ed99c71 | Steven G. Johnson | 23 December 2020, 19:05:30 UTC | rm ill-advised Fix2/Fix3 special casing | 23 December 2020, 19:05:30 UTC |
b9c30ff | Steven G. Johnson | 23 December 2020, 18:31:31 UTC | rebase underscore currying | 23 December 2020, 18:31:31 UTC |
0e15141 | Mustafa M | 23 December 2020, 07:17:49 UTC | Remove stale contrib files (#38970) | 23 December 2020, 07:17:49 UTC |
549a73b | Elliot Saba | 22 December 2020, 08:49:32 UTC | [Artifacts] Note the platform we're looking for in our error message (#38960) | 22 December 2020, 08:49:32 UTC |
c708ca2 | Elliot Saba | 22 December 2020, 00:12:51 UTC | Source tarballs should contain `StdlibArtifacts.toml` files (#38962) Without this, our full source dist tarballs still try to download these. | 22 December 2020, 00:12:51 UTC |
c3d3c6a | Sacha Verweij | 21 December 2020, 23:36:42 UTC | Carry MBedTLS patch that works around CMake 3.18.2 bug. (#38933) MBedTLS fails to build with CMake 3.18.2, due to a bug in CMake 3.18.2 causing failure to find python3. For more information, please see https://github.com/ARMmbed/mbedtls/issues/3690 and https://gitlab.kitware.com/cmake/cmake/-/issues/21204. This pull request applies the MBedTLS patch that works around this issue from https://github.com/ARMmbed/mbedtls/pull/3691, which has been merged into MBedTLS's development branch but not yet into a tagged release. The latest minor release as of that merge was 2.25.0, so it's probably safe to conjecture that this patch will be live as of minor release 2.26.0; it's also possible that this patch will land on patch releases for 2.24.0 and 2.25.0, but this author doesn't know. | 21 December 2020, 23:36:42 UTC |
684d5fc | Dilum Aluthge | 20 December 2020, 21:12:54 UTC | Support passing `auto` as the value to the `JULIA_NUM_THREADS` environment variable | 21 December 2020, 19:16:39 UTC |
5cd07f8 | Gunnar Farnebäck | 21 December 2020, 11:50:43 UTC | Repair jl_init. (#38950) * Export jl_get_libdir and use it from jl_init to find libjulia. | 21 December 2020, 11:50:43 UTC |
c303262 | Shuhei Kadowaki | 21 December 2020, 09:34:19 UTC | minor inferrability tweak for `manifest_deps_get` (#38954) | 21 December 2020, 09:34:19 UTC |
58d6684 | Mustafa M | 20 December 2020, 15:11:53 UTC | Correct exponent_max to conform to IEEE definition (#38929) * Exponent exponent_max returns actual maximum exponent, not the sentinel. Co-authored-by: mschauer <moritzschauer@web.de> | 20 December 2020, 15:11:53 UTC |
1887550 | Bartosz Białas | 19 December 2020, 18:09:15 UTC | Ensure errors in startup.jl don't block the REPL (#38261) | 19 December 2020, 18:09:15 UTC |
742a4a0 | Keno Fischer | 19 December 2020, 13:48:27 UTC | Merge pull request #35539 from JuliaLang/kf/stdlibtests Test suite tweaks for rr | 19 December 2020, 13:48:27 UTC |
7a3c153 | Keno Fischer | 16 December 2020, 02:54:44 UTC | Exclude `threads` test from rr tracing | 19 December 2020, 07:45:31 UTC |
2d0211b | Keno Fischer | 22 April 2020, 20:01:46 UTC | Disable thread affinity test under rr RR changes the affinity, so 0,1 may not be available in the affinity mask. It's a bit complicated to fix this test and would require parsing the existing affinity mask, so just disable it for now. | 19 December 2020, 07:45:31 UTC |
