64-w64-mingw32/13.2.0/;D:/env/msys64/mingw64/bin/../bin/gcc/;D:/env/msys64/mingw64/bin/../bin/gcc/x86_64-w64-mingw32/13.2.0/../../../../x86_64-w64-mingw32/bin/x86_64-w64-mingw32/13.2.0/;D:/env/msys64/mingw64/bin/../bin/gcc/x86_64-w64-mingw32/13.2.0/../../../../x86_64-w64-mingw32/bin/../bin/;D:/env/msys64/mingw64/bin/../bin/gcc/x86_64-w64-mingw32/13.2.0/../../../x86_64-w64-mingw32/13.2.0/;D:/env/msys64/mingw64/bin/../bin/gcc/x86_64-w64-mingw32/13.2.0/../../../../bin/;D:/a/msys64/mingw64/bin/x86_64-w64-mingw32/13.2.0/;D:/a/msys64/mingw64/bin/../bin/;D:/env/msys64/mingw64/bin/../bin/gcc/x86_64-w64-mingw32/13.2.0/../../../../x86_64-w64-mingw32/bin/;D:/env/msys64/mingw64/bin/../bin/gcc/x86_64-w64-mingw32/13.2.0/../../../;D:/a/msys64/mingw64/bin/ make: Leaving directory '/d/jl/julia/deps' ```	08 February 2024, 18:47:15 UTC
a0989f4	Shruti Dalvi	08 February 2024, 16:25:54 UTC	Added tests for kron! for diagonal matrices (#53184) Co-authored-by: Jishnu Bhattacharya <jishnub.github@gmail.com>	08 February 2024, 16:25:54 UTC
95df060	Alex Arslan	08 February 2024, 16:15:15 UTC	Qualify public, unexported bindings in REPL help (#52524) Fixes #52472, which was caused by `names` being changed to also return public, unexported symbols in #50105. Note that this restores previous behavior. A case could be made to instead add the public, unexported bindings as suggestions with the appropriate qualification. Not entirely sure how to test this so I'd welcome any suggestions. --------- Co-authored-by: Jameson Nash <vtjnash@gmail.com>	08 February 2024, 16:15:15 UTC
3dadada	Sukera	08 February 2024, 12:53:28 UTC	Allow using an unexported custom `AbstractTestSet` in `@testset` (#53212) Co-authored-by: Sukera <Seelengrab@users.noreply.github.com>	08 February 2024, 12:53:28 UTC
667cdde	Shuhei Kadowaki	08 February 2024, 12:39:13 UTC	compiler: fix the XXX test case in ssair.jl (#53245) This should live in test/compiler/inline.jl instead.	08 February 2024, 12:39:13 UTC
2b939f7	Christian Guinard	08 February 2024, 12:26:11 UTC	Add Debian terminfo directory default to find_terminfo_file (#51809) Without this, `Base.current_terminfo` returns `TermInfo(String[]; 0 flags, 0 numbers, 0 strings)` because it cannot find the correct terminfo files. In the regular terminal, there was basic color and formatting functionality because `$TERM`s starting with `xterm` are treated specially, but in a tmux session, where the `$TERM = tmux-256color`, there was no color at all. This also has the side-effect of enabling strikethrough and italics in the default gnome terminal emulator. The readme found in `/etc/terminfo/README` (as mentioned in `terminfo(5) "Fetching Compiled Descriptions"`) says the following: ``` /etc/terminfo/README This directory is for system-local terminfo descriptions. By default, ncurses will search ${HOME}/.terminfo first, then /etc/terminfo (this directory), then /lib/terminfo, and last not least /usr/share/terminfo. ``` I believe the manual might be different depending on the distro.	08 February 2024, 12:26:11 UTC
815cfd5	Richie Wilynnton	08 February 2024, 12:13:23 UTC	Added documentation to libdl module (#52825) Part of #52725 --------- Co-authored-by: RichieWilynnton <richiewilynton@gmail.com> Co-authored-by: Steven G. Johnson <stevenj@mit.edu> Co-authored-by: Jameson Nash <vtjnash@gmail.com> Co-authored-by: inky <git@wo-class.cn> Co-authored-by: Max Horn <max@quendi.de>	08 February 2024, 12:13:23 UTC
