| e9183b9 | Tim Holy | 28 February 2022, 11:48:25 UTC | Add valgrind suppression file for codegen Running `test-precompile` under valgrind generates about 12,000 warnings, of which the vast majority come from LLVM/codegen. This adds a suppression file to make it easy to exclude these warnings. | 28 February 2022, 11:48:25 UTC |
| fe1264a | Simon Byrne | 26 February 2022, 06:08:19 UTC | define round/trunc/ceil/floor to Bool (#25085) | 26 February 2022, 06:08:19 UTC |
| a2e4226 | Kristoffer Carlsson | 25 February 2022, 18:06:42 UTC | return immediately from `ht_keyindex` if dictionary is empty (#44341) Avoids hashing the input key for empty dictionaries. Inspired by rust-lang/hashbrown#305. | 25 February 2022, 18:06:42 UTC |
| 2e2c16a | N5N3 | 25 February 2022, 17:18:10 UTC | Fix `@simd` for non 1 step CartesianPartition (#42736) | 25 February 2022, 17:18:10 UTC |
| 9af12d3 | Dilum Aluthge | 25 February 2022, 16:38:08 UTC | Add some additional information to `LinearAlgebra.versioninfo()` (#44220) | 25 February 2022, 16:38:08 UTC |
| 9b8253c | Jerry Ling | 25 February 2022, 14:51:03 UTC | irrational `sign()` (#37949) | 25 February 2022, 14:51:03 UTC |
| 76fc067 | Yichao Yu | 25 February 2022, 14:48:24 UTC | Make sure all the relocations are filled in for partially cloned target (#44262) We collect the relocations (i.e. the GOT slots that is used in the code) for each target in `tgt.relocs`. Needing a relocation, however, does not imply that the function is cloned for this target within the group (It does mean that at least one target in the group has it cloned). The previous version would miss the relocation in this case. This was triggerred with the following cloning situation caller: clone_1 callee: clone_1, clone_1.clone_3 Since caller.clone_1 may call either callee.clone_1 or callee.clone_1.clone_3 a relocation for callee will be used and is required to be initialized. In addition to target 1, target 2 (and in fact target 3) within group 1 will also use caller.clone_1. However, since callee isn't cloned for target 2 the previous version wouldn't have saved this slot in the relocation array. | 25 February 2022, 14:48:24 UTC |
| c9cb015 | Kristoffer Carlsson | 25 February 2022, 14:47:23 UTC | add docs for creating a symbol from an enum instance (#44342) | 25 February 2022, 14:47:23 UTC |
| c6b5fa2 | Gerhard Aigner | 25 February 2022, 09:37:43 UTC | Indicate defaults in comandline help (#44223) Co-authored-by: Gerhard Aigner <aigner@trium.de> Co-authored-by: Jameson Nash <vtjnash@gmail.com> | 25 February 2022, 09:37:43 UTC |
| b6ffbeb | Tim Holy | 25 February 2022, 09:34:34 UTC | Add devdocs on 32-bit docker builds (#44300) Co-authored-by: Mosè Giordano <giordano@users.noreply.github.com> | 25 February 2022, 09:34:34 UTC |
| 163c032 | Fredrik Ekre | 25 February 2022, 08:36:39 UTC | Distributed: Propagate package environment to local workers. (#43270) Local workers now inherit the package environment of the main process, i.e. the active project, LOAD_PATH, and DEPOT_PATH. This behavior can be overridden by passing the new `env` keyword argument, or by passing `--project` in the `exeflags` keyword argument. Fixes #28781, and closes #42089. | 25 February 2022, 08:36:39 UTC |
| 6b29ebd | Elliot Saba | 24 February 2022, 22:52:14 UTC | [macOS] Codesign binary-dist tarballs (#44305) Because we're starting to distribute macOS tarballs as well, let's codesign them by default, when possible. | 24 February 2022, 22:52:14 UTC |
