| b5f13c3 | Alex Arslan | 25 May 2020, 19:42:28 UTC | Set VERSION to 1.4.3 prerelease (#36007) Now that 1.4.2 has been tagged | 25 May 2020, 19:42:28 UTC |
| 44fa15b | Kristoffer Carlsson | 23 May 2020, 18:35:54 UTC | Set VERSION to 1.4.2 (#35882) | 23 May 2020, 18:35:54 UTC |
| 3c50b7f | Ian McInerney | 15 May 2020, 08:12:38 UTC | Add GCC10 to GCC version mappings (#35849) | 15 May 2020, 08:12:38 UTC |
| f16c4b0 | Kristoffer Carlsson | 14 May 2020, 06:29:38 UTC | Merge pull request #35825 from JuliaLang/backports-release-1.4 Backports release 1.4.2 | 14 May 2020, 06:29:38 UTC |
| c0b1a0c | Kristoffer | 10 May 2020, 11:24:34 UTC | bump Pkg version to latest release.1.4 | 10 May 2020, 11:24:34 UTC |
| c9b9b8c | Jeff Bezanson | 18 April 2020, 20:18:47 UTC | fix #35134, regression in printing nested quote Exprs (#35479) (cherry picked from commit e00db7b3e5656a58116bb89f8eb83499f0c9b763) | 10 May 2020, 11:22:30 UTC |
| 8b3a4a2 | Shan Sikdar | 08 May 2020, 19:08:34 UTC | Have matrix square root account for Rii=Rjj=0 to prevent NaN values (#35758) (cherry picked from commit 880c73199b5d304912b18b3e4f82f95b37c19e6f) | 10 May 2020, 10:05:21 UTC |
| 36d2600 | Tim Besard | 05 May 2020, 21:17:06 UTC | Fix and test PCRE error message getter. (#35728) (cherry picked from commit 9c2492eb87308348521a3833175f79e4bc015d9e) | 10 May 2020, 09:53:28 UTC |
| 0203e94 | Keno Fischer | 03 May 2020, 06:29:50 UTC | Fix dropped bytes when bulk reading from uv streams with oversized buffers (#35695) In readbytes!(s::UVStream, buf::AbstractVector{UInt8}, nb), we wrap `buf` in a PipeBuffer of maxsize nb and swap it in as the main buffer for the stream `s`. However, when we inform libuv of the size of the buffer, we instead use the size of the underlying array, which can be larger. In that case, we probably either drop bytes or override something that the user did not want overriden. (cherry picked from commit 3a84b51caf211159e402c2562962525c2b9249ea) | 10 May 2020, 09:53:27 UTC |
| e76bf14 | Rafael Fourquet | 27 April 2020, 22:55:48 UTC | fix buggy rand(RandomDevice(), Bool) (#35590) (cherry picked from commit 0bab06f8a7f70d4518ad30e19a4c0a1968c875ad) | 10 May 2020, 09:53:25 UTC |
| 9651235 | Jeff Bezanson | 27 April 2020, 17:47:56 UTC | exclude types with free variables from `type_morespecific` (#35555) (cherry picked from commit 0cedde636ed866364ad94ae0b4dd33918f447c40) | 10 May 2020, 09:53:24 UTC |
| 75c7c60 | Alex Arslan | 24 April 2020, 06:19:18 UTC | Add a fast path for returning "" from repeat(str, 0) (#35579) Currently the case where `r == 0` falls through the same logic as every other non-negative value of `r` (aside from 1). This works for signed integers. However, this does not work for unsigned integers: in the loop where we unsafely fill in the output string, we're looping from 0 to `r - 1`, which for unsigned integers wraps around and causes us to request the address of the output string at a location that is well beyond what was allocated. Fixes #35578. (cherry picked from commit 1dcb42f29caffc98ad0722577844ce5c54c5ba0c) | 10 May 2020, 09:53:23 UTC |
| 3be9f10 | Valentin Churavy | 22 April 2020, 22:00:35 UTC | implement LibuvStream interface for BufferStream (#35545) It is noted as a non-OS stream, but has also been a subtype of LibuvStream since forever. #32309 did not add the `readerror` field. (cherry picked from commit de042102d7d607d19224600ca4a990cdf3cf67dc) | 10 May 2020, 09:53:23 UTC |
