e65c526 | Sascha Mann | 07 December 2020, 21:54:12 UTC | Iterator interface: Change i to item in example (#38740) Many iterators have an internal counter of iterations for the abort condition, e.g. `Squares` in the example. When implementing such an iterator while following this example, the naming can be a bit misleading because `i` in the example is _not_ a counter but instead the next item of the iteration. However, `i` is also commonly used as a counter in iterations. I think renaming it to `item` makes it a bit clearer. (cc @miguelraz) | 07 December 2020, 21:54:12 UTC |
b58e879 | ro-ble | 07 December 2020, 21:48:14 UTC | Hint on how to type the ÷ sign (#38720) The ÷ sign is the only operator in this chapter which has no dedicated key on the keyboard. With the Unicode input being at the very end of the documentation, newcomers probably do not know how to write the integer division sign. Therefore, I added this short hint on how to type it. | 07 December 2020, 21:48:14 UTC |
aee7ec1 | Benjamin Lorenz | 07 December 2020, 20:52:00 UTC | [libjulia] correct soname to use only major version number (#38748) The soname was changed in #38160 to be libjulia.so.1.6 instead of libjulia.so.1, but probably not on purpose. This contradicts what is written in the comment in Make.inc and it breaks e.g. libcxxwrap_julia_jll which is linked against libjulia.so.1. See https://github.com/JuliaPackaging/Yggdrasil/issues/2233 | 07 December 2020, 20:52:00 UTC |
18b7fee | Keno Fischer | 07 December 2020, 19:25:25 UTC | Allow a slightly wider class of constants to be inlined (#38732) Currently we only inline constants (i.e. replace Const() yielding statements by the constants themselves) if the constant in question is isbits. We do this to avoid blowing up the size of the IR by inlining constants that are too large. However, there are some useful small constants that this definition ignores. E.g. I was working with structs of the form: ``` struct Dims d::NTuple{N, Int} where N end ``` and the IR was littered with various pure statements that the optimizer wouldn't inline because this struct isn't considered isbits (side-note, I want to change that, but that's a separate issue). Here I'm proposing changing the criteria for a struct to be eligible for inlining slightly: Rather than just looking for isbits, we recurse through any immutable, isbits fields and as long as the size of all structs combined isn't too large, we allow the inlining. | 07 December 2020, 19:25:25 UTC |
5e048f3 | Keno Fischer | 07 December 2020, 19:24:48 UTC | Fix inference precision regression for functions returning PartialStruct (#38731) Consider the included test case: ``` f_partial_struct_constprop(a, b) = (a[1]+b[1], 2) g_partial_struct_constprop() = Val{f_partial_struct_constprop((1,), (1,))[1]}() ``` The function `f_partial_struct_constprop` used to infer to `Tuple{Int, Int}` be eligible for constprop and then get improved to `Const((2,2))`. However, at some point we sharpened the original inference to `PartialStrict((::Int, 2))`, which caused an early-out that prevented the consideration of this method for constant propagation purposes, regressing the overall inference result to `Tuple{Int, Int}`. Fix that by allowing `PartialStruct` to pass through the early out. I'm not seeing any impact on inference times. | 07 December 2020, 19:24:48 UTC |