| f20d5de | Jeff Bezanson | 24 February 2022, 22:08:37 UTC | fix #43411, wrapped `NamedTuple` can be bitstype more often (#44311) | 24 February 2022, 22:08:37 UTC |
| 21e5a26 | Kristoffer Carlsson | 24 February 2022, 21:35:41 UTC | also set the version in `pkgorigins` (#44318) | 24 February 2022, 21:35:41 UTC |
| 8ae9dd5 | Fredrik Ekre | 24 February 2022, 19:13:15 UTC | Distributed: Update documentation for addprocs(...). (#44331) | 24 February 2022, 19:13:15 UTC |
| 6850940 | Mosè Giordano | 24 February 2022, 18:36:07 UTC | [OpenBLAS_jll] Update to v0.3.20 (#44321) | 24 February 2022, 18:36:07 UTC |
| df81bf9 | Tim Holy | 24 February 2022, 07:30:09 UTC | Cache external CodeInstances (#43990) Prior to this PR, Julia's precompiled `*.ji` files saved just two categories of code: unspecialized method definitions and type-specialized code for the methods defined by the package. Any novel specializations of methods from Base or previously-loaded packages were not saved, and therefore effectively thrown away. This PR caches all the code---internal or external---called during package definition that hadn't been previously inferred, as long as there is a backedge linking it back to a method owned by a module being precompiled. (The latter condition ensures it will actually be called by package methods, and not merely transiently generated for the purpose of, e.g., metaprogramming or variable initialization.) This makes precompilation more intuitive (now it saves all relevant inference results), and substantially reduces latency for inference-bound packages. Closes #42016 Fixes #35972 Issue #35972 arose because codegen got started without re-inferring some discarded CodeInstances. This forced the compiler to insert a `jl_invoke`. This PR fixes the issue because needed CodeInstances are no longer discarded by precompilation. | 24 February 2022, 07:30:09 UTC |
| fb85cb3 | Valentin Churavy | 24 February 2022, 00:35:22 UTC | Merge pull request #44316 from pchintalapudi/pc/no-getpointerelement Remove uses of getPointerElementType | 24 February 2022, 00:35:22 UTC |
| 0201ac8 | Jameson Nash | 23 February 2022, 22:03:06 UTC | expose `Filesystem.touch` for `Filesytem.File` objects (#44307) | 23 February 2022, 22:03:06 UTC |
| 1b47efb | Juan Ignacio Polanco | 23 February 2022, 20:07:30 UTC | Make sure `range(start; step, length)` uses TwicePrecision when possible, fixes #44292. (#44313) | 23 February 2022, 20:07:30 UTC |
| 262c6f1 | Max Horn | 23 February 2022, 19:10:59 UTC | Fix warnings about unused static function (#44255) | 23 February 2022, 19:10:59 UTC |
| 7b2d1e0 | Prem Chintalapudi | 23 February 2022, 18:48:30 UTC | Remove uses of getPointerElementType | 23 February 2022, 19:05:43 UTC |
| ed53276 | Colin Caine | 23 February 2022, 18:57:11 UTC | Define `isempty(g::Generator)` and `Base.isdone(g::Generator)` (#44308) Co-authored-by: Fredrik Ekre <ekrefredrik@gmail.com> | 23 February 2022, 18:57:11 UTC |
| 4737a6e | Valentin Churavy | 23 February 2022, 18:56:47 UTC | Merge pull request #44317 from imciner2/im/gcwarning gc: Cast to jl_value_t* to avoid compiler warning | 23 February 2022, 18:56:47 UTC |
| 0079041 | Ian McInerney | 23 February 2022, 15:27:25 UTC | gc: Cast to jl_value_t* to avoid compiler warning | 23 February 2022, 15:27:25 UTC |
| 629eb86 | gbaraldi | 23 February 2022, 12:38:45 UTC | Fix M1 tail call issue when building (#44279) | 23 February 2022, 12:38:45 UTC |
| 333a3a3 | Felix Kastner | 23 February 2022, 12:32:17 UTC | fix `BLAS.spr!` docstring (#44303) | 23 February 2022, 12:32:17 UTC |