| fe43406 | Shan Sikdar | 10 April 2020, 14:54:26 UTC | add some checks for invalid expressions (#35363) add null check for module-default-defs, fix #34544 fix #35367 as well change jl_eval_module_expr to check 3rd arg is block (cherry picked from commit 8057c6035479fa29f52da4fa10335cc1ab90a805) | 10 May 2020, 09:53:21 UTC |
| d04036f | Jameson Nash | 10 April 2020, 01:36:28 UTC | precompile: avoid a potential crash when adding methods (#35378) fixes #29859 (cherry picked from commit 8a55a27ea7f99be30bf17fc2c5e810cab4ded635) | 10 May 2020, 09:53:20 UTC |
| 889de01 | Elliot Saba | 12 March 2020, 22:02:40 UTC | Fix MacOS Notarization (#35080) We need to individually sign each Mach-O file independently, otherwise notarization bails. Note that you must sign the overall `.app` after each individual file is signed. I've also added more Entitlements. I've gone for a "gentle shotgun" approach, turning on everything that I think we could reasonably want. I wait expectantly for someone to complain that they need Location access turned on for some package. ;) (cherry picked from commit 418f1114a4a97524849bc3cb3077306a0b27691b) | 10 May 2020, 09:53:19 UTC |
| ef4fe83 | Alex Arslan | 15 April 2020, 16:24:20 UTC | Set VERSION to 1.4.2-pre (#35478) | 15 April 2020, 16:24:20 UTC |
| 381693d | Kristoffer Carlsson | 14 April 2020, 17:20:20 UTC | Set VERSION to 1.4.1 (#35467) | 14 April 2020, 17:20:20 UTC |
| ec31c9a | Kristoffer Carlsson | 14 April 2020, 08:03:06 UTC | Merge pull request #35229 from JuliaLang/backports-release-1.4.1 Backports for Julia 1.4.1 | 14 April 2020, 08:03:06 UTC |
| 7bd73b6 | Rafael Fourquet | 12 April 2020, 09:23:42 UTC | [backport #35274] fix #35272, searchsorted(3:-1:1, 2.5, rev=true) (#35435) | 12 April 2020, 09:23:42 UTC |
| 44fa722 | Keno Fischer | 08 April 2020, 18:55:11 UTC | [LateGCLowering] Fix skipped Select lifting (#35387) Fixes #35341 (cherry picked from commit 7a4ea219b5e4a2df3dc354476127e06d18b5506d) | 10 April 2020, 09:41:39 UTC |
| 6d635d7 | Jeff Bezanson | 08 April 2020, 17:29:39 UTC | fix #35391, macro hygiene bug in gc_preserve (#35394) (cherry picked from commit 8256be444d4420b485893993cd3f391be397ad1b) | 10 April 2020, 09:41:39 UTC |
| 7f5adff | Kristoffer Carlsson | 09 April 2020, 13:45:59 UTC | Merge pull request #35405 from JuliaLang/vc/bump_bb_1.4 [1.4.1] Update LLVM and OpenBLAS BB release | 09 April 2020, 13:45:59 UTC |
| 7e5ee11 | Valentin Churavy | 08 April 2020, 19:48:10 UTC | bump OpenBLAS release | 08 April 2020, 19:48:10 UTC |
| 8f46c77 | Valentin Churavy | 07 April 2020, 15:57:50 UTC | bump LLVM release | 08 April 2020, 16:06:28 UTC |
| c53191a | Elliot Saba | 06 April 2020, 22:09:07 UTC | Version `UnicodeData.txt` filename (#35383) We've been downloading multiple different versions of `UnicodeData.txt` across different branches and trying to keep a consistent checksum. (cherry picked from commit 922891ccbfcd74a0d10eca4f197b976ec08c20fd) | 07 April 2020, 06:24:25 UTC |
| 48ee62c | Valentin Churavy | 04 April 2020, 14:51:59 UTC | Merge pull request #35313 from JuliaLang/vc/bp_llvm8 [LLVM8] patch maintenance | 04 April 2020, 14:51:59 UTC |
| 61d70cb | Jeff Bezanson | 30 March 2020, 15:17:09 UTC | fix #35305, need escaping when printing string macro calls (cherry picked from commit b6dd448e3afab4a4a0015bb2e6122b97f649e5da) | 04 April 2020, 09:28:13 UTC |
| d455bda | Markus Kuhn | 15 December 2019, 20:58:40 UTC | adding escape_raw_string() Implementation of the raw-string escaping convention. Can also be used to escape command-line arguments for Windows C/C++/Julia applications, which use the same \" escaping convention as Julia non-standard string literals. (cherry picked from commit cdc5e791888e61b8f78ee9ccb2fcd21b94c7493e) | 04 April 2020, 09:28:05 UTC |
