0a5f35b | Sacha Verweij | 25 February 2017, 03:45:18 UTC | Merge pull request #20719 from Sacha0/docdepvecops Remove deprecated vectorized operators/functions from the manual | 25 February 2017, 03:45:18 UTC |
6ac7e24 | Steven G. Johnson | 25 February 2017, 00:48:50 UTC | inline x^literal only for hardware-based number types x (closes #20768) (#20782) | 25 February 2017, 00:48:50 UTC |
581a264 | Jeff Bezanson | 24 February 2017, 23:41:46 UTC | Merge pull request #20352 from JuliaLang/jb/fix20332 fix #20332, `MethodInstance` display error | 24 February 2017, 23:41:46 UTC |
f6f7d8a | Sacha Verweij | 21 February 2017, 18:15:51 UTC | Remove deprecated array/vectorized operators/functions from the manual. Also revise that section. [ci skip] | 24 February 2017, 19:31:59 UTC |
d321300 | H-225 | 24 February 2017, 17:19:36 UTC | Improve UX of calling abstract-type constructors (#20596) * Make calling abstract type with only sysimg method display a more specific message * Do not show the Base/sysimg.jl method for Type{T} in tab-completion * Hide sysimg.jl method from closest candidates * Better type name in tests * Remove line number of sysimg.jl method from tests * Fix tests not checking for qualified names of the type * *Actually* add a test for qualified abstract types * Improve tests - Wrap long lines. - Make them easier to follow. - Fix test for tab-completion * Should be checking if the method string *starts with* the sample, and not `!=` to it. * Refactor sysimg method check * Add a comment to the sysimg.jl method for clarity * Cleanup `is_default_method` * Fix tests; well that was easy... - since the 'default' method from sysimg.jl is now hidden, there was one less method candidate, and therefore exactly three, and therefore, no elipsis anymore. - removed duplicate test for error message displayed for a MethodError on an abstract type without any constructors defined. * Add comments to the tests; make them easier to understand * Update doctests that fail due to now-hidden sysimg.jl default method * Fix rogue syntax error with '&&' * Use @test instead of @assert * Fix use of old unfactored code | 24 February 2017, 17:19:36 UTC |
dce17b6 | Jeff Bezanson | 31 January 2017, 19:53:29 UTC | fix #20332, `MethodInstance` display error | 24 February 2017, 17:17:05 UTC |
ddd40d9 | Jameson Nash | 24 February 2017, 16:57:12 UTC | Merge pull request #19890 from JuliaLang/jn/19872 only use accurate powf function | 24 February 2017, 16:57:12 UTC |
1e44282 | Jeff Bezanson | 24 February 2017, 15:43:51 UTC | Merge pull request #20777 from JuliaLang/jb/FEfixes fix two front-end issues | 24 February 2017, 15:43:51 UTC |
00b7b60 | Curtis Vogt | 24 February 2017, 15:40:37 UTC | Handle systems where `hostname` is unavailable (#20765) * Handle systems where `hostname` is unavailable Fixes #20758 * Comment on hostname regex * Fix grammar | 24 February 2017, 15:40:37 UTC |
078835c | Fengyang Wang | 24 February 2017, 14:15:05 UTC | Document rounding away from zero (#20775) | 24 February 2017, 14:15:05 UTC |
b0ca50b | Tim Holy | 24 February 2017, 11:20:35 UTC | Merge pull request #20712 from JuliaLang/teh/doc_cartesian Add CartesianIndex & CartesianRange docstrings, misc doc improvements | 24 February 2017, 11:20:35 UTC |
0dca594 | Stefan Karpinski | 24 February 2017, 08:30:44 UTC | hashing of general objects was far too simple (fixes #20744) (#20767) * hashing of general objects was far too simple (fixes #20744) * hash_uint(3h - object_id(x)) is a somewhat better hash If we'd used `3h - object_id(x)` previously, the original issue would never have come up since `hash(a, hash(b, hash(c, h)))` wouldn't have produced `3A - 3B + 3C - h` but would instead have given the value `27h -9A - 3B - C`, which is asymmetrical in `a` and `c`. Passing the result to `hash_uint` fixes that issue since that function is non-linear, but defense in depth, right? We may as well do the same operations in such a way that the passed-through value gets the most operations done on it. | 24 February 2017, 08:30:44 UTC |