| 114e33b | Colin Caine | 23 February 2022, 08:02:06 UTC | Fix typo in manual's description of `isdone` (#44306) Closes #44283 | 23 February 2022, 08:02:06 UTC |
| 26b0b6e | Shuhei Kadowaki | 23 February 2022, 05:08:22 UTC | inference: remove `CachedMethodTable` (#44240) Since we couldn't confirm any performance benefit from `CachedMethodTable` in the current infrastructure (see the benchmark results in #44240), now I'd like to propose to eliminate that entirely and save a bit of space. | 23 February 2022, 05:08:22 UTC |
| aa2421d | Viral B. Shah | 23 February 2022, 03:36:20 UTC | Documentation: added example to the function hasfield. (From #41782) (#44274) Co-authored-by: Johnny Chen <johnnychen94@hotmail.com> (cherry picked from commit 9c0e5b0d186ea95a06d5b0bdc4bc19d1a17b444d) | 23 February 2022, 03:36:20 UTC |
| 0db596e | pchintalapudi | 23 February 2022, 03:12:06 UTC | Remove uses of PointerType::getElementType for opaque pointers (#44310) Co-authored-by: Jameson Nash <vtjnash@gmail.com> | 23 February 2022, 03:12:06 UTC |
| dee1cd2 | Mosè Giordano | 23 February 2022, 00:16:23 UTC | [contrib] Update `download_cmake.sh` to CMake 3.19.3 and support aarch64 (#44267) * [contrib] Update `download_cmake.sh` to CMake 3.19.3 and support Linux aarch64 * [contrib] Fix `download_cmake.sh` for macOS, it was broken already before This also works on aarch64-apple-darwin. | 23 February 2022, 00:16:23 UTC |
| afd9926 | Jishnu Bhattacharya | 22 February 2022, 20:39:17 UTC | Implement the conversion of AbstractUnitRanges to OrdinalRanges (#40038) Co-authored-by: Jameson Nash <vtjnash@gmail.com> | 22 February 2022, 20:39:17 UTC |
| 4e57966 | Nathan Daly | 22 February 2022, 20:33:25 UTC | Fix thread-safety violation in Allocations Profiler: Create a new per-thread backtrace buffer in ptls (#44116) * Fix thread-safety violation in Allocations Profiler: Re-use the shared `ptls->bt_data` buffer from the thread-local storage for the buffer, to ensure that each thread has a separate buffer. This buffer is shared with the exception throwing mechanism, but is safe to share since julia exception throwing never interleaves with allocations profiling. * Approach two: Create a separate per-thread allocations backtrace buffer. * Update src/gc-alloc-profiler.cpp Co-authored-by: Jameson Nash <vtjnash@gmail.com> * Update src/gc-alloc-profiler.cpp Co-authored-by: Jameson Nash <vtjnash@gmail.com> * fix type error (#44235) * Update src/gc-alloc-profiler.cpp Co-authored-by: Jameson Nash <vtjnash@gmail.com> Co-authored-by: Pete Vilter <7341+vilterp@users.noreply.github.com> | 22 February 2022, 20:33:25 UTC |
| e492025 | Dilum Aluthge | 22 February 2022, 19:19:29 UTC | Distributed test suite: wrap another non-thread-safe test in `poll_while` (#44280) | 22 February 2022, 19:19:29 UTC |
| 8df488b | pchintalapudi | 22 February 2022, 16:20:22 UTC | Fix invalidations of various analysis results (#44284) | 22 February 2022, 16:20:22 UTC |
| 6fa255d | William Moses | 22 February 2022, 15:31:40 UTC | Avoid pointer indirection for blas call (#44271) | 22 February 2022, 15:31:40 UTC |
| cfd6119 | DilumAluthgeBot | 22 February 2022, 05:36:30 UTC | 🤖 Bump the Pkg stdlib from e31a3dc7 to 3aa15055 (#44297) Co-authored-by: Dilum Aluthge <dilum@aluthge.com> | 22 February 2022, 05:36:30 UTC |