ccfbc90 | Keno Fischer | 22 April 2020, 19:12:06 UTC | Don't detach rr workers in SharedArrays tests By default, I'm having the test suite set up to put any workers spawned with addprocs into their own rr sessions. This doesn't work with SharedArrays, because those workers share memory with the main process. Add a new JULIA_RR environment variable that specifies which rr to use for the test suite and a flag to the testsuite's addprocs command to determine whether or not to detach the workers from the current rr session. | 19 December 2020, 07:45:31 UTC |
6ae3ded | Keno Fischer | 21 April 2020, 06:03:09 UTC | FileWatching: Dump open file descriptors on failure If the FDs we allocated are too high, we leaked fds somewhere. On linux it's fairly easy to dump out information about all active FDs, so do that in the failure case, so the logs can help in debugging. | 19 December 2020, 07:45:31 UTC |
89fa9c6 | Keno Fischer | 21 April 2020, 06:01:37 UTC | threads: Pass env through to thread_exec We want to start running rr on CI, but at the moment the threads test is clearing ENV, causing it to lose the LD_PRELOAD that improves rr performance. Fix that by instead passing through the environment and setting the relevant JULIA_NUM_THREADS flag on top of it. | 19 December 2020, 07:45:31 UTC |
676ccf4 | Erik Schnetter | 18 December 2020, 19:53:12 UTC | Correct formatting for `sizehint!` documentation (#38939) | 18 December 2020, 19:53:12 UTC |
114c784 | Tamas K. Papp | 18 December 2020, 17:56:57 UTC | mention that InteractiveUtils is loaded automatically in interactive mode. (#38937) | 18 December 2020, 17:56:57 UTC |
aab283b | Katharine Hyatt | 18 December 2020, 17:40:43 UTC | Missing tests for arrays (#38926) | 18 December 2020, 17:40:43 UTC |
76952a8 | Colin Caine | 18 December 2020, 17:32:19 UTC | Add 2-arg versions of findmax/min, argmax/min (#35316) Defines a descending total order, `isgreater` (not exported), where unordered values like NaNs and missing are last. This makes defining min, argmin, etc, simpler and more consistent. Also adds 2-arg versions of findmax/min, argmax/min. Defines and exports the `isunordered` predicate for testing whether a value is unordered like NaN and missing. Fixes #27613. Related: #27639, #27612, #34674. Thanks to @tkf, @StefanKarpinski and @drewrobson for their assistance with this PR. Co-authored-by: Jameson Nash <jameson@juliacomputing.com> Co-authored-by: Takafumi Arakaki <takafumi.a@gmail.com> | 18 December 2020, 17:32:19 UTC |
a88c435 | Jeff Bezanson | 18 December 2020, 16:59:41 UTC | fix #38888, pessimistic sparam inference with concrete upper bound (#38899) | 18 December 2020, 16:59:41 UTC |
17de527 | Steven G. Johnson | 18 December 2020, 15:34:25 UTC | Unicode-compliant islower/uppercase (#38574) * Unicode-compliant islower/uppercase * don't test isletter for non-L* letters * include titlecase in alphas test * add news | 18 December 2020, 15:34:25 UTC |
d20ca48 | Milan Bouchet-Valat | 18 December 2020, 12:28:12 UTC | Make `findall` faster for AbstractArrays (#37177) The `findall` fallback is quite slow when predicate is a small function compared with generating a logical index using `broadcast` and calling `findall` on it to compute integer indices. The gain is most visible when predicate is true for a large proportion of entries, but it's there even when all of them are false. The drawback of this approach is that it requires allocating a vector of `length(a)/8` bytes whatever the number of returned indices. | 18 December 2020, 12:28:12 UTC |
9921e8c | Kristoffer Carlsson | 18 December 2020, 03:49:24 UTC | improve performance of fieldname by moving out error paths in separate functions (#38921) | 18 December 2020, 03:49:24 UTC |