c1a7bb2	Ian Butterworth	08 February 2024, 12:05:47 UTC	🤖 [master] Bump the Pkg stdlib from f3b81f1aa to 6dd0e7c9e (take 2) (#53220) Replaces https://github.com/JuliaLang/julia/pull/53216 (I can't push to that repo branch) Adds - making pkgimages for REPLExt - Fixes PkgCompletionProvider location Stdlib: Pkg URL: https://github.com/JuliaLang/Pkg.jl.git Stdlib branch: master Julia branch: master Old commit: f3b81f1aa New commit: 6dd0e7c9e Julia version: 1.11.0-DEV Pkg version: 1.11.0 Bump invoked by: @IanButterworth Powered by: [BumpStdlibs.jl](https://github.com/JuliaLang/BumpStdlibs.jl) Diff: https://github.com/JuliaLang/Pkg.jl/compare/f3b81f1aac77acf08f5d847ead29ad0a228dec67...6dd0e7c9e99d578aa5477e2c78c91a161ce4c357 ``` $ git log --oneline f3b81f1aa..6dd0e7c9e 6dd0e7c9e Revert "add test for recurring precompile" (#3779) dbf114fa9 Merge pull request #3777 from JuliaLang/kc/repl_extension a49d47981 fixup e3edf3917 avoid prompting for git creds on CI (#3778) 510454343 make the `REPL` specific code into an extension 12d2de14e make `pkgstr` independent of REPL ``` --------- Co-authored-by: Dilum Aluthge <dilum@aluthge.com> Co-authored-by: Valentin Churavy <vchuravy@users.noreply.github.com> Co-authored-by: Kristoffer Carlsson <kcarlsson89@gmail.com> Co-authored-by: Jameson Nash <vtjnash@gmail.com>	08 February 2024, 12:05:47 UTC
ebe947e	Lars Göttgens	08 February 2024, 11:57:22 UTC	Add docstring for InteractiveUtils module (#53206) Handles the `InteractiveUtils` part of https://github.com/JuliaLang/julia/issues/52725. cc @stevengj	08 February 2024, 11:57:22 UTC
e68e432	Kristoffer Carlsson	08 February 2024, 11:56:45 UTC	allow printing some TOML dictionaries inline by marking them with an IdSet (#53233) This was one way I came up with to "mark" what dictionaries one wants to print inline. I am not sure `IdSet` is even considered public but it seemed the most reasonable for this.	08 February 2024, 11:56:45 UTC
4c2df21	James Wrigley	08 February 2024, 11:55:17 UTC	Add `Pipe` to the documentation (#53202) The reasoning is that `Pipe` is specifically documented in other functions that are part of the public API (e.g. `redirect_stdio()`), so it should be documented too. Same goes for `link_pipe!()` since it's mentioned in the `Pipe` docstring. The other function that could be documented is `open_pipe!()`, but that specifically applies to `PipeEndpoint`'s and I could imagine that we want it to be difficult to create a half-initialized `Pipe` :sweat_smile: I wrote a docstring for `link_pipe!()` based off the libuv docs: https://docs.libuv.org/en/v1.x/pipe.html#c.uv_pipe	08 February 2024, 11:55:17 UTC
b5bd10e	Florian	08 February 2024, 11:54:08 UTC	Add `track_content` option to allow hashing of `include_dependency`s (#51798) Continuation of #49866. Fixes #52462 So far any `include_dependency` was tracked by `mtime`. A package using `include_dependency` can't be relocated without recompilation if the dependency also needs to be relocated. With `include_dependency(path, track_content=true)` the tracking works like for `include`, i.e. recompilation is only triggered when the file's content changes. In case `path` is a directory we use the string `join(readdir(path))` as content. --------- Co-authored-by: Valentin Churavy <vchuravy@users.noreply.github.com>	08 February 2024, 11:54:08 UTC
15e2af2	jam-khan	08 February 2024, 11:52:35 UTC	Added docstring for Artifacts.jl (#52913) This is a part of issue #52725. --------- Co-authored-by: Steven G. Johnson <stevenj@mit.edu> Co-authored-by: Max Horn <max@quendi.de>	08 February 2024, 11:52:35 UTC
c5ad467	Knut Andreas Meyer	08 February 2024, 11:44:55 UTC	Document --heap-size-hint in Command-line Interface (#50480) Almost a direct copy of the output in `julia --help`. Closes #50588 Co-authored-by: Max Horn <max@quendi.de>	08 February 2024, 11:44:55 UTC
2673c14	Matt Bauman	08 February 2024, 11:39:42 UTC	Clarify `==` docs for collections (#52495) Alright, here's my spitball at fixing https://github.com/JuliaLang/julia/issues/52484. This also clarifies the non-missing behaviors, and adds a reference to `all` which should be a useful breadcrumb. I don't want to say `all(x .== y)` because that has all sorts of broadcasting semantics, but we could perhaps add a comparison to `all(splat(==), zip(x, y))` or `all(map(==, x, y))`, but even those feel sketchy to reference as they may miss other properties (like array shape). --------- Co-authored-by: Jameson Nash <vtjnash@gmail.com> Co-authored-by: inky <git@wo-class.cn> Co-authored-by: Max Horn <max@quendi.de>	08 February 2024, 11:39:42 UTC
