5f013d8 | Olivier MATHIEU | 10 October 2019, 22:28:33 UTC | Add takewhile, dropwhile to iterators (#33437) | 10 October 2019, 22:28:33 UTC |
1ed2e1d | Huang, Zhaoquan | 10 October 2019, 21:31:46 UTC | Update dtypes.h to fix C4146 in Visual C++ (#33511) | 10 October 2019, 21:31:46 UTC |
99483b5 | Jeff Bezanson | 10 October 2019, 21:31:16 UTC | improve docs for `SpinLock` to steer people away from it (#33514) Inspired by e.g. #33507 | 10 October 2019, 21:31:16 UTC |
b300471 | Fredrik Ekre | 10 October 2019, 17:40:19 UTC | Use default community health files from JuliaLang/.github, see (#33522) https://help.github.com/en/articles/creating-a-default-community-health-file-for-your-organization | 10 October 2019, 17:40:19 UTC |
5c1dca4 | Mustafa M | 10 October 2019, 15:41:39 UTC | Don't eagerly convert isabspath to String (#33491) | 10 October 2019, 15:41:39 UTC |
2f7f0bf | Alexander Seiler | 10 October 2019, 14:56:45 UTC | Fix multiplication of Triangular and Diagonal matrix (#33334) * Fix multiplication of Triangular and Diagonal matrix * Adopt more multiplication methods * Add `similar` methods for `Adjoint` and `Transpose` * Enable type tests | 10 October 2019, 14:56:45 UTC |
581f81c | Kristoffer Carlsson | 10 October 2019, 09:45:31 UTC | use simpler copy algorithm for small BitArrays in broadcasting (#33517) | 10 October 2019, 09:45:31 UTC |
09795cd | Jacob Quinn | 09 October 2019, 22:33:56 UTC | Merge pull request #33483 from artemsolod/time_mod_arithm modulo when constructing Time values | 09 October 2019, 22:33:56 UTC |
8e155c7 | Morten Piibeleht | 09 October 2019, 21:54:27 UTC | Fix attaching docstrings to where call syntax (#33446) * Allow documenting where-expressions I.e. docstrings can now also be attached to expressions that look as follows: """ """ foo(::T) where T Note that it was already fine to attach docstrings to such methods when the docstring was attached to the function body: """ """ function foo(::T) where T ... end * Better fix * Add tests | 09 October 2019, 21:54:27 UTC |
b75fce2 | lhnguyen-vn | 09 October 2019, 19:29:00 UTC | Fixed printing for Complex{Int} with maximum negative imaginary part (Issue #33433) (#33498) | 09 October 2019, 19:28:59 UTC |
e0bdc76 | Tim Besard | 09 October 2019, 17:01:09 UTC | Add safepoint annotations to GC callbacks. (#33508) | 09 October 2019, 17:01:09 UTC |
d913db4 | Jeff Bezanson | 09 October 2019, 04:26:29 UTC | fix #33367, defining enums is slow (#33494) Each definition used a new closure and call to `ntuple`, which is not necessary since the code only runs at load time. | 09 October 2019, 04:26:29 UTC |
80dca30 | Matt Bauman | 08 October 2019, 19:36:51 UTC | Merge pull request #27457 from non-Jedi/reducedim_custom_eltype Test dimensional reduce with non-bitstype | 08 October 2019, 19:36:51 UTC |
f250511 | Adam Beckmeyer | 06 June 2018, 14:33:15 UTC | Add test for custom type dimensional reduction | 08 October 2019, 15:59:42 UTC |
76d9b4b | Mustafa M | 08 October 2019, 15:28:41 UTC | Fix return type docstring for realpath and homedir (#33492) | 08 October 2019, 15:28:41 UTC |
0bcceac | Mathieu Besançon | 08 October 2019, 15:28:26 UTC | matrix is an array (#33500) | 08 October 2019, 15:28:26 UTC |
353b6c0 | Matt Bauman | 08 October 2019, 14:13:19 UTC | Better documentation and docs search for punctuation (#33344) * Better documentation and docs search for punctuation * Update doc/src/base/punctuation.md Co-Authored-By: James Bradbury <jekbradbury@gmail.com> * Update doc/src/base/punctuation.md Co-Authored-By: James Bradbury <jekbradbury@gmail.com> * Fixup crosslinks working around https://github.com/JuliaDocs/Documenter.jl/issues/781 | 08 October 2019, 14:13:19 UTC |