9e27eee | Mustafa M | 18 December 2020, 02:53:08 UTC | Remove stale WinRPM, Vagrant, Appveyor CI files and update build instructions on Windows (#38910) * Delete winrpm, vagrant and appveyor build files * Update Windows build instructions | 18 December 2020, 02:53:08 UTC |
e2cf172 | Valentin Churavy | 17 December 2020, 23:24:43 UTC | Merge pull request #38852 from JuliaLang/vc/libgit2_va Mark ccall to git_libgit2_opts as variadic | 17 December 2020, 23:24:43 UTC |
e2fe24a | Valentin Churavy | 17 December 2020, 22:17:37 UTC | Merge pull request #38934 from JuliaLang/vc/llvm_11_patches [LLVM] Add patches and update binaries | 17 December 2020, 22:17:37 UTC |
a2f0337 | Elliot Saba | 17 December 2020, 22:07:12 UTC | Remove useless nghttp2_jll import from LibGit2_jll (#38935) This was a copy-paste error. | 17 December 2020, 22:07:12 UTC |
5327824 | Jameson Nash | 17 December 2020, 21:41:01 UTC | tasks,debugging: make it possible to get the backtrace of a task (#32283) This should work for any non-copy stack task. To make it work better, this now switches to the JL_HAVE_UNW_CONTEXT by default for some platforms. Also export the list of all live (currently running or suspended) tasks which have real stacks (the non-copy-stack tasks) which were started by the current thread. | 17 December 2020, 21:41:01 UTC |
fad55ab | Jameson Nash | 17 December 2020, 21:40:22 UTC | Switch back to LLVM ORC v2, take two (#38804) Refs #38008 & #38785 (This reverts commit 599d3299c9666d8aae13843c3551eaeb2d25f07d, and fixes a load ordering issue with our JuliaOJIT::addModule) | 17 December 2020, 21:40:22 UTC |
535a9b1 | Valentin Churavy | 12 December 2020, 18:24:36 UTC | Mark ccall to git_libgit2_opts as variadic | 17 December 2020, 21:22:18 UTC |
8a84395 | Valentin Churavy | 17 December 2020, 18:03:17 UTC | bump LLVM BB | 17 December 2020, 19:38:25 UTC |
c39d661 | Valentin Churavy | 15 December 2020, 15:34:00 UTC | [LLVM] Add patches for PPC knownbits and AArch64 globalisel | 17 December 2020, 19:38:25 UTC |
996dd55 | Carsten Bauer | 17 December 2020, 16:01:59 UTC | Improve position of MKL.jl mention in build documentation (#38923) * Improve position of MKL.jl mention * Clarify that SuiteSparse will not use MKL if going down the MKL.jl path Co-authored-by: Viral B. Shah <ViralBShah@users.noreply.github.com> | 17 December 2020, 16:01:59 UTC |
9c2d813 | Viral B. Shah | 17 December 2020, 14:33:41 UTC | Document the need for disabling BB if building without network access (#38918) * Document the need for disabling BB if building without network access Fix #33569 * Update doc/build/build.md Co-authored-by: Dilum Aluthge <dilum@aluthge.com> Co-authored-by: Keno Fischer <keno@juliacomputing.com> Co-authored-by: Dilum Aluthge <dilum@aluthge.com> | 17 December 2020, 14:33:41 UTC |
b36338a | Viral B. Shah | 17 December 2020, 11:42:33 UTC | Update non-BB dsfmt build to match with the BB one. (#38917) * Update non-BB dsfmt build to match with the BB one. Update URLs to https Fix #17945 * Add -DDSFMT_SHLIB | 17 December 2020, 11:42:33 UTC |
9ec3147 | Max Horn | 17 December 2020, 11:39:12 UTC | Restore the version information for libjulia.dylib (#38829) Fixes #38782 | 17 December 2020, 11:39:12 UTC |
fc577d0 | Valentin Churavy | 17 December 2020, 11:38:22 UTC | Execute `backtrace` once before testing formatting (#38886) On some platforms (PowerPC) the call to the `jlplt` is not a tail-call and so it will be part of the backtrace. This means we are off-by-one and won't skip the Julia function `backtrace` messing up tests that check precise formatting. | 17 December 2020, 11:38:22 UTC |