| 6c08ec4 | Kristoffer Carlsson | 31 March 2020, 14:30:31 UTC | fix method error printing for missing (#35315) fix method error printing for missing (cherry picked from commit 0b0b126ce73defed660aab11942d9307915f1926) | 04 April 2020, 09:27:29 UTC |
| 069f137 | Tim Holy | 29 March 2020, 09:06:53 UTC | Fix offset when widening collections (cherry picked from commit 1da24809378ab7c849f59c09270769e24d15a7b4) | 04 April 2020, 09:25:11 UTC |
| bc6adcd | Jeff Bezanson | 25 March 2020, 15:03:34 UTC | fix a bug preventing inlining of getindex of const globals (#35239) (cherry picked from commit 1f787492c3ac44797941f10c8bd452a1a7eef0f8) | 04 April 2020, 09:17:39 UTC |
| 625be9e | Elliot Saba | 04 April 2020, 00:18:11 UTC | Update `UnicodeData` checksum (#35356) Upstream has apparently modified this file as of March 30th 2020, which is annoying as it has broken all older builds of Julia attempting to download and check this file. (cherry picked from commit c9805191e32fde9d4aa55e59493178f47444926e) | 04 April 2020, 09:12:51 UTC |
| f73a14c | Jeff Bezanson | 31 March 2020, 14:54:53 UTC | fix #35295, typo in change to `clipboard` on linux (#35310) (cherry picked from commit 33df293a6a46d83bdd3a9caa08b311ef74452d60) | 04 April 2020, 09:12:50 UTC |
| bc6fa8f | Joseph Wilson | 30 March 2020, 15:32:02 UTC | Linux: Remove `--nodetach` option for `xsel` copy. (#35180) The `--nodetach` option causes `clipboard(x)` to hang indefinitely on (most? All which I have tried) linux systems. Removing this option fixes `clipboard(x)`, but briefly spawns a child process. Since `clipboard(x)` is designed for interactive use, this shouldn't be an issue. (cherry picked from commit 5dd66767f7884fd7de104d9b8144841774ecbdf0) | 04 April 2020, 09:12:48 UTC |
| 1156956 | Rafael Fourquet | 24 March 2020, 16:50:08 UTC | fix #35225, issubset on empty ranges (#35244) (cherry picked from commit 64ee8ab993db6ed806ca06c950dc6475f2e7dc0c) | 04 April 2020, 09:12:46 UTC |
| 7538e15 | Valentin Churavy | 30 March 2020, 16:04:03 UTC | bump LLVM BB release | 30 March 2020, 16:04:03 UTC |
| b4be5f2 | Valentin Churavy | 30 March 2020, 14:50:37 UTC | yank LLVM assert checksums | 30 March 2020, 14:50:37 UTC |
| 9131e12 | Milan Bouchet-Valat | 27 March 2020, 08:50:26 UTC | change libLLVM suffix to match build-configuration Backport of 5e0cab69966eb989f380492dda7590bca3a3ba8d. | 29 March 2020, 21:44:51 UTC |
| c6673ae | Valentin Churavy | 29 March 2020, 21:37:40 UTC | [LLVM8] add patches for D75072 and D65174 (cherry picked from commit d20184f90e8ddacb40222154714bc6140eaaa4a3) | 29 March 2020, 21:44:24 UTC |
| 8f1218a | Thomas Faingnaert | 14 February 2020, 14:01:04 UTC | Add AddrSpacePtr type (cherry picked from commit 8cb8f0c485a81082813dde82b7caba29e2343c36) | 23 March 2020, 08:58:36 UTC |
| e54a0de | Morten Piibeleht | 23 March 2020, 08:10:04 UTC | Bump Documenter to 0.24.7 (#35227) (cherry picked from commit d37778535283b5484ec76759d227612190a3b63d) | 23 March 2020, 08:58:36 UTC |
| 23e8527 | Jeff Bezanson | 23 March 2020, 03:46:25 UTC | fix #35201, syntax error with named splatting inside function call (#35202) (cherry picked from commit 59d7522eaecc86783078eac8746516e3a0d8a308) | 23 March 2020, 08:58:36 UTC |
| 7ccda0e | Jeff Bezanson | 18 March 2020, 01:32:06 UTC | fix bug in t-function for `>:` (#35136) (cherry picked from commit 2a5bb59facdce6c8d987a3752f2b50e2e06d7bef) | 23 March 2020, 08:58:36 UTC |