| 01855a7 | Milan Bouchet-Valat | 22 February 2022, 03:35:19 UTC | Mark stack as non-executable (#43481) The linker is unable to detect that executable stack is not required and so marks libjulia.so as requiring it Pass `-Wl,-z,noexecstack` to ensure that the stack is marked as not executable. This improves security and allows Julia to run on systems where SELinux has been configured to disallow executable stacks. AFAIK no change is needed on OSes other than Linux as they do not allow executable stacks by default. | 22 February 2022, 03:35:19 UTC |
| 76cf176 | Tim Holy | 21 February 2022, 21:29:06 UTC | Doc build steps needed for analyzegc (#43901) | 21 February 2022, 21:29:06 UTC |
| c5bc69c | Ian Butterworth | 21 February 2022, 15:37:54 UTC | Fix bad precompile statements via an output lock (#44252) | 21 February 2022, 15:37:54 UTC |
| ac922b0 | Viral B. Shah | 21 February 2022, 15:04:07 UTC | Regenerate all the checksums (#44275) | 21 February 2022, 15:04:07 UTC |
| d75eea1 | Tamas K. Papp | 21 February 2022, 15:03:01 UTC | Add tests for issue #28869. (#44291) for the issue (and a trivial typo fix). | 21 February 2022, 15:03:01 UTC |
| 0b48b91 | Aditya Puranik | 21 February 2022, 09:47:52 UTC | Fix aliasing bug in copy!(x, x) for x::AbstractSet and x::AbstractDict, fixes #41268 (#44265) | 21 February 2022, 09:47:52 UTC |
| 0a294f9 | Shuhei Kadowaki | 21 February 2022, 09:44:26 UTC | doc: minor tweak on `startswith`/`endswith` docs (#44288) This change makes it more clear whether `startswith(prefix)` is equivalent to: - `s -> startswith(s, prefix)` or - `s -> startswith(prefix, s)` | 21 February 2022, 09:44:26 UTC |
| 1777553 | Shuhei Kadowaki | 21 February 2022, 06:15:52 UTC | inference: override return type inference by const-prop more carefully (#44282) A return type derived by const-prop' inference can be wider than that of non const-prop' inference in rare cases e.g. when there are cycles but cached result is still accurate. This commit checks if the const-prop'ed result is really more accurate than non-const result. fix https://github.com/Ferrite-FEM/Tensors.jl/issues/178 | 21 February 2022, 06:15:52 UTC |
| f879d20 | Valentin Churavy | 20 February 2022, 17:01:56 UTC | Merge pull request #44276 from JuliaLang/vc/newpm3 Make another wave of passes NewPM compatible | 20 February 2022, 17:01:56 UTC |
| f45b6ad | Mosè Giordano | 20 February 2022, 16:17:43 UTC | [CPUID] Add ISA entries for A64FX and M1 (#44194) * [CPUID] Rework how current ISA is determined * [CPUID] Add ISA entry for A64FX * [CPUID] Add ISA entry for Apple Silicon M1 * [CPUID] Simplify collection of full set of features for architecture * [CPUID] Remove AES from A64FX ISA, not all chips appear to have it | 20 February 2022, 16:17:43 UTC |
| 6936433 | Valentin Churavy | 20 February 2022, 15:13:30 UTC | Merge pull request #44270 from dnadlinger/debuginfo-fix-arm32 debuginfo: Fix build on 32-bit ARM | 20 February 2022, 15:13:30 UTC |
| b0a3130 | Valentin Churavy | 20 February 2022, 02:07:08 UTC | [LLVM][PM] Make LowerPTLS NewPM compatible | 20 February 2022, 14:55:30 UTC |
| 7ad142b | Valentin Churavy | 20 February 2022, 01:59:07 UTC | [LLVM][PM] Make MultiVersioning NewPM compatible | 20 February 2022, 14:55:30 UTC |
| a292df3 | Valentin Churavy | 20 February 2022, 01:14:23 UTC | [LLVM][PM] Make GCInvariantVerifier NewPM compatible | 20 February 2022, 14:55:30 UTC |
| e69fe9b | Valentin Churavy | 20 February 2022, 00:55:02 UTC | [LLVM][PM] Make LowerExcHandlers NewPM compatible | 20 February 2022, 14:55:30 UTC |
| e723d37 | Ian Butterworth | 20 February 2022, 14:40:11 UTC | Profile: Use grace period to avoid trailing signals from timer (#44268) | 20 February 2022, 14:40:11 UTC |
| 6409a8a | N5N3 | 20 February 2022, 10:20:01 UTC | Add more negative stride support to BLAS Level 1/2 functions (#42957) | 20 February 2022, 10:20:01 UTC |