b5dd8a1 | Kristoffer Carlsson | 17 December 2020, 11:33:45 UTC | Merge pull request #38906 from JuliaLang/teh/precompiles_revise Reduce latency of Revise.revise() | 17 December 2020, 11:33:45 UTC |
3853060 | Sacha Verweij | 17 December 2020, 02:08:36 UTC | Correct pcre2-cet-flags.patch whitespace. (#38913) pcre2-cet-flags.patch lacks some trailing whitespace that exists in one of the target files. Consequently, the patch does not apply when the patch tool is strict about whitespace. This teensy diff makes the whitespace in pcre2-cet-flags.patch strictly match the whitespace in the target file. Co-Authored-By: Nathan Daly <NHDaly@gmail.com> Co-authored-by: Nathan Daly <NHDaly@gmail.com> | 17 December 2020, 02:08:36 UTC |
a56af7d | Katharine Hyatt | 17 December 2020, 02:06:18 UTC | Missing tests for cmp on UInt8s (#38914) | 17 December 2020, 02:06:18 UTC |
8f6432e | Jeff Bezanson | 16 December 2020, 19:36:43 UTC | fix #34170, assertion failure with obvious_subtype (#38904) | 16 December 2020, 19:36:43 UTC |
9e52bdf | Steven G. Johnson | 16 December 2020, 17:50:20 UTC | tab-complete \ldq and \rdq for “ and ” (#38860) | 16 December 2020, 17:50:20 UTC |
cc1623b | Tim Holy | 16 December 2020, 13:03:20 UTC | Internalize Revise precompiles into Base For some reason (perhaps #32705?) most or all of these fail if they are emitted as precompile statements, so this moves them into Base itself. This drops the time for a revision down to 1.85s. | 16 December 2020, 13:03:20 UTC |
4c4e078 | Tim Holy | 16 December 2020, 11:56:52 UTC | Add precompiles to reduce time to first Revise.revise() Perhaps the most annoying thing about Revise now is that the first revision is quite slow, about 3.1s on my machine. This PR drops the time to about 2.4s. Basically the idea is to precompile statements that Revise will need. Discovered via the new snoopi_deep/Core.Compiler.Timings framework. | 16 December 2020, 11:56:52 UTC |
f184f05 | Katharine Hyatt | 16 December 2020, 07:58:02 UTC | add a test for popat with default (#38903) | 16 December 2020, 07:58:02 UTC |
fbaf7a8 | Keno Fischer | 16 December 2020, 05:13:27 UTC | Allow DCE of pure functions that are not inlined (#38130) E.g. because the return value is too large. | 16 December 2020, 05:13:27 UTC |
ce338db | Mustafa M | 16 December 2020, 04:39:57 UTC | Add section regarding Pkg mode to REPL docs (#38842) * Add section regarding Pkg mode to REPL docs Co-authored-by: Stephen Vavasis <vavasis@uwaterloo.ca> | 16 December 2020, 04:39:57 UTC |
341d6c9 | Shuhei Kadowaki | 16 December 2020, 04:14:23 UTC | rm unnecessary `widenconst_bestguess` call (#38871) the equivalent widening logic is imposed within `typeinf_local`, and so as far as I understand we don't this need `widenconst_bestguess` in `typeinf_edge` | 16 December 2020, 04:14:23 UTC |
1df2c11 | Keno Fischer | 16 December 2020, 04:13:24 UTC | Turn off client-side port reuse on Darwin (#38901) For scalability, in the Distributed code and when supported by the operating system, we bind all client sockets to the same port (the server ports still differ). In general, each tcp connection is identified by the 4-tuple (source_ip, source_port, dest_ip, dest_port) (also known as a 5-tuple if the protocol is included). Re-using the client port saves on the number of available ports, but since the 5-tuple is still different due to the varying server ports, it doesn't end