14aaf76 | Kristoffer Carlsson | 08 October 2019, 07:59:44 UTC | Merge pull request #33493 from JuliaLang/sf/fix_7z_zlib_problems Fix 7z and zlib problems | 08 October 2019, 07:59:44 UTC |
5bf7a14 | Elliot Saba | 07 October 2019, 20:29:07 UTC | Add `$(EXE)` to the end of `7z` when installing it | 07 October 2019, 20:29:07 UTC |
7f1e190 | Elliot Saba | 07 October 2019, 20:28:53 UTC | Fix `zlib` <-> `libz` confusion within build system | 07 October 2019, 20:28:53 UTC |
d5d5718 | Jeff Bezanson | 07 October 2019, 19:33:06 UTC | add some type info to Base to avoid excess recursion in inference (#33476) Fixes #33336. This addresses some commonly-occurring cases where having too little type info makes inference see a lot of recursion in Base that is not actually possible. | 07 October 2019, 19:33:06 UTC |
b5f6fdc | Artem Solod | 06 October 2019, 21:43:51 UTC | modulo when constructing Time values | 06 October 2019, 21:43:51 UTC |
0aa59a0 | Jeff Bezanson | 05 October 2019, 06:45:22 UTC | fix #33466, serialization of `IdDict` (#33473) Can still read old data, but new files will not be readable by older versions. | 05 October 2019, 06:45:22 UTC |
cff873b | Keno Fischer | 05 October 2019, 06:29:31 UTC | Clean up printing of testset start (#33408) This has bothered me for a while, but looks like it didn't bother anyone else enough to clean it up, so here I go :). Printing when a test starts is useful when looking through logs because you can correlate errors on a given worker with the specific test it was running. However, it doesn't have to be so ugly and break the nice regular formatting of the rest of the test. Before: ``` Test (Worker) | Time (s) | GC (s) | GC % | Alloc (MB) | RSS (MB) From worker 3: running testset int... From worker 2: running testset math... int (3) | 9.51 | 0.21 | 2.2 | 507.51 | 232.42 math (2) | 81.06 | 0.75 | 0.9 | 3391.19 | 323.18 Test Summary: | Pass Total Overall | 1637318 1637318 SUCCESS ``` After: ``` Test (Worker) | Time (s) | GC (s) | GC % | Alloc (MB) | RSS (MB) int (3) | started at 2019-09-27T18:35:02.761 math (2) | started at 2019-09-27T18:35:02.883 int (3) | 11.91 | 0.28 | 2.4 | 723.85 | 278.07 math (2) | 40.51 | 0.87 | 2.1 | 3287.85 | 325.49 Test Summary: | Pass Total Overall | 2030534 2030534 SUCCESS ``` | 05 October 2019, 06:29:31 UTC |
792d3e9 | Keno Fischer | 05 October 2019, 06:23:50 UTC | Add some missing Makefile dependency links (#33355) Noticed that some files weren't being rebuilt when I modified the list of builtins, causing bad state and strange error messages. | 05 October 2019, 06:23:50 UTC |
7466d11 | Jeff Bezanson | 05 October 2019, 02:53:35 UTC | improve `instanceof_tfunc` to take declared parameter bounds into account (#33472) | 05 October 2019, 02:53:35 UTC |
0725e0e | Keno Fischer | 05 October 2019, 02:49:38 UTC | Merge pull request #33458 from JuliaLang/kf/converterror Slightly improved printing for convert errors | 05 October 2019, 02:49:38 UTC |
cf87c35 | Keno Fischer | 02 October 2019, 22:44:23 UTC | Slightly improved printing for convert errors We often get MethodErrors involving `convert` between slightly different types (e.g. where the programmer expected to have constructed the correct type, but made a slight mistake). This attempts to slightly improve the formatting of the error message to make differences more obvious. Of course significant work remains to make this error message more usable. | 04 October 2019, 21:05:11 UTC |
3b1b767 | Keno Fischer | 04 October 2019, 20:53:54 UTC | Merge pull request #33354 from JuliaLang/kf/baditerate Don't crash in _apply if the user writes a bad iterate method | 04 October 2019, 20:53:54 UTC |
9d98131 | Keno Fischer | 04 October 2019, 01:48:35 UTC | Merge pull request #33448 from JuliaLang/tb/gc_collection Support for fine-grained selection of the GC collection type. | 04 October 2019, 01:48:35 UTC |