057c316 | Valentin Churavy | 07 December 2020, 19:15:51 UTC | Merge pull request #38736 from JuliaLang/vc/sret [LLVM] Set type of StructRet attribute | 07 December 2020, 19:15:51 UTC |
048bde0 | Likan Zhan | 07 December 2020, 14:57:57 UTC | Add syntax highlighting to code examples in logging.jl (#36655) * Update logging.jl To make the highlighting style consistent: ``` ==> ```julia * Update base/logging.jl Co-authored-by: Viral B. Shah <ViralBShah@users.noreply.github.com> Co-authored-by: Matt Bauman <mbauman@gmail.com> | 07 December 2020, 14:57:57 UTC |
cf9bcc7 | Steven G. Johnson | 07 December 2020, 13:15:10 UTC | clarification to sincos and sincospi docs | 07 December 2020, 13:15:10 UTC |
d06c2a9 | Keno Fischer | 07 December 2020, 02:22:17 UTC | Try to avoid deadlock in Distributed test (#38714) What I believe is happening here is that, since put! is a yield point, the remote call finishes and the async task gets run before the `wait` statement is reached, thus causing a deadlock. I don't have 100% proof that this is the issue, but I did see the test hanging during this test, so I figured it's worth a try. May help #38712 (particularly, the second of the linked test failures). | 07 December 2020, 02:22:17 UTC |
9a0de1d | Keno Fischer | 07 December 2020, 02:21:46 UTC | Propagate more errors in Distributed test (#38717) | 07 December 2020, 02:21:46 UTC |
1cf5bd5 | Kristoffer Carlsson | 07 December 2020, 01:22:47 UTC | bump Pkg to fix hash failure in some artifacts (#38733) | 07 December 2020, 01:22:47 UTC |
1fe854e | Valentin Churavy | 07 December 2020, 00:11:12 UTC | Merge pull request #38738 from KlausC/krc/SHAfix fixed bug in SHA.jl | 07 December 2020, 00:11:12 UTC |
39a6c0d | Jameson Nash | 06 December 2020, 23:00:01 UTC | document correct usage for shell_escape_wincmd (#38513) Note that most resources online are wrong, and even `cmd /c help cmd` prints the wrong list, so it is important to be clear here about the actual guarantees this function can afford. Refs #38352 | 06 December 2020, 23:00:01 UTC |
e609b8b | Mustafa M | 06 December 2020, 21:53:05 UTC | win: allow key Inno variables to be set during compile time (#37241) | 06 December 2020, 21:53:05 UTC |
814a0d3 | Valentin Churavy | 06 December 2020, 00:49:14 UTC | set type of StructRet attribute | 06 December 2020, 20:42:02 UTC |
6eef7b6 | Valentin Churavy | 06 December 2020, 20:10:47 UTC | [Compiler] call error() instead of error(), since the later is not callable yet (#38708) | 06 December 2020, 20:10:47 UTC |
6371070 | C. Augusto Proiete | 06 December 2020, 19:16:43 UTC | Enable themes for Windows common controls and dialogs (#38705) | 06 December 2020, 19:16:43 UTC |
d5b8db5 | Valentin Churavy | 06 December 2020, 19:15:54 UTC | win32: implement enumeration of modules ourselves (#33062) | 06 December 2020, 19:15:54 UTC |
aa4f341 | Lyndon White | 06 December 2020, 18:56:08 UTC | Add missing news from #36146 (#37909) * Add missing news from #36146 * qualify DevNull * say devnull * Update NEWS.md Co-authored-by: Kristoffer Carlsson <kricarl@student.chalmers.se> Co-authored-by: Kristoffer Carlsson <kricarl@student.chalmers.se> Co-authored-by: Viral B. Shah <ViralBShah@users.noreply.github.com> | 06 December 2020, 18:56:08 UTC |
c649cbe | Klaus Crusius | 06 December 2020, 17:46:42 UTC | fixed bug in SHA.jl | 06 December 2020, 17:46:42 UTC |
9606ffd | William Dearden | 06 December 2020, 05:48:08 UTC | Update noteworthy differences from R due to R update (#27330) Co-authored-by: Milan Bouchet-Valat <nalimilan@club.fr> | 06 December 2020, 05:48:08 UTC |