up causing any trouble. However, on Darwin, we run into a bit of a pickle. When a connection exits, the server side of the socket enters TIME_WAIT state, while the client side of the socket immediately enters CLOSED state allowing the port to be reused (if we weren't setting SO_REUSEPORT anyway). Now, ordinarily the server port number is not allowed to be re-used if there is any remaining PCB (Process control block, basically the kernel data structure associated to a 5-tuple), that references said port (including those in TIME_WAIT state). However, this is very annoying for servers in general (since it would mean that servers can't be restarted for some number of seconds until after the last connection terminated), so it is common to pass SO_REUSEADDR on server sockets. In fact, libuv does that for us automatically. Unfortunately, as a result, there is nothing that prevents us from re-using the same 5-tuple before the TIME_WAIT state has timed out (the client just got immediately recycled, and we explicitly bypassed the TIME_WAIT restriction on the server). Linux appears to handle this fine, but Darwin does not. On Darwin, when we try to connect with a re-used 5-tuple, there is some chance (depending on the kernel hash table for PCBs), that we will get the state PCB rather than the fresh one, causing the connection to drop. This is the cause of #38812 and our mac CI reliability issued in the Distributed test. It is worth pointing out that this is not directly related to SO_REUSEPORT. Since the client port is immediately available for recycle, it is perfectly legal for us to re-use that port immediately, even if SO_REUSEPORT is not set (which would exhibit the same problem). However, because doing that is not particularly useful, we just randomize the port if SO_REUSEPORT is not set. This doesn't fix the situation either, but should hopefully reduce the incidence rate below the rate at which it is problematic for CI. We could turn off SO_REUSEADDR on the server entirely, but that is undesirable for reasons mentioned above. Since Linux does not have the same issue, I have some hope that Apple might consider this a bug and adjust the behavior. I have an open support request with them to that extent. However, in the meantime this will hopefully help our CI reliability. | 16 December 2020, 04:13:24 UTC |
2362037 | Steven G. Johnson | 16 December 2020, 03:50:20 UTC | update to utf8proc 2.6.1 (#38900) | 16 December 2020, 03:50:20 UTC |
0eb4723 | Keno Fischer | 16 December 2020, 02:04:26 UTC | Improve apply_type_nothrow precision (#38071) | 16 December 2020, 02:04:26 UTC |
da38c6b | Henrique Becker | 15 December 2020, 20:03:16 UTC | Note that `&&` and `||` are `and` and `or`. (#37463) Suggested by this post in Discourse: https://discourse.julialang.org/t/and-or-bitwise-or-shortcircuit-what-we-get-searching-docs/46263 | 15 December 2020, 20:03:16 UTC |
74a08fe | Matt Bauman | 15 December 2020, 19:53:26 UTC | fix #38897, docstring for .op= (#38898) Also change "equivalent to" to a more rough "akin to" since `x .op y` isn't always identical to `broadcast(op, x, y)`. | 15 December 2020, 19:53:26 UTC |
fa6077e | Ian Butterworth | 15 December 2020, 19:38:51 UTC | measure compile time only when using time macros (fix #38877) (#38885) | 15 December 2020, 19:38:51 UTC |
ac1c0c1 | Katharine Hyatt | 15 December 2020, 19:36:38 UTC | Test isstored (#38894) | 15 December 2020, 19:36:38 UTC |