| 9dbe7e8 | Cédric St-Jean | 22 March 2020, 19:58:50 UTC | Remove outdated paragraph from deepcopy docstring (#35192) (cherry picked from commit 6baa701c571ff3dd976e45fb9892d5430993edb5) | 23 March 2020, 08:58:36 UTC |
| 0e24a7f | Rafael Fourquet | 20 March 2020, 21:26:35 UTC | add compat annotation for map!(f, values(dict)) (from #31223) (#35184) (cherry picked from commit 644753491de0d5d933c96e6f1f5876b3c7603cf4) | 23 March 2020, 08:58:36 UTC |
| 71fcb9a | Jeff Bezanson | 20 March 2020, 19:30:48 UTC | speed up llvmcall unique name generation (#35144) (cherry picked from commit 5d5f799ed43b3ef458e16fba4f34533f5719c614) | 23 March 2020, 08:58:36 UTC |
| f1593e1 | Curtis Vogt | 16 March 2020, 19:44:17 UTC | Support git config credential resetting (#35064) (cherry picked from commit 0f1b1192735e1c05c5aa0eab85bef92250abe05c) | 23 March 2020, 08:15:15 UTC |
| 5cf6725 | Elliot Saba | 06 March 2020, 21:33:03 UTC | walkdir: avoid symlink loops when `follow_symlinks == false` (#35006) * walkdir: avoid symlink loops when `follow_symlinks == false` Because `isdir()` attempts to dereference symlinks, attempting to `walkdir()` trees that contain symlink loops errors out. This change modifies `walkdir()` to treat all symlinks as files when `follow_symlinks == false`. * rm: When checking `filemode()`, use `lstat()` to avoid following symlinks (cherry picked from commit 178ac974b5e9f649c0ffb3ed78752dc9241b7a5a) | 23 March 2020, 08:15:14 UTC |
| 4e11ce9 | Viral B. Shah | 05 March 2020, 02:18:58 UTC | Rebase of #28526 on master (#34989) * fix \ SparseVector * split (\) for sparsevector; test for Factor and Symmetric Sparse Matrix Co-authored-by: Chi Po Choi <choi.chi.po@gmail.com> (cherry picked from commit d2f967766ca1622e244fdb8136c2f5374d697fee) | 23 March 2020, 08:15:13 UTC |
| 1d13caa | Keno Fischer | 04 March 2020, 22:49:52 UTC | Sockets: fix return value of getpeername/getsockname (#34986) (cherry picked from commit 598209de2161df196419cb39a2d595c6abf8763b) | 23 March 2020, 08:15:12 UTC |
| ec8f59c | Carsten Bauer | 02 March 2020, 05:51:43 UTC | Fix inverse of SVD of complex matrix (#34872) Closes #34866 (cherry picked from commit f7093312ca8747765c96b8571f47b4bfa2e53708) | 23 March 2020, 08:15:10 UTC |
| de3914e | Keno Fischer | 28 February 2020, 03:34:01 UTC | Fix regression in inlining of invoke (#34906) When we added the check to prevent inlining through ambiguous methods, we failed exclude this check for calls to `invoke` (which skip ambiguous methods, since the method is explicitly selected). Fixes #34900. (cherry picked from commit 6abc8521bfe3af422866cb91b196367706d239b6) | 23 March 2020, 08:15:10 UTC |
| b8e9a9e | Kristoffer Carlsson | 21 March 2020, 16:36:32 UTC | Set VERSION to 1.4.0 (#35168) | 21 March 2020, 16:36:32 UTC |
| 268f9d2 | Kristoffer Carlsson | 19 March 2020, 11:33:15 UTC | Merge pull request #34992 from JuliaLang/backports-release-1.4 Backports for Julia 1.4.0 | 19 March 2020, 11:33:15 UTC |
| 8c2a944 | Elliot Saba | 19 March 2020, 05:43:34 UTC | Bump LLVM tarballs on 1.4 (#35120) * Switch LLVM BB download to new format * use cxx11-specific downloads instead of GCC-version-specific downloads * Default to `cxx11` when it's indeterminate * handle CXX expanded BB builds * Add LLVM_full checksums Co-authored-by: Valentin Churavy <v.churavy@gmail.com> | 19 March 2020, 05:43:34 UTC |