65382c7 | Harmen Stoppels | 05 December 2020, 06:28:25 UTC | Do not use zero(...) in generic_lu! algorithm to support Unitful (#38659) * make generic_lu! accept unitful matrices * Add Furlong test for lu | 05 December 2020, 06:28:25 UTC |
8629e21 | Valentin Churavy | 04 December 2020, 23:10:16 UTC | Merge pull request #38707 from JuliaLang/vc/news update news [ci skip] | 04 December 2020, 23:10:16 UTC |
08ac65f | Valentin Churavy | 04 December 2020, 21:54:38 UTC | update news [ci skip] | 04 December 2020, 21:54:38 UTC |
2ced28b | Jeremie Knuesel | 04 December 2020, 20:28:14 UTC | Reword doc on undef arrays with missing (#38260) The intent is to make clear that undef constructors are not the right way to initialize arrays to missing. This way, other array implementations (such as SentinelArrays.jl) are free to provide faster undef constructors that actually leave data uninitialized (without creating an incompatiblity with specified behavior from Base). | 04 December 2020, 20:28:14 UTC |
dc38a37 | DilumAluthgeBot | 04 December 2020, 20:22:41 UTC | [automated] Bump the Downloads stdlib from 2f0d46f to 02e77ae (#38700) Co-authored-by: Dilum Aluthge <dilum@aluthge.com> | 04 December 2020, 20:22:41 UTC |
0253ebf | Valentin Churavy | 04 December 2020, 18:14:31 UTC | Merge pull request #38487 from JuliaLang/jn/lock-finalizers threads: avoid deadlock from recursive lock acquire | 04 December 2020, 18:14:31 UTC |
5948ee6 | Kristoffer Carlsson | 04 December 2020, 17:37:46 UTC | bump Pkg to latest set of 1.6 backports (#38678) Also bump Tar: its tests were broken by this change. Co-authored-by: Stefan Karpinski <stefan@karpinski.org> | 04 December 2020, 17:37:46 UTC |
f90cbe0 | Jacob Quinn | 04 December 2020, 17:29:18 UTC | Merge pull request #38699 from goretkin/datetime-doc-wording `Dates` documentation wording | 04 December 2020, 17:29:18 UTC |
e6696a1 | Gustavo Goretkin | 04 December 2020, 17:06:20 UTC | Avoid first-person pronouns, except for in the FAQ | 04 December 2020, 17:06:20 UTC |
650bb37 | Gustavo Goretkin | 04 December 2020, 16:54:44 UTC | Simplify wording This use of "measure" is "an unspecified portion or quantity" (`wiktionary.com`) or "an adequate or due portion" or "a moderate degree" (`merriam-webster.com`). In the context of technical writing about quantities of time, this idiomatic usage might give pause (even to a native English user). | 04 December 2020, 16:54:44 UTC |
56dd7d7 | Tim Besard | 04 December 2020, 13:18:37 UTC | Merge pull request #38679 from JuliaLang/tb/excstack_without_handler Report the top of the exception stack when there's no handler. | 04 December 2020, 13:18:37 UTC |
9e6a4a5 | Daniel Karrasch | 04 December 2020, 11:15:12 UTC | Fix & extend LAPACK docs (#38442) | 04 December 2020, 11:15:12 UTC |
aaab77e | Tim Besard | 03 December 2020, 14:32:40 UTC | Report the top of the exception stack when there's no handler. | 04 December 2020, 10:34:11 UTC |
63c8e47 | Keno Fischer | 04 December 2020, 09:05:12 UTC | Run Distributed tests on node 1 (#38694) We're still seeing fairly regular timeouts in the Distributed test. I'd like to try moving it into the node1 section, like we do with e.g. the threads test that spawns lots of threads. Particularly our small buildbots can get a bit overloaded with all the other tests, so I'm thinking the timeout might be genuine. If it is, then this will fix it. If we still see timeouts, we have some other underlying problem. | 04 December 2020, 09:05:12 UTC |