c28611f | Jeff Bezanson | 15 December 2020, 19:31:11 UTC | fix #38837, inference regression in tuple `map` (#38887) | 15 December 2020, 19:31:11 UTC |
bd0ddf3 | Miguel Raz Guzmán Macedo | 15 December 2020, 15:54:10 UTC | Add doctest for IdDict with example (#38881) This came up from a Slack discussion and was based on a surprising [Python behavior](https://twitter.com/DahlitzF/status/1338384990040682498). @haampie provided the example and I thought the docstrings could help out another beginner to understand why a different data structure is justified. | 15 December 2020, 15:54:10 UTC |
4dede6d | Stefan Karpinski | 15 December 2020, 14:38:13 UTC | LibGit2: improve error when CA root cert can't be set (#38827) This also fixes an insecure behavior: even if `set_ssl_cert_locations` failed, `REFCOUNT` was still incremented, which meant that subsequent calls to `ensure_initialized` didn't call `initialize` and so there was never a successful call to `set_ssl_cert_locations`. Without this libgit2 defaults to not verifying host identities and that is insecure. To prevent this, this patch locks on `ensure_initialized` and decrements `REFCOUNT` if initialize throws an error, ensuring that `initialize` succeeds at least once, including the call to `set_ssl_cert_locations`. | 15 December 2020, 14:38:13 UTC |
5bf1665 | Valentin Churavy | 15 December 2020, 13:55:59 UTC | Merge pull request #38883 from JuliaLang/vc/cmdlineargs_ppc Don't override target passed on command-line with target from sysimg for processor_fallback | 15 December 2020, 13:55:59 UTC |
2da5cb2 | Valentin Churavy | 15 December 2020, 03:27:41 UTC | Don't override target passed on command-line with target from sysimg | 15 December 2020, 03:27:52 UTC |
80ace52 | Keno Fischer | 15 December 2020, 02:48:25 UTC | Make Vararg not a DataType (#38136) Currently `Vararg` is a DataType, but is special cased in a bunch of places to give it special behavior (e.g. in subtyping and of course in tuple construction). However, unlike all other DataTypes, it cannot appear as a type parameter, which caused trouble in PR #38071. Having it be a DataType is a bit of a pun of convenience in the first place - it's a lot more similar to a tvar (which can be considered an implementation detail of UnionAll in the same way Vararg is an implementation detail of Tuple), which has its own non-type object. This PR does the same to Vararg, and moves it from being an abstract DataType with special cased behavior to its own custom type (called `Core.TypeofVararg`). There are a few small behavior differences, but they are mostly internal. In particular, we no longer have `Vararg <: Type` and Vararg objects no longer have the .parameters field. Also, things like `Vararg{T} where T` are technically illegal now since Vararg is not a type. However, since a lot of people are using that pattern, I've brought it back with a deprecation (which is of course off by default). The only things that's disallowed is `Vararg{N, N} where N`, but I haven't seen anybody use that. | 15 December 2020, 02:48:25 UTC |
b376982 | Keno Fischer | 15 December 2020, 00:10:47 UTC | Don't clobber rax in x86_64 trampoline (#38882) rax is used to pass vararg metadata. r11 can be used for dynamic linker shenanigans between functions. Messing with rax here can cause `jl_` to print nonsense. | 15 December 2020, 00:10:47 UTC |
8d80154 | Valentin Churavy | 14 December 2020, 23:26:14 UTC | Merge pull request #38839 from JuliaLang/jn/38838 win: skip bad paths in dllist | 14 December 2020, 23:26:14 UTC |