3b1c1fd | Nathan Daly | 04 October 2019, 01:47:52 UTC | Restrict `signed(x) = reinterpret(...)` to BitIntegers. (#33459) This commit restricts the `reinterpret` implementation of `signed(x)` to only apply to `Base.BitInteger`s. This brings its behavior in-line with `unsigned(x)`. Previously, although `unsigned(x)` did restrict the `reinterpret` implementation to only apply to `Base.BitInteger`s, `signed(x)` by default called `reinterpret(Signed, x)` for any type `T <: Unsigned`. This of course wouldn't work for `BigInt`, which is presumably why `unsigned(x)` was restricted (since `BigInt <: Signed` but not `BigInt <: BitInteger`), but although _Base_ contains no non-BitInteger types in `Unsigned` (besides `Bool`), of course users _could_ create such types, so this should be restricted symmetrically. Below is an example of the previous, unexpected broken behavior: ```julia struct WeirdUnsigned <: Unsigned msg::String val::UInt16 end WeirdUnsigned(x::Int) = WeirdUnsigned("missing", UInt16(x)) Base.Signed(x::WeirdUnsigned) = signed(x.val) ``` ```julia julia> signed(WeirdUnsigned(4)) ERROR: bitcast: expected primitive type value for second argument Stacktrace: [1] reinterpret at ./essentials.jl:417 [inlined] [2] signed(::WeirdUnsigned) at ./int.jl:162 [3] top-level scope at REPL[20]:1 ``` Whereas after this commit, this works as expected, calling the `Signed` constructor: ```julia julia> signed(WeirdUnsigned(4)) 4 julia> typeof(signed(WeirdUnsigned(4))) Int16 ``` ------------------- This should be a non-breaking change. If there exist any `primitive type` implementations that were relying on the reinterpret implementation, they may see a slight performance regression, as this will by default now call `Signed(x::T)`, instead of `reinterpret(typeof(Signed(x)), x)` -- so they may want to manually define the reinterpret implemention themselves. | 04 October 2019, 01:47:52 UTC |
75a1210 | Keno Fischer | 21 September 2019, 20:28:44 UTC | Don't crash in _apply if the user writes a bad iterate method Instead check that we indeed got a tuple of appropriate length back. | 03 October 2019, 22:17:32 UTC |
018d9ae | Alexandre Brilhante | 03 October 2019, 19:46:13 UTC | Add check=false for calculating determinant of singular (#32086) Symmetric/Hermitian matrices, fixes #32079. | 03 October 2019, 19:46:13 UTC |
979bf43 | Nick Robinson | 03 October 2019, 19:26:50 UTC | Improve `show` for 0-dim array (#33206) | 03 October 2019, 19:26:50 UTC |
bf8c324 | Tamas K. Papp | 03 October 2019, 19:07:19 UTC | Allows the results of `mapfoldr` to be inferred on tuples. (#33235) | 03 October 2019, 19:07:19 UTC |
60a60a5 | Matt Bauman | 03 October 2019, 16:09:21 UTC | Better help answers for strange inputs (#33444) Before: ```julia help?> `` ERROR: LoadError: Invalid at-var syntax `Core.at-cmd`. help?> {a ERROR: LoadError: Invalid at-var syntax `incomplete: premature end of input`. help?> 123456789012345678901234567890 ERROR: LoadError: Invalid at-var syntax `Core.at-int128_str`. ``` After: ```julia help?> `` at-cmd str ... help?> {a No documentation found. help?> 123456789012345678901234567890 at-int128_str str ... ``` | 03 October 2019, 16:09:21 UTC |
1739ca0 | Nathan Daly | 03 October 2019, 14:51:56 UTC | Export jl_expand_and_resolve (#32902) Instead of exporting the internal function jl_resolve_globals_in_ir, this factors out the code from generated functions into `jl_expand_and_resolve`, which lowers an expression into a CodeInfo and then resolves globals in the codeinfo in light of provided typenames. | 03 October 2019, 14:51:56 UTC |
888d32a | Jeff Bezanson | 03 October 2019, 08:42:27 UTC | fix #33370, avoid collisions between gensyms and anon function names (#33426) | 03 October 2019, 08:42:27 UTC |