| 61b3ea4 | Elliot Saba | 04 March 2020, 05:35:55 UTC | Fix `$(PYTHON)` paths for native windows executable or cygwin (#34984) (cherry picked from commit 0e8143e1e7d125b74e474d0fbd3a5a5e2d574779) | 04 March 2020, 09:46:46 UTC |
| b8ce43a | Ian Butterworth | 04 March 2020, 08:49:42 UTC | Add missing 1.4 news note for new function pkgdir() (#33128) (#34663) * add note for new function pkgdir() (#33128) Co-Authored-By: Fredrik Ekre <ekrefredrik@gmail.com> Co-authored-by: Viral B. Shah <viral@juliacomputing.com> Co-authored-by: Fredrik Ekre <ekrefredrik@gmail.com> | 04 March 2020, 08:49:42 UTC |
| 467d3a1 | Kenta Sato | 01 March 2020, 16:49:01 UTC | fix unsigned(::Ptr) and signed(::Ptr) (#34941) (cherry picked from commit cdf580aff8642e34889cb1c3c8445992663d16fa) | 04 March 2020, 08:43:05 UTC |
| b99ed72 | Kristoffer Carlsson | 24 February 2020, 16:51:22 UTC | Set VERSION to 1.4.0-rc2 (#34841) | 24 February 2020, 16:51:22 UTC |
| 7612062 | Kristoffer Carlsson | 22 February 2020, 09:08:41 UTC | Merge pull request #34517 from JuliaLang/backports-release-1.4 Backports for Julia 1.4 RC2 | 22 February 2020, 09:08:41 UTC |
| b4edc25 | Tim Holy | 15 February 2020, 12:20:38 UTC | Convert range type in `reduced_index` The `reduced_indices` and `reduced_indices0` methods sometimes assert that the return axes type is the same as the input. Consequently, the implementation of `reduced_index` had better return a range of the same type as the input. This corrects the error in https://github.com/JuliaArrays/OffsetArrays.jl/issues/92. I'll put a workaround in OffsetArrays.jl too. (cherry picked from commit eda4131859229c7d9de4461e154a73bb89a6b783) | 21 February 2020, 22:06:46 UTC |
| e0d0bb1 | Jeff Bezanson | 21 February 2020, 16:02:10 UTC | fix #34826, unwrap VecElement when needed in `julia_const_to_llvm` (#34829) (cherry picked from commit a3f7fa8611f8d0036d01f449ef893c65050d9f6a) | 21 February 2020, 19:34:00 UTC |
| e161506 | Sebastian Micluța-Câmpeanu | 21 February 2020, 16:00:05 UTC | Fix #34316 (#34817) Fix hypot with both arguments 0 Remove hypot from Furlongs in order to test the defined fallback (cherry picked from commit 8e3aa66ab7084e04b77ce8f08fddd39340e70351) | 21 February 2020, 19:33:59 UTC |
| 6ab792f | Kristoffer | 20 February 2020, 15:40:18 UTC | use Pkg release 1.4 | 20 February 2020, 15:40:18 UTC |
| 9fe1a5e | Jameson Nash | 19 February 2020, 18:51:30 UTC | Profile: fix indent computation BoundsError (#34806) fixes #34723 (cherry picked from commit f5afdf908033f667791df05c0fa800258ca9bb67) | 19 February 2020, 20:31:16 UTC |
| 967bc47 | Jameson Nash | 19 February 2020, 18:50:55 UTC | codegen: fix segfaults with VecElement with union element (#34805) We were being too aggressive at unwrapping these: since they are already unwrapped once, they need to be treated almost as a normal struct. fixes #29864 (cherry picked from commit 8720530fae5e7f34003057fe487c871ca81560a0) | 19 February 2020, 20:31:15 UTC |
| d9d6364 | Jameson Nash | 19 February 2020, 18:43:38 UTC | codegen: fix bad declaration of memcmp stdlib function signature (#34792) We were lying to LLVM, and it now catches that in the verifier. Stop lying to LLVM so that it will again be happy to generate good code for us (e.g. optimizing this to bcmp). (cherry picked from commit d634f73f2b7445b8bf331707242ea0421676a293) | 19 February 2020, 20:31:15 UTC |