e65515e | strickek | 14 December 2020, 21:50:02 UTC | Shows evaluated variables in tests that use `contains` (#38817) Co-authored-by: strickek <klaus@stricker.at> | 14 December 2020, 21:50:02 UTC |
6c95660 | Valentin Churavy | 14 December 2020, 19:46:23 UTC | Merge pull request #38875 from JuliaLang/vc/ppc_artifacts [Artifacts] Test c_simple on ppc | 14 December 2020, 19:46:23 UTC |
b6c09a3 | Valentin Churavy | 14 December 2020, 19:46:09 UTC | Merge pull request #38850 from JuliaLang/vc/ppc_fastmath [PPC] relax fastmath test | 14 December 2020, 19:46:09 UTC |
8c01444 | Jameson Nash | 14 December 2020, 19:23:45 UTC | inference: stop re-converging worlds after optimization (#38820) The validity did not change, so we should not need to update it. This also ensures we copy over all result information earlier, so we can destroy the InferenceState slightly sooner, and slightly cleaner data flow. | 14 December 2020, 19:23:45 UTC |
2ff110b | Jameson Nash | 14 December 2020, 18:01:41 UTC | optimize invokelatest call performance (#38835) Applications typically shouldn't use this function in performance sensitive places, as it hints that their design is flawed, but might as well make it faster anyways. and optimize invokelatest kwcall too, while we are at it | 14 December 2020, 18:01:41 UTC |
33cfcc4 | Valentin Churavy | 14 December 2020, 17:57:28 UTC | Merge pull request #37723 from JuliaLang/vc/use_system Fix cross-compilation of libjulia | 14 December 2020, 17:57:28 UTC |
6322b8c | Shuhei Kadowaki | 14 December 2020, 17:55:37 UTC | fix erroneous code path within `show_sym` (#38830) fix erroneous code path within `show_sym` currently, `show_sym` can recur into `show(::IO, String)`, but it's a bit "erroneous" since `is_syntactic_operator(::String)` is not defined. | 14 December 2020, 17:55:37 UTC |
0d229f8 | Shuhei Kadowaki | 14 December 2020, 17:54:46 UTC | improve various minor inferrabilities (#38832) they will help inference avoid entering into frames with `Nothing` args, and may improve codegen even a bit use `AbstractString` for rubustness against future changes of `RegexMatch` | 14 December 2020, 17:54:46 UTC |
6ef077c | Jeff Bezanson | 14 December 2020, 17:49:33 UTC | fix #38423, another stack overflow in method definition (#38810) also fixes #36544 and fixes #36804 | 14 December 2020, 17:49:33 UTC |
d3e48cf | Valentin Churavy | 14 December 2020, 16:39:14 UTC | [Artifacts] Test c_simple on ppc | 14 December 2020, 16:39:14 UTC |
17f9d60 | Valentin Churavy | 14 December 2020, 16:33:22 UTC | [OpenBLAS_jll] use normalized ppc arch (#38874) | 14 December 2020, 16:33:22 UTC |
b1dbee0 | Valentin Churavy | 12 December 2020, 18:00:55 UTC | Remove fastmath tests that are defined to be undef/poison | 14 December 2020, 15:43:14 UTC |
0c256d3 | Valentin Churavy | 14 December 2020, 15:39:35 UTC | Merge pull request #38863 from JuliaLang/vc/mmap_ppc [Mmap] Skip ReadOnlyMemoryError test on PowerPC | 14 December 2020, 15:39:35 UTC |
6ddc7f1 | DilumAluthgeBot | 13 December 2020, 23:16:05 UTC | [automated] Bump the Downloads stdlib from a6e4926 to 09f18ad (#38861) Co-authored-by: Dilum Aluthge <dilum@aluthge.com> | 13 December 2020, 23:16:05 UTC |
1044621 | Viral B. Shah | 13 December 2020, 19:55:50 UTC | Merge branch 'zierenberg-master' | 13 December 2020, 19:55:50 UTC |
9d6e3f2 | Stefan Karpinski | 22 May 2020, 19:44:08 UTC | Update div.jl | 13 December 2020, 19:51:58 UTC |
7722d6b | Stefan Karpinski | 22 May 2020, 19:42:52 UTC | Update base/div.jl | 13 December 2020, 19:51:58 UTC |