b685cf6 | Elliot Saba | 03 October 2019, 08:35:11 UTC | Update GMP and MPFR builds to use better compiler flags (#33096) | 03 October 2019, 08:35:11 UTC |
c3d4a32 | Tim Besard | 03 October 2019, 05:55:57 UTC | Add a Julia interface. | 03 October 2019, 05:56:08 UTC |
d378da8 | Tim Besard | 02 October 2019, 07:24:41 UTC | Support for fine-grained selection of the GC collection type. This makes it possible to force an incremental collection, ignoring heuristics. | 03 October 2019, 05:56:08 UTC |
c72a295 | Tim Besard | 03 October 2019, 05:11:14 UTC | Merge pull request #33450 from JuliaLang/tb/gc_safe_printf Use jl_safe_printf for GC debug output. | 03 October 2019, 05:11:14 UTC |
a440ae7 | Simon Byrne | 02 October 2019, 22:30:45 UTC | Ryu: use UInt128 arithmetic instead of rolling own (#33441) | 02 October 2019, 22:30:45 UTC |
d911a9f | Jeff Bezanson | 02 October 2019, 22:04:04 UTC | better fix for #33337; revert #33353 (#33440) | 02 October 2019, 22:04:04 UTC |
3f37178 | Jeff Bezanson | 02 October 2019, 21:56:05 UTC | Merge pull request #33380 from JuliaLang/cjf/bt-lookup-offset Clean up return->call address translation for backtraces | 02 October 2019, 21:56:05 UTC |
fd473cc | Jeff Bezanson | 02 October 2019, 20:14:27 UTC | improve some keyword arg lowering error messages. fixes #33250 (#33442) | 02 October 2019, 20:14:27 UTC |
4de3ce3 | Jeff Bezanson | 02 October 2019, 19:35:04 UTC | fix #33155, add static parameters to at-locals (#33443) | 02 October 2019, 19:35:04 UTC |
b6991f2 | Mathieu Besançon | 02 October 2019, 14:48:04 UTC | docstring setindex (#32738) * docstring setindex * add bound check * fix selectdim implementation | 02 October 2019, 14:48:04 UTC |
d9ede17 | Tim Besard | 02 October 2019, 14:37:15 UTC | Use jl_safe_printf for all GC debug output. | 02 October 2019, 14:37:15 UTC |
1de5628 | Tim Besard | 02 October 2019, 14:02:52 UTC | Use jl_safe_printf for GC_TIME output. | 02 October 2019, 14:02:52 UTC |
32e3c9e | Fredrik Ekre | 02 October 2019, 12:28:27 UTC | Revert "Revert "Bump Pkg to 1.4-DEV. (#33364)" (#33410)" (#33434) This reverts commit 0b55ce12bf882d7db1e8eb68c91cb03f7e135c33. | 02 October 2019, 12:28:27 UTC |
16a09b4 | Kristoffer Carlsson | 02 October 2019, 09:37:06 UTC | fix not messing up `current_taks().storage[:SOURCE_PATH]` by allowing the finally in include to run (#33447) | 02 October 2019, 09:37:06 UTC |
ef0c910 | Steven G. Johnson | 02 October 2019, 07:36:11 UTC | fix exception type for sparse LDLᵀ (#33372) * fix exception type for sparse LDLᵀ * update tests for SingularException * whoops, is_ll means this is Cholesky * ZeroPivotException | 02 October 2019, 07:36:11 UTC |
035acf4 | Jameson Nash | 01 October 2019, 21:17:44 UTC | Merge pull request #33373 from JuliaLang/jn/profile-more-stats Profile: more formatting / information printing | 01 October 2019, 21:17:44 UTC |
311b0b9 | Yuri | 01 October 2019, 20:37:00 UTC | Fix a typo (handles => handled) (#33436) | 01 October 2019, 20:37:00 UTC |
2924113 | Iblis Lin | 01 October 2019, 20:36:14 UTC | llvm: add SONAME suffix to solve dlopen conflict on FreeBSD (#33058) close #32462 | 01 October 2019, 20:36:14 UTC |
8784619 | Matt Bauman | 01 October 2019, 20:12:35 UTC | Fixup crosslinks working around https://github.com/JuliaDocs/Documenter.jl/issues/781 | 01 October 2019, 20:14:55 UTC |
350f514 | Jeff Bezanson | 01 October 2019, 18:53:36 UTC | Array fixes for non-power-of-2 sized elements, fixes #26026 (#33283) | 01 October 2019, 18:53:36 UTC |
b288444 | Valentin Churavy | 01 October 2019, 17:44:33 UTC | Merge pull request #33439 from JuliaLang/aa/fix-whitespace Remove trailing whitespace | 01 October 2019, 17:44:33 UTC |