| 94838dc | Dilum Aluthge | 20 February 2022, 09:31:41 UTC | CI: enable assertions (both Julia assertions and LLVM assertions) for the `llvmpasses` job (#44278) | 20 February 2022, 09:31:41 UTC |
| e454858 | Dilum Aluthge | 20 February 2022, 09:20:18 UTC | In the `sysimage.mk` Makefile, set the `JULIA_NUM_THREADS=1` environment variable when running the `generate_precompile.jl` script (#44281) | 20 February 2022, 09:20:18 UTC |
| 7615063 | pchintalapudi | 20 February 2022, 07:25:23 UTC | Make more passes NewPM compatible (#44277) * Make PropagateAddrspaces and RemoveAddrspaces NewPM compatible * Fix LateGCLower analysis invalidation * Fix missing return | 20 February 2022, 07:25:23 UTC |
| 2714b92 | Julian Samaroo | 20 February 2022, 04:52:38 UTC | Profile.Allocs: jl_raw_alloc_t.task is untyped pointer (#44269) | 20 February 2022, 04:52:38 UTC |
| 2816a6f | Shuhei Kadowaki | 20 February 2022, 04:16:59 UTC | effects: make `:terminates_globally` functional on recursive functions (#44210) As with loop, it's hard to prove termination of recursive call automatically. But with this commit we can manually specify termination at least. ```julia Base.@assume_effects :terminates_globally function recur_termination1(x) x == 1 && return 1 1 < x < 20 || throw("bad") return x * recur_termination1(x-1) end @test fully_eliminated() do recur_termination1(12) end Base.@assume_effects :terminates_globally function recur_termination2(x) x == 1 && return 1 1 < x < 20 || throw("bad") return _recur_termination2(x) end _recur_termination2(x) = x * recur_termination2(x-1) @test fully_eliminated() do recur_termination2(12) end ``` | 20 February 2022, 04:16:59 UTC |
| ad52223 | pchintalapudi | 20 February 2022, 02:23:28 UTC | Make AllocOpt NewPM compatible (#44273) | 20 February 2022, 02:23:28 UTC |
| d849ed7 | Shuhei Kadowaki | 20 February 2022, 02:05:33 UTC | fix EscapeAnalysis.jl documentation (#44257) | 20 February 2022, 02:05:33 UTC |
| c9ac2ea | David Nadlinger | 19 February 2022, 20:31:41 UTC | debuginfo: Fix build on 32-bit ARM This slipped through in 955d4271, as we aren't building for 32 bit ARM during CI right now. GitHub: Fixes #44254. | 20 February 2022, 00:15:44 UTC |
| f3e3554 | Valentin Churavy | 19 February 2022, 23:33:10 UTC | Merge pull request #44245 from JuliaLang/vc/llvm_14 Fix build against LLVM 14 | 19 February 2022, 23:33:10 UTC |
| 627016a | Viral B. Shah | 19 February 2022, 22:13:58 UTC | Update LBT to 5.0.1 for source build (#44258) | 19 February 2022, 22:13:58 UTC |
| 623ceb7 | Elliot Saba | 19 February 2022, 17:47:41 UTC | Ensure that `open(::Function, ::Cmd)` waits for termination (#44078) On Windows, we observed occasional issues where an error within the function callback to the `open(::Function, ::Cmd)` method would cause problems due to assuming that the opened process had finished by the time the `open()` call was finished. In most cases this was true, however on Windows, it was found that we need to explicitly `wait()` upon the process object to ensure that all file handles held by the subprocess were properly closed by the time `open()` is finished. Co-authored-by: Dilum Aluthge <dilum@aluthge.com> | 19 February 2022, 17:47:41 UTC |
| 928f63c | Seth Axen | 19 February 2022, 17:46:55 UTC | Fix dot(::Adjoint, ::Adjoint) for numbers that don't commute under multiplication (#44219) Co-authored-by: Fredrik Ekre <ekrefredrik@gmail.com> | 19 February 2022, 17:46:55 UTC |