41a6e7b	Mosè Giordano	08 February 2024, 10:03:05 UTC	[LibCURL_jll] Upgrade to v8.6.0 (#53222)	08 February 2024, 10:03:05 UTC
5cb5cd8	Artem Pelenitsyn	08 February 2024, 09:25:36 UTC	code_warntype docs: more neutral reference to @code_warntype (#51361) fix #51358 --------- Co-authored-by: Fredrik Ekre <ekrefredrik@gmail.com> Co-authored-by: Max Horn <max@quendi.de>	08 February 2024, 09:25:36 UTC
4414599	Ian Butterworth	08 February 2024, 09:24:39 UTC	Fix PkgId repr (#52795) Fixes #52793 ``` % ./julia --startup-file=no -E 'repr(Base.PkgId(Base.UUID("295af30f-e4ad-537b-8983-00126c2a3abe"), "Revise"))' "Base.PkgId(UUID(\"295af30f-e4ad-537b-8983-00126c2a3abe\"), \"Revise\")" ``` ``` julia> Base.PkgId(Base.UUID("295af30f-e4ad-537b-8983-00126c2a3abe"), "Revise") Revise [295af30f-e4ad-537b-8983-00126c2a3abe] ```	08 February 2024, 09:24:39 UTC
4d0a469	N5N3	08 February 2024, 06:22:48 UTC	SubArray: avoid invalid elimination of singleton indices (#53228) close #53209	08 February 2024, 06:22:48 UTC
2bd4cf8	Jishnu Bhattacharya	08 February 2024, 06:22:24 UTC	Avoid allocations in views of views (#53231) Currently, views-of-views construct their re-indexed indices by slicing into the parent indices. This PR changes this to use views of the parent indices instead. This makes the operation faster and non-allocating if the `parentindices` for the original view are `Vector`s. ```julia julia> a = rand(200, 200); julia> av = view(a, collect.(axes(a))...); julia> @btime view($av, axes($av)...); 312.393 ns (4 allocations: 3.25 KiB) # master 7.354 ns (0 allocations: 0 bytes) # PR ``` Indexing into the resulting view seems equally fast in simple cases: ```julia julia> av2 = view(av, axes(av)...); julia> @btime sum($av2); 66.883 μs (0 allocations: 0 bytes) # master 66.888 μs (0 allocations: 0 bytes) # PR julia> av2 = view(av, collect.(axes(av))...); julia> @btime sum($av2); 66.886 μs (0 allocations: 0 bytes) # master 66.891 μs (0 allocations: 0 bytes) # PR ``` --------- Co-authored-by: N5N3 <2642243996@qq.com>	08 February 2024, 06:22:24 UTC
a6ce761	Steve Kelly	08 February 2024, 04:50:19 UTC	add docs and jldoctests for `Base.shell_split` and `Base.rstrip_shell` (#52520) These were undocumented and I have seen `Base.shell_split` used in the package ecosystem. This adds some docs for clarity although neither of these functions are part of Base's exports. --------- Co-authored-by: Jameson Nash <vtjnash@gmail.com>	08 February 2024, 04:50:19 UTC
72d3abe	Max Horn	07 February 2024, 21:28:09 UTC	Fix potential instability/invalidation in dlpath (#53232) While debugging some invalidations and instabilities in code in packages I work on, one of the issues I stumbled over is this code in this `dlpath` method: ```julia function dlpath(libname::Union{AbstractString, Symbol}) handle = dlopen(libname) path = dlpath(handle) dlclose(handle) return path end ``` The `dlopen` modified in this PR can in principle return `nothing`. But there is no `dlpath` method for this. If one loads just a plain Julia, all is fine, but under certain conditions (deep in a call chain analyzed with Cthulhu.jl) it ended up not being able to prove that `path` will be a string, and only inferred it as `Any`. But if `throw_error` is set to `true` (the default, and used in the relevant code path) then `dlopen` cannot return `nothing`, it always returns a `String`. But the Julia compiler can't know that, as the exception is thrown by a C helper. So instead modify the Julia code a bit to help Julia deduce this fact by itself.	07 February 2024, 21:28:09 UTC
18df941	Navid C. Constantinou	07 February 2024, 20:44:57 UTC	doc: cosmetic whitespace changes in documentation.md (#53183)	07 February 2024, 20:44:57 UTC
e2c8809	Valentin Churavy	07 February 2024, 19:36:12 UTC	Disambiguate get_world_counter and get_inference_world (#53088) Add brief docs to `Base.get_world_counter` and add `Base.tls_wold_age`. Furthermore to disambiguate `get_world_counter` from the inference world rename the abstract interpreter accessor to `get_inference_world`	07 February 2024, 19:36:12 UTC