| a90cfd2 | Keno Fischer | 19 February 2020, 18:20:48 UTC | partr: Fix deadlock (#34807) When there is no work to do, the first thread to be idle will attempt to run the event loop once, waiting for any notifications (which will usually create new work). However, there is an interesting corner case where a notification arrives, but no work was scheduled. That doesn't usually happen, but there are a few situations where it does: 1) Somebody used a libuv primitive outside of julia, so the callback doesn't schedule any julia work. 2) Another thread forbily interrupted the waiting thread because it wants to take over the event loop for various reasons 3) On Windows, we occasionally get spurious wake ups of the event loop. The existing code in partr assumed that we were in situation 2, i.e. that there was another thread waiting to take over the event loop, so it released the event loop and simply put the current thread to sleep in the expectation that another thread will pick it up. However, if we instead are in one of the other two conditions, there may not be another thread there to pick up the event loop. Thus, with no thread owning the event loop, julia will stop responding to events and effectively deadlock. Since both 1 and 3 are rare, and we don't actually enter the event loop until there was no work for 4 milliseconds (which is fairly rare), this condition rarely happens, but is occasionally observable on Windows, where it caused #34769. To test that this fix works, we manually create situation 1 in the test by creating an idle callback, which will prevent the event loop from blocking, but only schedules julia work after it's been called 100 times. This reproduces the observed failure from the issue and is fixed by this PR. Fixes #34769 Co-authored-by: Jeff Bezanson <jeff@juliacomputing.com> Co-authored-by: Jameson Nash <jameson@juliacomputing.com> (cherry picked from commit f36edc2c2429140663baeedddd4cf11712c4469d) | 19 February 2020, 20:31:14 UTC |
| f4b4629 | Ryan Birmingham | 18 February 2020, 18:55:32 UTC | make alignment deal with float32 like float64 (#34763) (cherry picked from commit b0d1c1abb69981e80640fb1501ce88d597b9cf74) | 19 February 2020, 20:31:13 UTC |
| 2592da4 | Elliot Saba | 19 February 2020, 19:48:30 UTC | Merge pull request #34801 from JuliaLang/sf/LLVM_asserts_jll_1.4 Add LLVM asserts checksums | 19 February 2020, 19:48:30 UTC |
| 81e14fb | Elliot Saba | 18 February 2020, 18:08:27 UTC | Add LLVM asserts checksums | 18 February 2020, 18:08:27 UTC |
| 27d2693 | Jameson Nash | 13 February 2020, 21:49:36 UTC | ccall: avoid some bad symbol mangling Other bad cases are probably less common, but still exist. The code to fix those is included in #25984. Fixes #34680 (cherry picked from commit 66ecd90bd6359a3094bb21964830b0f75d71bcda) | 17 February 2020, 14:36:04 UTC |
| 6b89ea6 | Jameson Nash | 13 February 2020, 19:40:52 UTC | build: fix missing function prototype on older gcc The ffs function is not always available under that name, use instead the compiler name for it which may always be used. (cherry picked from commit ef17985df6ee670dec4d266b7ed9cf56beca295e) | 17 February 2020, 14:35:56 UTC |
| 317b4ef | Takafumi Arakaki | 14 February 2020, 20:44:17 UTC | Remove Iterators.reverse(::Tuple) specialization (#34732) (cherry picked from commit 48f942bdd659ae55d7d3e34dcf15dbb8bd871d71) | 17 February 2020, 14:35:41 UTC |
| df2da49 | Jeff Bezanson | 14 February 2020, 18:24:21 UTC | fix #34752, inference bug in varargs with constant prop (#34755) (cherry picked from commit 438485e5a93f240ab68a28190a798c263e5a7a25) | 17 February 2020, 14:35:26 UTC |
| 4f81f94 | Jameson Nash | 04 February 2020, 22:17:45 UTC | [GCLowering] strip all alias information Even non LoadInst / StoreInst (esp. vectorized loads and stores) can have this information (cherry picked from commit 40d0bc55cf74c18e5df0a0517de7ad69d14238f5) | 06 February 2020, 08:37:12 UTC |