70fc72a | Jeff Bezanson | 24 February 2017, 05:37:30 UTC | Merge pull request #20773 from JuliaLang/jb/typeshowbug fix bug where non-tuple types were shown as `NTuple` | 24 February 2017, 05:37:30 UTC |
dfe2dfb | Jeff Bezanson | 24 February 2017, 04:14:39 UTC | fix #20000, parsing of parameter syntax in macro calls: `@m(a; b=c)` | 24 February 2017, 04:14:39 UTC |
391c744 | Jeff Bezanson | 24 February 2017, 04:12:04 UTC | fix error reporting from `julia_to_scm` In this context we need to throw flisp errors instead of julia errors. This was causing corruption, plus error expressions got replaced with `invalid syntax` errors. | 24 February 2017, 04:12:04 UTC |
3ccfd1c | Jeff Bezanson | 24 February 2017, 00:52:31 UTC | Merge pull request #20734 from JuliaLang/teh/man_amb_fix Be consistent about function names in padding example | 24 February 2017, 00:52:31 UTC |
f95bc77 | Jeff Bezanson | 24 February 2017, 00:51:40 UTC | Merge pull request #20757 from iblis17/bsd-sharedarr sharedarray: improve error msg on bsd family | 24 February 2017, 00:51:40 UTC |
554a031 | Jeff Bezanson | 24 February 2017, 00:49:31 UTC | Merge pull request #20756 from iblis17/mmap-anon mmap: fix MAP_ANONYMOUS flags value for bsd family | 24 February 2017, 00:49:31 UTC |
c49908c | Jeff Bezanson | 23 February 2017, 21:59:56 UTC | fix bug where non-tuple types were shown as `NTuple` | 23 February 2017, 21:59:56 UTC |
7968717 | Andreas Noack | 23 February 2017, 21:08:52 UTC | Make setindex! work for views of BitArrays by loosening signature (#20717) for Arrays to StridedArrays. This is not the fastest solution but it should be correct. Since StridedArrays are not defined when bitarray.jl is loaded, the relevant defitions are moved to multidimensional.jl Fixes #20715 | 23 February 2017, 21:08:52 UTC |
468b610 | Jeff Bezanson | 23 February 2017, 20:16:01 UTC | Merge pull request #20750 from JuliaLang/jb/randomfixes fixes for 4 miscellaneous issues | 23 February 2017, 20:16:01 UTC |
0c5eac2 | Jameson Nash | 23 February 2017, 19:10:23 UTC | make sure that the indirection through the `Val{p}` type doesn't stop inlining of `^` | 23 February 2017, 19:10:45 UTC |
cc7990b | Jameson Nash | 23 February 2017, 18:10:52 UTC | make several pure-marked math functions actually pure | 23 February 2017, 18:52:29 UTC |
d3ee10b | Fredrik Ekre | 23 February 2017, 18:37:15 UTC | Dict merge(!) with combiner function (#20708) | 23 February 2017, 18:37:15 UTC |
4b76c95 | Fredrik Ekre | 23 February 2017, 18:31:45 UTC | Goodbye, Human (#20748) | 23 February 2017, 18:31:45 UTC |
930c2c1 | Katharine Hyatt | 23 February 2017, 18:17:53 UTC | Merge pull request #20730 from JuliaLang/ksh/diffstat Add a DiffStat type, methods, and tests | 23 February 2017, 18:17:53 UTC |
4275c74 | Tony Kelman | 23 February 2017, 17:29:32 UTC | Fix doc build in cygwin cross-compile (#20713) need cygpath_w to translate absolute paths from cygwin style to windows for mingw julia executable to understand them fixes buildbot failure seen at https://build.julialang.org/builders/package_win6.2-x64/builds/1346/steps/make%20binary-dist/logs/stdio was broken by me in #20519, didn't notice until now because of #20643 preventing the buildbots from making new nightlies | 23 February 2017, 17:29:32 UTC |
f2b155a | Tony Kelman | 23 February 2017, 17:24:43 UTC | Merge pull request #20025 from JuliaLang/yyc/codegen/llvm-stall Backport LLVM patches to fix X86 partial register stall | 23 February 2017, 17:24:43 UTC |
a17b138 | Simon Byrne | 23 February 2017, 16:58:38 UTC | Expand docs for mod (#20633) * Expand docs for mod Address #14826. * simplify doc text * Another rewrite * more changes * typo * some more * missing the | 23 February 2017, 16:58:38 UTC |