d24316a	Nathan Zimmerberg	07 February 2024, 15:53:14 UTC	Fix Int8(-128) // Int8(-128) (#51944) Fixes #51731 and #51730 Also fixes one of the previously broken tests: ```julia @test_broken Rational{Int64}(UInt(1), typemin(Int32)) == Int64(1) // Int64(typemin(Int32)) ``` This PR ensures the `Rational{T}` constructor with concrete `T` will only throw if the final numerator and denominator cannot be represented by `T`, or are both zero. If the `T` in `Rational{T}` is not concrete, this PR tries to ensure the numerator and denominator are promoted to the same type. This means `-1*Rational{Integer}(-1, 0x01) == 1` doesn't throw now. A side effect of this is that `Rational{Integer}(Int8(-1), 0x01)` now throws. Also, related to <https://github.com/JuliaLang/julia/pull/25702#issuecomment-359821951>, now `divgcd` doesn't change the types of its inputs and can handle `typemin`, but it still throws a `DivideError` if both inputs are zero.	07 February 2024, 15:53:14 UTC
736eeda	Michael Goerz	07 February 2024, 15:03:21 UTC	docs: remove outdated discussion about externally changing module bindings (#53170) As of Julia 1.9, bindings in modules can be changed directly. See https://discourse.julialang.org/t/clarify-the-documentation-about-modifying-module-variables/109668/3	07 February 2024, 15:03:21 UTC
bead1d3	Steven G. Johnson	07 February 2024, 14:50:51 UTC	fix spurious overflow for Float16(::Rational) (#52395) Fixes #52394. Also fixes `Float32` for `UInt128`, since currently `Float32((typemax(UInt128)-0x01) // typemax(UInt128))` gives `Nan32`.	07 February 2024, 14:50:51 UTC
d765ad1	Arjun R Maniyani	07 February 2024, 13:56:14 UTC	Added documentation for the Markdown Macro (#52607) Co-authored-by: Jameson Nash <vtjnash@gmail.com>	07 February 2024, 13:56:14 UTC
e22db88	Mosè Giordano	07 February 2024, 13:08:24 UTC	[deps] Do not build OpenBLAS Bfloat16 kernels in from source build (#53221) Not clear why OpenBLAS build fails, GCC 10 should be sufficient to compile the Bfloat16 kernels and [from what I can tell](https://github.com/JuliaLang/julia/issues/53172#issuecomment-1930144786) that's the compiler version used in CI, but I don't know how to verify it since this is a nightly job. If someone who knows more about the setup can chime in, that'd be great. In the meantime, disabling these kernels should fix #53172.	07 February 2024, 13:08:24 UTC
d453af8	Steve Kelly	07 February 2024, 04:53:09 UTC	Fix escaping of docstring in `@__DIR__` (#53225) This issue was introduced in #52442. The markdown parser had issues with the triple-quotes. Usually you can use quadruple-quotes to fix this, but this does not appear to be supported, so we escape each with backslash. Here is the corrected docstring rendered: ```julia-repl help?> @__DIR__ @__DIR__ -> String Macro to obtain the absolute path of the current directory as a string. If in a script, returns the directory of the script containing the @__DIR__ macrocall. If run from a REPL or if evaluated by julia -e <expr>, returns the current working directory. Example ≡≡≡≡≡≡≡ The example illustrates the difference in the behaviors of @__DIR__ and pwd(), by creating a simple script in a different directory than the current working one and executing both commands: julia> cd("/home/JuliaUser") # working directory julia> # create script at /home/JuliaUser/Projects open("/home/JuliaUser/Projects/test.jl","w") do io print(io, """ println("@__DIR__ = ", @__DIR__) println("pwd() = ", pwd()) """) end julia> # outputs script directory and current working directory include("/home/JuliaUser/Projects/test.jl") @__DIR__ = /home/JuliaUser/Projects pwd() = /home/JuliaUser ```	07 February 2024, 04:53:09 UTC
41bbfb4	Michael Abbott	07 February 2024, 02:39:00 UTC	Add explicit mention of branch cuts to `log`, `sqrt`, `angle` docstrings (#52970) I realised that `angle` doesn't even specify that it's in -π to π. These are all pretty standard but perhaps we should be explicit.	07 February 2024, 02:39:00 UTC