9755fd4 | Elliot Saba | 04 December 2020, 05:17:14 UTC | Add empty `get-csl` target (#38693) | 04 December 2020, 05:17:14 UTC |
88f445a | Jeff Bezanson | 04 December 2020, 05:16:37 UTC | fix #38501, wrap interpolated literal strings in `:string` exprs (#38692) | 04 December 2020, 05:16:37 UTC |
b78e505 | Steven G. Johnson | 04 December 2020, 04:19:20 UTC | add missing link | 04 December 2020, 04:19:20 UTC |
8dd959e | Simon Byrne | 04 December 2020, 01:14:18 UTC | Fix overflow for hypot in Float16 (#38322) * Fix overflow for hypot in Float16 Fixes #38311 * hypot tests for ax > floatmax(ax)/2 * hypot Float16 promote to Float32 Co-authored-by: achuchmala <ach@spyro-soft.com> | 04 December 2020, 01:14:18 UTC |
0db72ee | DilumAluthgeBot | 04 December 2020, 00:20:27 UTC | [automated] Bump the NetworkOptions stdlib from 462628a to 3d8a45c (#38684) Co-authored-by: Dilum Aluthge <dilum@aluthge.com> | 04 December 2020, 00:20:27 UTC |
2aaf9ea | Carsten Bauer | 03 December 2020, 23:51:43 UTC | Sticky shell REPL mode (#38646) * make shell REPL mode sticky * fix tests * fix precompile statement generation script | 03 December 2020, 23:51:43 UTC |
9631a9f | Jameson Nash | 03 December 2020, 23:51:10 UTC | add __CET__ check guards to trampoline assembly (#38683) Fixes #38680 | 03 December 2020, 23:51:10 UTC |
7d5c0c5 | Jeff Bezanson | 03 December 2020, 22:42:50 UTC | fix part of #38650, `struct` should be a hard scope (#38663) | 03 December 2020, 22:42:50 UTC |
8ffcc0e | Elliot Saba | 03 December 2020, 21:24:20 UTC | Fix `stdlib/Makefile` rules for JLLs (#38688) These had gotten out of sync during rebases, this should clean things up | 03 December 2020, 21:24:20 UTC |
fe1253e | Jeff Bezanson | 03 December 2020, 21:04:20 UTC | fix #38664, regression in `===` codegen for Bool (#38686) | 03 December 2020, 21:04:20 UTC |
72bebdf | Jameson Nash | 30 November 2020, 23:28:22 UTC | move timing_stack field from jl_task_t onto the stack also | 03 December 2020, 18:53:39 UTC |
59aedd1 | Jameson Nash | 18 November 2020, 19:04:50 UTC | threads: avoid deadlock from recursive lock acquire Finalizers can't safely acquire many essential locks (such as the iolock, to cleanup libuv objects) if they are run inside another lock. Therefore, inhibit all finalizers on the thread until all locks are released (previously, this was only true for our internal locks). | 03 December 2020, 18:53:39 UTC |
49b8e61 | Tim Besard | 03 December 2020, 17:00:20 UTC | Fix out-of-tree compilation of loader library. (#38677) | 03 December 2020, 17:00:20 UTC |
7193a07 | Kristoffer Carlsson | 03 December 2020, 16:59:21 UTC | add missing checksums for utf8proc and libwhich (#38674) | 03 December 2020, 16:59:21 UTC |
e7504fc | DilumAluthgeBot | 03 December 2020, 16:57:14 UTC | [automated] Bump the Downloads stdlib from 98922d7 to 2f0d46f (#38672) Co-authored-by: Dilum Aluthge <dilum@aluthge.com> | 03 December 2020, 16:57:14 UTC |
7c17bb3 | Logan Kilpatrick | 03 December 2020, 15:55:45 UTC | Add examples for endswith and startswith (#38255) | 03 December 2020, 15:55:45 UTC |
f309d27 | strickek | 03 December 2020, 15:55:21 UTC | Update docview.jl - printmatch() (#38500) printfuzzy does not exist, delete it | 03 December 2020, 15:55:21 UTC |