| 4061e8f | Jeff Bezanson | 19 February 2022, 17:45:05 UTC | fix #44239, regression in keyword args in getindex (#44246) | 19 February 2022, 17:45:05 UTC |
| 749a658 | Jakob Nybo Nissen | 19 February 2022, 15:43:33 UTC | Add docstring to `String` (#43943) The type `String` itself did not have a docstring, instead the type is documented through two of its constructors. Add docstring to `String`, and remove redundant information from the constructor docstrings. In particular, mention that `String`s, while they are interpreted as being UTF8 encoded, do not assume strings are valid UTF8, and may be arbitrary byte sequences. | 19 February 2022, 15:43:33 UTC |
| 5bd0545 | Max Horn | 19 February 2022, 15:41:17 UTC | Fix get/set_fpcr_aarch64 (#44256) On Aarch64, the `fpcr` register is 64bit wide, although the top 32bit are currently unused and reserved for future usage. Nevertheless, we should safe and restore the full 64 bit, not just 32 bit. This also silences a compiler warning about this. Reference: <https://developer.arm.com/documentation/ddi0595/2021-06/AArch64-Registers/FPCR--Floating-point-Control-Register> | 19 February 2022, 15:41:17 UTC |
| f852c88 | Takafumi Arakaki | 19 February 2022, 15:31:44 UTC | Fix missing `jl_effective_threads` declaration (#44251) | 19 February 2022, 15:31:44 UTC |
| 700c69b | Valentin Churavy | 18 February 2022, 19:42:27 UTC | fix deprecated AlignedLoad | 19 February 2022, 15:29:13 UTC |
| 074d761 | Valentin Churavy | 18 February 2022, 19:37:41 UTC | fix some missing includes | 19 February 2022, 15:29:13 UTC |
| ed9851b | Valentin Churavy | 18 February 2022, 19:37:31 UTC | AttrBuilder now takes LLVMContext | 19 February 2022, 15:29:12 UTC |
| f544321 | Thomas Christensen | 19 February 2022, 08:49:10 UTC | Extend doc-string for `@deprecate` (#42892) | 19 February 2022, 08:49:10 UTC |
| 3453775 | Takafumi Arakaki | 19 February 2022, 03:24:22 UTC | Use uv_thread_getaffinity when --threads=auto (#42340) Co-authored-by: Jameson Nash <vtjnash@gmail.com> | 19 February 2022, 03:24:22 UTC |
| 482b04c | Keno Fischer | 19 February 2022, 02:53:17 UTC | Elaborate what egal means (#44248) Several people were confused by what `egal` means here. Put `===` in parentheses so people know what is meant. | 19 February 2022, 02:53:17 UTC |
| f5d9b86 | Jeff Bezanson | 18 February 2022, 21:32:57 UTC | fix bug in `addenv` for environment entries with embedded `=` (#44212) Co-authored-by: Jameson Nash <vtjnash@gmail.com> Co-authored-by: Fredrik Ekre <ekrefredrik@gmail.com> | 18 February 2022, 21:32:57 UTC |
| 15dcd5b | Nathan Daly | 18 February 2022, 21:25:34 UTC | Fix build warning in gc-alloc-profiler.cpp: cast to (size_t) (#44180) `std::vector::size()` returns a `size_t`, so we need to cast the int `jl_n_threads` to a `size_t` for the comparison. | 18 February 2022, 21:25:34 UTC |
| 923fe92 | Abel Siqueira | 18 February 2022, 20:45:07 UTC | Fix CITATION.cff (#44236) | 18 February 2022, 20:45:07 UTC |
| 2fe57fd | Valentin Churavy | 18 February 2022, 19:58:13 UTC | Merge pull request #42463 from JuliaLang/vc/newpm2 [LLVM] Convert more passes to NewPM | 18 February 2022, 19:58:13 UTC |
| 6c51d9e | Simeon Schaub | 18 February 2022, 18:04:07 UTC | fix nothrow check for `get_binding_type` (#44229) This got missed in #43671. | 18 February 2022, 18:04:07 UTC |
| 39e849a | Simeon Schaub | 18 February 2022, 18:03:35 UTC | set type of statement when processing `GlobalRef`s (#44200) Co-authored-by: Shuhei Kadowaki <40514306+aviatesk@users.noreply.github.com> Co-authored-by: Jameson Nash <vtjnash@gmail.com> | 18 February 2022, 18:03:35 UTC |