ef33864 | Alex Arslan | 01 October 2019, 17:30:49 UTC | Remove trailing whitespace This broke the trailing whitespace checker, which is now failing for all PRs against master, regardless of whether they contain trailing whitespace. | 01 October 2019, 17:30:49 UTC |
06d6b8e | Mustafa M | 01 October 2019, 16:01:34 UTC | Internally use libuv for mkdir on all platforms (#33422) Internally use libuv for mkdir on all platforms | 01 October 2019, 16:01:34 UTC |
cf98390 | Anshul Singhvi | 01 October 2019, 15:29:39 UTC | Add a note about Julia (#33412) | 01 October 2019, 15:29:39 UTC |
cb075dc | Olivier MATHIEU | 01 October 2019, 15:10:35 UTC | Fix abstract_type args error (#33369) See ```scheme (define (abstract-type-def-expr name params super) (receive (params bounds) (sparam-name-bounds params) `(block (global ,name) (const ,name) (scope-block (block ,@(map (lambda (v) `(local ,v)) params) ,@(map (lambda (n v) (make-assignment n (bounds-to-TypeVar v #t))) params bounds) (abstract_type ,name (call (core svec) ,@params) ,super)))))) ``` @ https://github.com/JuliaLang/julia/blob/v1.2.0/src/julia-syntax.scm#L880-L889 and ```lisp (struct_type ,name (call (core svec) ,@params) (call (core svec) ,@(map quotify field-names)) ,super (call (core svec) ,@field-types) ,mut ,min-initialized))) ``` https://github.com/JuliaLang/julia/blob/v1.2.0/src/julia-syntax.scm#L848-L850 | 01 October 2019, 15:10:35 UTC |
99d49b2 | P H Garcia | 01 October 2019, 15:10:04 UTC | explains that meta can be 'esc' (#33377) On Windows, in the REPL, alt-enter maximizes the screen. It took me a while to figure out that pressing Esc and then the key also works for the meta key. | 01 October 2019, 15:10:04 UTC |
aaa581a | Dilum Aluthge | 01 October 2019, 14:58:19 UTC | Clarify the difference between "light" and "full" source distributions (#32810) | 01 October 2019, 14:58:19 UTC |
4f19dc2 | Jameson Nash | 01 October 2019, 14:55:57 UTC | Profile,test: avoid profiling compiler time (#33314) | 01 October 2019, 14:55:57 UTC |
4fa6a62 | Alex Arslan | 01 October 2019, 06:25:34 UTC | Fix source build of p7zip on FreeBSD (#33431) | 01 October 2019, 06:25:34 UTC |
2859b11 | Jeff Bezanson | 01 October 2019, 02:50:28 UTC | improve display of keyword argument methods in stack traces (#33118) part of #33065 | 01 October 2019, 02:50:28 UTC |
6cf8dba | Nathan Daly | 30 September 2019, 22:02:43 UTC | Add `--trace-compile=` flag to `--help-hidden` output. (#31072) Since the `--trace-compile` flag is an unsupported, experimental flag, we don't want to document it in `--help`, which has forward maintainability requirements. Instead, we will put the help string in `--help-hidden`, which contains documentation for unsupported flags. | 30 September 2019, 22:02:43 UTC |
5e6a906 | Moritz Schauer | 30 September 2019, 21:56:21 UTC | Update docstring for `isnan` (#33258) | 30 September 2019, 21:56:21 UTC |
f696e17 | Kristoffer Carlsson | 30 September 2019, 08:09:35 UTC | fix using non zero axes arrays in threads loops (#33396) | 30 September 2019, 08:09:35 UTC |
d036575 | Simon Byrne | 29 September 2019, 04:13:34 UTC | Document internal Ryu functions, remove duplicate functionality. (#33406) Also moved some functions from Math to Base. | 29 September 2019, 04:13:34 UTC |
0b55ce1 | Jameson Nash | 29 September 2019, 03:50:02 UTC | Revert "Bump Pkg to 1.4-DEV. (#33364)" (#33410) This reverts commit 77af253c16719a149a5f6ae0d444b4321def3e35. | 29 September 2019, 03:50:02 UTC |
4849593 | Mustafa M | 28 September 2019, 21:03:44 UTC | Update from NSIS to Inno Setup installer (#33375) | 28 September 2019, 21:03:44 UTC |