96a45ae | Kristoffer Carlsson | 03 December 2020, 12:04:51 UTC | hook up Artifacts to the TOML cache in Base (#38658) | 03 December 2020, 12:04:51 UTC |
e150b86 | Dilum Aluthge | 03 December 2020, 08:21:19 UTC | Document how to build Julia from source with Git checkouts of one or more stdlibs (#38671) | 03 December 2020, 08:21:19 UTC |
9528ac2 | Kristoffer Carlsson | 02 December 2020, 23:57:51 UTC | bump Pkg to latest set of 1.6 backports (#38655) | 02 December 2020, 23:57:51 UTC |
2e3364e | Elliot Saba | 02 December 2020, 23:56:32 UTC | [loader]: Re-export symbols for C embedding, rename to `libjulia-internal` (#38160) | 02 December 2020, 23:56:32 UTC |
5b6a081 | Jeff Bezanson | 02 December 2020, 23:25:11 UTC | reuse anon functions in channels test to reduce compilation time (#38662) | 02 December 2020, 23:25:11 UTC |
97bdd8b | Shuhei Kadowaki | 02 December 2020, 18:34:36 UTC | define at-invokelatest macro (#37971) - `@invokelatest f(args...; kwargs...)` will simply be expanded into `Base.invokelatest(f, args...; kwargs...)` | 02 December 2020, 18:34:36 UTC |
319ae4a | Steven G. Johnson | 02 December 2020, 17:17:42 UTC | better sub/superscript tab completion (#38649) * better sub/superscript tab completion * Update stdlib/REPL/src/REPLCompletions.jl Co-authored-by: Jeff Bezanson <jeff.bezanson@gmail.com> Co-authored-by: Jeff Bezanson <jeff.bezanson@gmail.com> | 02 December 2020, 17:17:42 UTC |
a923e6b | Andrzej Chuchmała | 02 December 2020, 17:02:40 UTC | Tests for multiplication of sparse triangular and dense matrices (#38505) | 02 December 2020, 17:02:40 UTC |
3be99c5 | Tim Holy | 02 December 2020, 16:40:20 UTC | Fix Markdown word wrap (#38502) * Revert "fix markdown rendering (#37235)" This reverts commit 36505aa1542fe84fcfefccfc80317b37470f8be3. * Fix Markdown word-wrap (fixes #38275) This changes one test, but IMO the extra spaces in the test result are not actually desirable. The alignment on Julia 1.5 is also off-by-one, whereas this seems well-aligned. | 02 December 2020, 16:40:20 UTC |
8c7464e | DilumAluthgeBot | 02 December 2020, 15:19:00 UTC | [automated] Bump the Tar stdlib to 74a2e675fea0a170f1aee6b547fdcd43d3f567b4 (#38614) Co-authored-by: Dilum Aluthge <dilum@aluthge.com> | 02 December 2020, 15:19:00 UTC |
a4465b8 | DilumAluthgeBot | 02 December 2020, 15:17:54 UTC | [automated] Bump the Statistics stdlib to 4b3ef9aaa79350510ca0be395458f66051c2f92d (#38613) Co-authored-by: Dilum Aluthge <dilum@aluthge.com> | 02 December 2020, 15:17:54 UTC |
c477b9b | DilumAluthgeBot | 02 December 2020, 15:16:19 UTC | [automated] Bump the Downloads stdlib to 98922d7b3526c73094671201fc36d591c022d6c2 (#38610) Co-authored-by: Dilum Aluthge <dilum@aluthge.com> | 02 December 2020, 15:16:19 UTC |
c2eeb27 | DilumAluthgeBot | 02 December 2020, 15:15:34 UTC | [automated] Bump the NetworkOptions stdlib to 462628a2790a1d1b531b9e29517887a39462a524 (#38612) Co-authored-by: Dilum Aluthge <dilum@aluthge.com> | 02 December 2020, 15:15:34 UTC |
3073964 | Valentin Churavy | 02 December 2020, 14:10:40 UTC | Merge pull request #38651 from DilumAluthgeBot/BumpStdlibs/LibCURL-8310487 🤖 Bump the LibCURL stdlib to 8310487 | 02 December 2020, 14:10:40 UTC |
63ca518 | Elliot Saba | 02 December 2020, 06:25:46 UTC | build-stats: install `llvm-tools` before trying to run if using BB (#38652) Now that we've split up the LLVM JLLs a bit, ensure that `llvm-tools` is installed if we're using BB LLVM and we try to run `make build-stats` | 02 December 2020, 06:25:46 UTC |