2a97fd9 | Stefan Karpinski | 22 May 2020, 19:42:45 UTC | Update base/div.jl | 13 December 2020, 19:51:58 UTC |
99cb5da | Johannes Zierenberg | 05 March 2020, 18:06:33 UTC | documentation fld(x,y) in div.jl corrected | 13 December 2020, 19:51:58 UTC |
888df7a | Johannes Zierenberg | 05 March 2020, 08:55:40 UTC | update documentation of fld(x,y) in div.jl Added an example with explanation why floating-point rounding can lead to an alleged violation of the fld(x,y)==x/y condition | 13 December 2020, 19:51:58 UTC |
d34cc20 | Mustafa M | 13 December 2020, 19:45:23 UTC | Use https links in a doc link (#38866) Co-authored-by: EternalLearner42 <46832510+EternalLearner42@users.noreply.github.com> | 13 December 2020, 19:45:23 UTC |
9f4a807 | Valentin Churavy | 13 December 2020, 19:32:33 UTC | Adjust calling convention of LAPACK functions (#38836) * Adjust calling convetion of dgebal and co * Adjust calling convention of BLAS & LAPACK | 13 December 2020, 19:32:33 UTC |
a6eee98 | Valentin Churavy | 13 December 2020, 04:33:46 UTC | [Mmap] Skip ReadOnlyMemoryError test on PowerPC | 13 December 2020, 04:33:46 UTC |
8f2b0e7 | Valentin Churavy | 13 December 2020, 00:55:52 UTC | Use DEPOT_PATH instead of HOME in cmdlineargs test (#38855) | 13 December 2020, 00:55:52 UTC |
ddd977d | Valentin Churavy | 12 December 2020, 20:57:04 UTC | Merge pull request #38853 from JuliaLang/aa/checksum Commit checksums for libosxunwind | 12 December 2020, 20:57:04 UTC |
e10d7a3 | Alex Arslan | 12 December 2020, 19:06:54 UTC | Commit checksums for libosxunwind This avoids a dirty repo state after building on macOS. | 12 December 2020, 19:06:54 UTC |
b3eaa34 | Shuhei Kadowaki | 12 December 2020, 09:30:00 UTC | improve inferrabilities within TOML module (#38831) * improve inferrability within TOML module * simplify with `@try` macro * apply suggestion, use `Int64` | 12 December 2020, 09:30:00 UTC |
bc14e28 | Valentin Churavy | 12 December 2020, 05:43:58 UTC | Merge pull request #38798 from JuliaLang/vc/ppc [LLVM] add patch for miscompilation with setjmp on PPC | 12 December 2020, 05:43:58 UTC |
649c18a | Valentin Churavy | 11 December 2020, 22:30:45 UTC | Update libLLVM, LLVM, and Clang | 12 December 2020, 04:06:32 UTC |
d00c851 | Elliot Saba | 12 December 2020, 04:01:16 UTC | Fix bugs in the checksum generator | 12 December 2020, 04:06:11 UTC |
fe141ca | Valentin Churavy | 10 December 2020, 22:10:21 UTC | [LLVM] add patch for #38773 | 12 December 2020, 03:27:35 UTC |
7f7fa18 | Valentin Churavy | 09 December 2020, 15:13:52 UTC | [LLVM] add patch for miscompilation with setjmp on PPC | 12 December 2020, 03:27:35 UTC |
7d137e9 | Artem Pelenitsyn | 12 December 2020, 00:14:55 UTC | devdocs: slots live in CodeInfo, not MethodInstance (fix #38840) (#38843) | 12 December 2020, 00:14:55 UTC |
26b73d6 | Leandro Martínez | 11 December 2020, 21:38:01 UTC | Accept AbstractChar in count and findall (::Char, ::String) (#38675) This change solves an issue raised in this post, in which searching for a char in a string results in a Method error: https://discourse.julialang.org/t/method-for-counting-characters-in-string-count-a-batman-vs-count-a-batman/51153 Co-authored-by: Stefan Karpinski <stefan@karpinski.org> Co-authored-by: Steven G. Johnson <stevenj@mit.edu> | 11 December 2020, 21:38:01 UTC |
3d3cf1b | Sebastian Stock | 11 December 2020, 21:20:24 UTC | range: avoid int overflow in type computation (correction to #38745) (#38814) | 11 December 2020, 21:20:24 UTC |