a3eb9d4 | Keno Fischer | 28 September 2019, 19:17:12 UTC | div with rounding modes [+ rounded division] (#33040) * Re-arrange fld/cld code In preparation for supporting other rounding modes in div, create a three-argument div function that takes a rounding mode as the last argument and make this the fundamental fallback for fld/cld. * Implemented rounded division * add various divrem combinations to avoid overflow * Whitespace/test fixes * Small tweaks to docstrings * Bugfixes * Add the exhaustive test * Tigthen up types for div fallback I think it's better to give a MethodError here than an approximate answer for non-AbstractFloat reals (e.g. custom integer types). | 28 September 2019, 19:17:12 UTC |
e5c3ffc | Rafael Fourquet | 27 September 2019, 22:04:05 UTC | add NEWS for #32208, improved rand(::Tuple) (#33404) | 27 September 2019, 22:04:05 UTC |
1c1000a | Rafael Fourquet | 27 September 2019, 14:42:27 UTC | Merge pull request #32208 from JuliaLang/rf/rand/tuple-pow2 faster rand(::Tuple) | 27 September 2019, 14:42:27 UTC |
54f5acc | Benoît Legat | 27 September 2019, 10:21:41 UTC | docstring typo: Rea -> Real (#33383) | 27 September 2019, 10:21:41 UTC |
59b56ce | Rafael Fourquet | 31 May 2019, 16:10:27 UTC | faster rand(::MersenneTwister, ::OneTo) This was an oversight. This makes rand(::MersenneTwister, ::Tuple) generally faster. | 27 September 2019, 10:06:44 UTC |
2ae542f | Rafael Fourquet | 31 May 2019, 16:09:16 UTC | faster rand(::Tuple) for power of two lengths | 27 September 2019, 10:06:40 UTC |
36f34af | Alexander Seiler | 26 September 2019, 19:59:19 UTC | Rename `eigenvalues` and `eigenvectors` in `show` method for `Eigen` and `GeneralizedEigen` (#33393) Signed-off-by: Alexander Seiler <seileralex@gmail.com> | 26 September 2019, 19:59:19 UTC |
7528266 | Chris Foster | 26 September 2019, 03:17:00 UTC | Improved jl_unw_stepn: skip internal frames, don't adjust ip for signal frames Add an option to jl_unw_stepn to avoid adjusting the instruction pointer when we know the cursor derives from a signal frame. (We could alternatively try to do this with unw_is_signal_frame, but that wouldn't work for windows and would be an extra function call for each jl_unw_step which seems a bit unnecessary when we already know the top frame is the signal frame.) Also generalize skipping of the first few backtrace frames as needed to hide the internal backtrace machinery itself by adding a `skip` option to jl_unw_stepn/jl_backtrace_from_here/record_backtrace. As part of this also move the workaround for 32-bit windows into the backtrace internals. Move Base.backtrace into error.jl as it doesn't need any lookup functionality from StackTraces to work correctly. | 26 September 2019, 04:55:47 UTC |
9a1dbc0 | Valentin Churavy | 25 September 2019, 15:05:23 UTC | Merge pull request #33368 from yhls/yhls/fixrenaming fix bug related to block renaming for DCE | 25 September 2019, 15:05:23 UTC |
4a2a59a | Katharine Hyatt | 25 September 2019, 12:40:29 UTC | Add note formatting and some missing xrefs (#32930) | 25 September 2019, 12:40:29 UTC |
c33cf3e | Jakob Nybo Nissen | 25 September 2019, 12:39:18 UTC | Changed display of throw_eachindex_mismatch (issue #32890) (#33194) | 25 September 2019, 12:39:18 UTC |
5457323 | Steven G. Johnson | 25 September 2019, 12:33:14 UTC | cross-reference ntoh and hton from bswap docs (#33376) | 25 September 2019, 12:33:14 UTC |
99d1d67 | Alexander Seiler | 25 September 2019, 09:01:56 UTC | Fix multiplication of triangular matrix and sparse vector (#33378) * Fix multiplication of triangular matrix and sparse vector * Use variable `n` everywhere | 25 September 2019, 09:01:56 UTC |
77af253 | Fredrik Ekre | 25 September 2019, 08:51:24 UTC | Bump Pkg to 1.4-DEV. (#33364) | 25 September 2019, 08:51:24 UTC |
141ff51 | Chris Foster | 25 September 2019, 08:06:15 UTC | Clear ARMv7 instruction pointer encoding bit for use in caller lookup | 25 September 2019, 08:06:15 UTC |