2eccd8e | Elliot Saba | 02 December 2020, 03:34:09 UTC | Fake JLLs for all packages (#38347) | 02 December 2020, 03:34:09 UTC |
ae2da90 | Dilum Aluthge | 02 December 2020, 01:31:39 UTC | 🤖 Bump the LibCURL stdlib to 8310487 | 02 December 2020, 01:31:39 UTC |
a178ac2 | jmert | 01 December 2020, 22:47:54 UTC | Add news for cispi (#38647) [ci skip] | 01 December 2020, 22:47:54 UTC |
f89a83a | Jerry Ling | 01 December 2020, 20:01:25 UTC | fix 1d/2d array method error message with unsupported keyword arg (#36325) Co-authored-by: Jeff Bezanson <jeff.bezanson@gmail.com> | 01 December 2020, 20:01:25 UTC |
626542e | Steven G. Johnson | 01 December 2020, 18:23:27 UTC | titlecase should not break words inside graphemes (#38575) | 01 December 2020, 18:23:27 UTC |
916762a | Nathan Boyer | 01 December 2020, 15:05:30 UTC | doc: add punctuation for pipes (|>) and composition (∘) (#37533) | 01 December 2020, 15:05:30 UTC |
a69ecaa | Shuhei Kadowaki | 01 December 2020, 14:49:48 UTC | improve inferrability of `list_deletefirst!` (#38236) | 01 December 2020, 14:49:48 UTC |
8c35003 | Rafael Fourquet | 01 December 2020, 12:25:53 UTC | update at-testset docs for verbose option (#38447) | 01 December 2020, 12:25:53 UTC |
c27d5e9 | Elliot Saba | 01 December 2020, 09:58:51 UTC | [cmdlineargs] Limit total number of threads in oversubscription test (#38633) This test was exhausting our limit of 1000 PIDs because the machine it was running on has 128 cores, which ends up trying to launch a Julia process with 1280 threads. | 01 December 2020, 09:58:51 UTC |
b1adb5d | Bogumił Kamiński | 01 December 2020, 08:34:01 UTC | doc: update docstring of getindex (#38537) The current docstring suggests that only a limited range of `inds` are allowed, which is not the case. | 01 December 2020, 08:34:01 UTC |
2e49e30 | Tim Besard | 01 December 2020, 08:17:10 UTC | Merge pull request #38619 from JuliaLang/tb/D92210 Don't sink ptrtoint/inttoptr sequences into non-noop addrspacecasts. | 01 December 2020, 08:17:10 UTC |
812cea8 | Tim Besard | 30 November 2020, 20:48:46 UTC | Bump BinaryBuilder LLVM artifacts. | 01 December 2020, 06:55:49 UTC |
59d676c | Steven G. Johnson | 01 December 2020, 06:00:36 UTC | streamline a[begin] lowering via firstindex(a, n) (#35779) | 01 December 2020, 06:00:36 UTC |
8ea6365 | Shuhei Kadowaki | 30 November 2020, 23:50:01 UTC | improve inferrability of `seek(::Stateful, sentinel)` (#38625) the current inference can't propagate constant values or `isa` condition of object fields, and this helps inference eliminate the possibility of `getindex(s.nextvalstate::Nothing, 1)` call | 30 November 2020, 23:50:01 UTC |
bc0b9a7 | Gunnar Farnebäck | 30 November 2020, 22:52:11 UTC | Properly handle move up zero lines in TerminalMenus. (#38630) | 30 November 2020, 22:52:11 UTC |
7711fd5 | Jeff Bezanson | 30 November 2020, 21:58:30 UTC | Merge pull request #38603 from JuliaLang/jn/38547 Profile: change interval to exclude time taken for profile itself | 30 November 2020, 21:58:30 UTC |
056be22 | Sheehan Olver | 30 November 2020, 21:55:27 UTC | Modify documentation for `view` to take into account range special case (#38536) | 30 November 2020, 21:55:27 UTC |
687378e | Simon Byrne | 30 November 2020, 21:48:45 UTC | allow default_addprocs_params to be specialized on ClusterManager (#38570) I made this change to allow the set to be expanded for my own package, but I noticed this also helps unify #38353 and existing ssh-only options. | 30 November 2020, 21:48:45 UTC |