| 533bbad | Valentin Churavy | 29 January 2020, 22:13:42 UTC | [GCLowering] handle vectorized loads Vectorized loads can come in 3 variants: - scalar - masked - gather All work about the same--if we run into a Loaded pointer, we need to follow back the base pointer and track it. And otherwise they're just normal def/use statements. But we also need to fix up after an LLVM mistake and put a valid value in the pass-through slot. Co-authored-by: Jameson Nash <vtjnash@gmail.com> (cherry picked from commit 7d1c9254e7b4a8f675a0bb6eeb7c6590a2fe7252) | 06 February 2020, 08:36:59 UTC |
| 3a22e2f | Jeff Bezanson | 03 February 2020, 19:41:05 UTC | add some helpful type information to Base and REPLCompletions (#34596) fixes #34098 (cherry picked from commit cced577b79316ea38482d84b69e7be9666d14041) | 05 February 2020, 09:27:45 UTC |
| eb75505 | Jeffrey Lin | 03 February 2020, 10:21:16 UTC | src/codegen.cpp: fix segfault on `@code_native` with eltype (#34434) (#34458) (cherry picked from commit 6a96a4dbdbce5cb8fc58639690b04329f7042610) | 05 February 2020, 09:27:45 UTC |
| 880b9b1 | Kenta Sato | 03 February 2020, 10:20:51 UTC | make package cache interit permission from source (#34573) * make package cache interit permission from source * remove temporary directries (cherry picked from commit 50754720d76699fd2863b33f8bb409235b2505f9) | 05 February 2020, 09:27:44 UTC |
| 9850ec4 | Takafumi Arakaki | 01 February 2020, 12:26:36 UTC | Construct MulAddMul at gemm_wrapper! call sites (#34601) * Construct MulAddMul at gemm_wrapper! call sites * Add branches manually in MulAddMul constructor This is suggested by chethega in: https://github.com/JuliaLang/julia/pull/29634#discussion_r314010599 * Update stdlib/LinearAlgebra/src/generic.jl Co-Authored-By: Kristoffer Carlsson <kristoffer.carlsson@chalmers.se> Co-authored-by: Kristoffer Carlsson <kristoffer.carlsson@chalmers.se> (cherry picked from commit 2da42e0f1365d1ab4c3e9213a6b488fef1dc693d) | 05 February 2020, 09:27:43 UTC |
| 5b4c5c2 | Jameson Nash | 30 January 2020, 20:49:05 UTC | alignment: subtly change meaning of datatype_align (#34473) Rather than meaning the actual alignment of the object, this now means the preferred alignment of the object. The actual alignment of any object is the minimum of this preferred alignment and the alignment supported by the runtime allocator. This aligns us with how LLVM treats alignment, and is probably reasonably sensible anyways. Also tries to audit our existing uses of CreateLoad/CreateStore for correctness, and upgrade some to include pointer-types. fixes #32414 (cherry picked from commit 268ac245734d5a3ef4ce9f9d4876aec3fedaa6e6) | 05 February 2020, 08:32:55 UTC |
| 2a64ee2 | Erik Schnetter | 24 January 2020, 14:13:51 UTC | jl_special_vector_alignment: Remove work-around for old LLVM version (<=3.8) (#34490) LLVM <= 3.8 segfaults when LLVM SIMD vectors are not powers of 2 (or a sum of 2 different powers of 2). This LLVM bug was corrected in 2017 (see <https://bugs.llvm.org/show_bug.cgi?id=27708>). This change removes this work-around, finally allowing arbitrary SIMD vector lengths in Julia. This changes Julia's ABI for SIMD types. All SIMD types are now represented as SIMD vectors in LLVM. The work-around represented some SIMD types as LLVM arrays instead to avoid the LLVM segfault. I don't think that SIMD vector lengths that are not a power of 2 were in widespread use, if at all. (cherry picked from commit e0740fe5a67177e9a9a9ee57cde9d13379f7bc69) | 05 February 2020, 08:32:55 UTC |