8ba936f | Chris Foster | 25 September 2019, 03:46:07 UTC | Revert API breaking name change lookupat->lookup from #33190 With return->call address tanslation done inside jl_unw_stepn, there should largely be no need to make this API breaking change, as all backtrace addresses which the user has access to will already be translated and can be passed directly into `StackTraces.lookup()`, as was done previously. There's one possible exception to this rule: If the user were looking up the debug info for a pointer to a C function *not* obtained from a backtrace, they previously would have had to use `StackTraces.lookup(ptr+1)` to "undo" the the `ptr-1` translation that `StackTraces.lookup` used to perform prior to #33190. This seems like an edge case however and probably not worth breaking the API over. | 25 September 2019, 06:14:18 UTC |
6bc2065 | Chris Foster | 24 September 2019, 08:44:58 UTC | Clean up return->call address translation for StackTraces.lookup / jl_unw_stepn Translate return->call address internally in jl_unw_stepn so that users never have to deal with this subtlety when calling StackTraces.lookup. Also rearrange the jl_unw_stepn API a little so that interpreter stack traces can always be fit in to bt_data when calling jl_unw_stepn incrementally for deep stacks + add test cases for that (Fixes #29695). For the tests, clean up printing of CodeInfo objects wrapped in InterpreterIP. | 25 September 2019, 06:14:18 UTC |
30647b6 | Jameson Nash | 24 September 2019, 19:27:32 UTC | Profile: print paths more concisely Heuristically decide on a short path representation while printing. Look for a Project.toml file in a parent directly, take the name from it, and substitute that for the prefix of the absolute path. For relative paths, determine whether they are relative to the root of Base, or not actually paths (such as REPL[1]). | 24 September 2019, 20:29:20 UTC |
a016689 | Jameson Nash | 24 September 2019, 19:27:32 UTC | Profile: collect additional tree statistics - Adds some additional print formatting (such as a header) - Permit sorting tree view on alternate fields (as the table already did) - Collects additional StackFrame information while building the tree (printing the overhead information in the tree view) | 24 September 2019, 20:29:20 UTC |
55759af | Leon Shen | 23 September 2019, 22:21:55 UTC | fix bug related to block renaming for DCE | 23 September 2019, 22:48:12 UTC |
4f17558 | Jeff Bezanson | 23 September 2019, 16:31:11 UTC | add a subtyping fast path for tuple of Unions <: Vararg (#33353) fixes #33337 | 23 September 2019, 16:31:11 UTC |
187ee96 | Nick Robinson | 23 September 2019, 15:15:27 UTC | Broadcast AbstractRNG like a scalar (#33213) * Broadcast AbstractRNG like a scalar * Add PR link to NEWS | 23 September 2019, 15:15:27 UTC |
cc6ae96 | Steven G. Johnson | 23 September 2019, 12:08:25 UTC | document random reproducibility policy (#33350) * document random reproducibility policy * Update index.md * Update index.md * Update index.md * Update index.md | 23 September 2019, 12:08:25 UTC |
318affa | Steven G. Johnson | 23 September 2019, 07:05:19 UTC | bug fixes in matrix log (#32327) * bug fixes in matrix log * patches to matrix log (#33245) * patches to matrix log Avoid integer overflow if `s > 63`. Correct logic for `s == 0`. Only use fancy divided difference formulae if eigenvalues are close - avoids dangerous roundoff error if they are in opposite sectors. * add tests | 23 September 2019, 07:05:18 UTC |
b13c64a | Dilum Aluthge | 23 September 2019, 01:23:26 UTC | Add support for "all except x" in JULIA_DEBUG (#32432) | 23 September 2019, 01:23:26 UTC |
36220ed | Takafumi Arakaki | 21 September 2019, 23:35:30 UTC | Document getfield(value, i::Int) (#33346) | 21 September 2019, 23:35:30 UTC |