87bfa51 | Jameson Nash | 30 November 2020, 21:28:08 UTC | show: remove dead displaysize call from show_ir (#38615) This has not been relevant to creating a string from IR in a couple years. | 30 November 2020, 21:28:08 UTC |
c31d126 | Sebastian Stock | 30 November 2020, 20:44:55 UTC | Disallow ranges with Integer eltype but non-integer step (#32439) | 30 November 2020, 20:44:55 UTC |
217a449 | Simon Byrne | 30 November 2020, 20:18:12 UTC | Use mpq functions for Rational{BigInt} (#38520) Fixes #9826. | 30 November 2020, 20:18:12 UTC |
3ffb4fb | Valentin Churavy | 30 November 2020, 18:55:36 UTC | Merge pull request #38626 from JuliaLang/vc/validate add aliasscope/popaliasscope to IR validator | 30 November 2020, 18:55:36 UTC |
1ea7dc5 | Valentin Churavy | 30 November 2020, 17:04:41 UTC | add aliasscope/popaliasscope to IR validator | 30 November 2020, 17:04:41 UTC |
00e7026 | DilumAluthgeBot | 30 November 2020, 16:28:47 UTC | [automated] Bump the LibCURL stdlib to 54626fe52d2882332fa98223e034a43768db22a4 (#38611) Co-authored-by: Dilum Aluthge <dilum@aluthge.com> | 30 November 2020, 16:28:47 UTC |
a35a46e | DilumAluthgeBot | 30 November 2020, 16:26:41 UTC | [automated] Bump the ArgTools stdlib to fa878696ff2ae4ba7ca9942bf9544556c0d86ce4 (#38609) Co-authored-by: Dilum Aluthge <dilum@aluthge.com> | 30 November 2020, 16:26:41 UTC |
bc94c50 | Morten Piibeleht | 30 November 2020, 13:05:21 UTC | doc: add compat for mktemp and mktempdir docstrings (#38452) | 30 November 2020, 13:05:21 UTC |
95d5fba | Lyndon White | 30 November 2020, 13:00:27 UTC | doc: add missing link in NEWS.md (#37908) | 30 November 2020, 13:00:27 UTC |
059ea24 | Fredrik Ekre | 30 November 2020, 12:49:32 UTC | Fix code formatting in addprocs docstring, fixes #38554. (#38620) | 30 November 2020, 12:49:32 UTC |
3074306 | Antonello Lobianco | 30 November 2020, 10:30:52 UTC | Update NEWS.md (#38607) Minor typo (the the default active project is now 1.6 not 1.5) | 30 November 2020, 10:30:52 UTC |
28232bf | Tim Besard | 30 November 2020, 09:26:11 UTC | Don't sink ptrtoint/inttoptr sequences into non-noop addrspacecasts. This miscompilation causes multiple issues with CUDA.jl, e.g., https://github.com/JuliaGPU/CUDA.jl/issues/558 Patch has been submitted upstream: https://reviews.llvm.org/D92210 | 30 November 2020, 09:26:11 UTC |
51d5fcf | Jameson Nash | 30 November 2020, 00:12:05 UTC | Profile: add check against bad stack walk by libunwind Seems to be happening especially often on linux32 (which may be missing unwind info in BinaryBuilder.jl for libraries such as LLVM). This can manifest as a stackwalk taking an excessive amount of time (since it becomes an infinite recursion). Additional piece of closing #38547 | 30 November 2020, 00:13:00 UTC |
fab29dd | Jameson Nash | 29 November 2020, 03:05:20 UTC | Profile: change interval to exclude time taken for profile itself Similar to current macOS and Windows operation May close #38547 | 30 November 2020, 00:13:00 UTC |
fedf447 | Tim Holy | 29 November 2020, 22:31:38 UTC | Support Core.Compiler.InterpreterIP in building stacktraces (#38566) The new inference timing captures a backtrace for each descent into inference as a mechanism to find out who it was triggered by. This allows us to convert those backtraces to stacktraces. | 29 November 2020, 22:31:38 UTC |