| daa0849 | Shuhei Kadowaki | 18 February 2022, 14:05:27 UTC | `AbstractInterpreter`: make it easier to overload pure/concrete-eval (#44224) fix #44174. | 18 February 2022, 14:05:27 UTC |
| b8e5d7e | Martin Holters | 18 February 2022, 08:47:33 UTC | Reduce number of `getindex(::Type, ...)` methods (#44127) Previously, there were special cases for `T[]`, `T[a]`, `T[a,b]` and `T[a,b,c]`. Together with the general case for more elements, that meant five methods to consider in cases like `T[x...]` where the length of `x` was not known at compile time. That was beyond the inference limit and such a call would be inferred as `Any`. So this change gets rid of all the special cases. The loop-based general case worked well if all arguments were of the same type, but otherwise suffered from type-instability inside the loop. Without the special cases for low element count this would be hit more often, so for the non-homogeneous case, the loop is replaced with a call to `afoldl` that basically unrolls the loop for up to 32 elements. | 18 February 2022, 08:47:33 UTC |
| ea3d788 | Valentin Churavy | 18 February 2022, 04:09:10 UTC | [LLVM][PM] Make FinalLowerGC NewPM compatible | 18 February 2022, 04:09:10 UTC |
| 7600bfd | Valentin Churavy | 17 February 2022, 19:17:54 UTC | [LLVM][PM] Make JuliaLICM NewPM compatible | 18 February 2022, 03:22:21 UTC |
| 8db9654 | Valentin Churavy | 02 October 2021, 22:19:33 UTC | [LLVM][PM] Make LateLowerGC NewPM compatible | 18 February 2022, 03:22:21 UTC |
| 1ad2396 | Thomas Christensen | 17 February 2022, 21:31:22 UTC | make `cat(As..., dims=Val((1,2,...))` work (#44211) | 17 February 2022, 21:31:22 UTC |
| 4abb47f | adkabo | 17 February 2022, 21:13:45 UTC | Add AbstractString to strings documentation (#43935) | 17 February 2022, 21:13:45 UTC |
| 17379e0 | Raymond Huffman | 17 February 2022, 20:00:18 UTC | Fixes to compile=all (#44213) - adds a null check to `jl_get_unspecialized` - prevents an infinite loop in `jl_compile_all_defs` | 17 February 2022, 20:00:18 UTC |
| 68bae54 | Valentin Churavy | 17 February 2022, 17:57:28 UTC | Merge pull request #44214 from JuliaLang/vc/new_pm_plugin [NewPM] add pass plugin | 17 February 2022, 17:57:28 UTC |
| 072c041 | Ian Butterworth | 17 February 2022, 17:52:29 UTC | Profile: Minor fixes. Signal handling fix. (#44199) | 17 February 2022, 17:52:29 UTC |
| 105c77a | Ian Butterworth | 17 February 2022, 17:46:29 UTC | Allow more time for profile print when Sockets and Threads watchdogs request info (#44218) | 17 February 2022, 17:46:29 UTC |
| 7b39515 | Kristoffer Carlsson | 17 February 2022, 07:33:22 UTC | Set VERSION to 1.9.0-DEV, move 1.8 NEWS to HISTORY, update SPDX (#44203) | 17 February 2022, 07:33:22 UTC |
| 7a1c20e | Johnny Chen | 17 February 2022, 03:58:34 UTC | Eagerly do boundscheck when indexing CartesianIndices with CartesianIndices (#42235) | 17 February 2022, 03:58:34 UTC |
| 2aaa25a | DilumAluthgeBot | 17 February 2022, 03:51:39 UTC | 🤖 Bump the Pkg stdlib from b4da4946 to e31a3dc7 (#44208) Co-authored-by: Dilum Aluthge <dilum@aluthge.com> | 17 February 2022, 03:51:39 UTC |
| 4930fc9 | Steven G. Johnson | 17 February 2022, 01:06:53 UTC | add reverse iteration for Cmd (#43498) | 17 February 2022, 01:06:53 UTC |
| 3887e01 | Valentin Churavy | 16 February 2022, 21:21:48 UTC | [NewPM] add pass plugin | 16 February 2022, 21:21:57 UTC |
| 7889b2a | Jeff Bezanson | 16 February 2022, 21:17:44 UTC | edit NEWS for 1.8 (#44209) | 16 February 2022, 21:17:44 UTC |