| ca11925 | KristofferC | 31 January 2020, 07:35:02 UTC | bump Pkg to newest on release-1.4 | 05 February 2020, 08:32:53 UTC |
| 42a7d5d | Elliot Saba | 23 January 2020, 00:10:29 UTC | Upgrade buildsystem to look for `python`, `python3` or `python2` (cherry picked from commit 181050b3b0675da8053c118700dbf27da73ef1f8) | 31 January 2020, 07:31:53 UTC |
| 7ba75c4 | Andreas Noack | 31 January 2020, 07:16:33 UTC | Make SuiteSparse wrapper thread safe. (#34546) Fixes #34500 (cherry picked from commit 51f1710d7eac046de2ba4cc580ceb030e37f57a7) | 31 January 2020, 07:26:29 UTC |
| 46c4328 | Tim Besard | 28 January 2020, 21:33:34 UTC | Make testall target work with JULIA_BUILD_MODE=debug again. (#34402) (cherry picked from commit aa13637bdde6ab517054d40ec4552e3eb3f64f23) | 31 January 2020, 07:26:26 UTC |
| c9b2c33 | Takafumi Arakaki | 27 January 2020, 13:00:46 UTC | Always use Symbol as _group value of log records (#34509) (cherry picked from commit 54c1667b7a8bb27c0cf5993fc97cd0caf0ed30d2) | 31 January 2020, 07:26:25 UTC |
| 0adcee9 | Fredrik Ekre | 24 January 2020, 08:11:58 UTC | Remove URLs with /latest and use correct, relative, (#34491) documenter links for the manual, fixes #34488. (cherry picked from commit 4db419b7f689a656be6b5daf266a12ab446cbf9f) | 31 January 2020, 07:26:24 UTC |
| 449d69d | Elliot Saba | 30 January 2020, 19:56:21 UTC | Merge pull request #34561 from JuliaLang/sf/bump_openblas_version [release-1.4] Bump OpenBLAS BB tarball version and update checksums | 30 January 2020, 19:56:21 UTC |
| 7151b45 | Elliot Saba | 28 January 2020, 23:19:59 UTC | Bump OpenBLAS BB tarball version and update checksums | 30 January 2020, 07:31:36 UTC |
| 0d3f01c | Elliot Saba | 30 January 2020, 07:00:11 UTC | tests: Use `realpath()` before `==` path comparisons (#34506) (#34581) These should probably be using `samefile`, if they were real code instead of just tests. Though it's unclear why real code would be doing this. Maybe just don't put your paths in hash-tables and you'll normally be fine. (cherry picked from commit a5c422fea8ca32baac92501768713b704d3611c3) | 30 January 2020, 07:00:11 UTC |
| b963fac | simeonschaub | 25 January 2020, 18:06:36 UTC | fix printing of LineNumberNode in block within ref (#34504) (cherry picked from commit 0234e009abb5adac4b36bd3ab47694ba2c2bf82a) | 26 January 2020, 14:06:44 UTC |
| 5eaa8d6 | simeonschaub | 24 January 2020, 23:30:22 UTC | fix bug in printing of block expressions in ref (#34503) (cherry picked from commit 5157d080c42b865ab70117e19a5e102a51531a16) | 26 January 2020, 14:06:43 UTC |
| 9196ba1 | Kristoffer Carlsson | 23 January 2020, 23:09:11 UTC | Update Documenter version to 0.24.4. (#34487) (cherry picked from commit 6af5f03d2ac7945b8cff3b5eee715a3d9446e42c) | 26 January 2020, 14:06:42 UTC |
| b0c33b0 | Alex Arslan | 23 January 2020, 17:23:55 UTC | Set VERSION to 1.4.0-rc1 (#34452) | 23 January 2020, 17:23:55 UTC |
| 8852200 | Tamas K. Papp | 22 January 2020, 12:38:06 UTC | fix typo in #34465 (#34476) thanks at-dkarrasch | 22 January 2020, 12:38:05 UTC |
| e4f6cd7 | Kristoffer Carlsson | 21 January 2020, 22:00:28 UTC | Merge pull request #34461 from JuliaLang/backports-release-1.4 Backports for Julia 1.4 RC 1 | 21 January 2020, 22:00:28 UTC |
| a718746 | Tamas K. Papp | 21 January 2020, 20:54:18 UTC | the PR was not mentioned in NEWS.md, fix (#34465) | 21 January 2020, 20:54:18 UTC |