0b2130e | Sacha Verweij | 23 February 2017, 16:24:44 UTC | Revise mention of deprecated array-scalar methods in a news item, indicating their deprecation. [ci skip] (#20743) | 23 February 2017, 16:24:44 UTC |
162a61d | Iblis Lin | 23 February 2017, 09:45:53 UTC | sharedarray: improve error msg on bsd family | 23 February 2017, 13:23:08 UTC |
f1adcc8 | Iblis Lin | 23 February 2017, 09:01:27 UTC | mmap: fix MAP_ANONYMOUS flags value for bsd family | 23 February 2017, 09:01:27 UTC |
12624fb | Alex Arslan | 23 February 2017, 05:41:28 UTC | Merge pull request #20674 from iblis17/test-file Fix FileMonitor test case on FreeBSD | 23 February 2017, 05:41:28 UTC |
9e947e5 | Jeff Bezanson | 23 February 2017, 05:01:14 UTC | fix #20610, add deprecation for `=>` expression head | 23 February 2017, 05:01:14 UTC |
03939e7 | Jeff Bezanson | 23 February 2017, 03:36:39 UTC | fix #20745, handle macrocall in `deparse` | 23 February 2017, 04:23:25 UTC |
51f7a84 | Jeff Bezanson | 23 February 2017, 03:35:59 UTC | fix #20729, handle null in ASTs better | 23 February 2017, 04:23:25 UTC |
2d5a51d | Jeff Bezanson | 23 February 2017, 03:35:26 UTC | fix #20733, inference with new definition of `getindex` for tuples | 23 February 2017, 04:23:21 UTC |
2a19c36 | Katharine Hyatt | 22 February 2017, 23:17:37 UTC | Add show method and test for FetchHead (#20727) | 22 February 2017, 23:17:37 UTC |
e4be955 | Katharine Hyatt | 22 February 2017, 22:23:14 UTC | Merge pull request #20721 from JuliaLang/ksh/showmethefactors More show methods and tests for factorizations | 22 February 2017, 22:23:14 UTC |
0cebddf | Steven G. Johnson | 22 February 2017, 20:19:42 UTC | bug fix for x^Val{p} (fixes problem noted in #20648) (#20732) | 22 February 2017, 20:19:42 UTC |
1c5d733 | Jameson Nash | 22 February 2017, 19:58:46 UTC | only use accurate powf function The powi intrinsic optimization over calling powf is that it is inaccurate. We don't need that. When it is equally accurate (e.g. tiny constant powers), LLVM will already recognize and optimize any call to a function named `powf`, and produce the same speedup. fix #19872 | 22 February 2017, 20:01:19 UTC |
5a1f971 | Jameson Nash | 22 February 2017, 20:00:06 UTC | update DomainError showerror handling to ignoring inlining | 22 February 2017, 20:00:11 UTC |
fa167f5 | Jameson Nash | 22 February 2017, 19:45:04 UTC | Merge pull request #20593 from JuliaLang/jn/union-codegen codegen support for efficient union representations | 22 February 2017, 19:45:04 UTC |
f3961b9 | Alex Arslan | 22 February 2017, 19:14:35 UTC | Merge pull request #20676 from iblis17/fbsd-term Fix Terminals.hascolor checking on FreeBSD | 22 February 2017, 19:14:35 UTC |
4d2efd8 | Sacha Verweij | 22 February 2017, 18:54:07 UTC | Mention devectorization deprecations in NEWS.md. (#20724) | 22 February 2017, 18:54:07 UTC |
0234645 | Katie Hyatt | 21 February 2017, 19:02:47 UTC | Show method and test for Schur | 22 February 2017, 18:35:31 UTC |
603726c | Tim Holy | 21 February 2017, 12:03:04 UTC | Be more explicit about data sharing for `reshape` | 22 February 2017, 11:47:48 UTC |
e399558 | Tim Holy | 21 February 2017, 12:02:37 UTC | Add Cartesian docstrings to stdlib and reorganize some sections | 22 February 2017, 11:47:43 UTC |
0e08411 | Tim Holy | 22 February 2017, 11:05:59 UTC | Be consistent about function names in padding example | 22 February 2017, 11:05:59 UTC |
61a291b | Jeff Bezanson | 22 February 2017, 04:15:47 UTC | Merge pull request #20722 from JuliaLang/jb/nontypebounds tolerate non-types as Tuple parameters better | 22 February 2017, 04:15:47 UTC |
ac7459d | Katie Hyatt | 22 February 2017, 00:49:21 UTC | Add a DiffStat type, methods, and tests Create a DiffStat type so that we can get more information about GitDiffs. Add a show method for this type and GitDiff. Also add methods to get the number of files changed/insertions/ deletions in the diff. | 22 February 2017, 00:57:37 UTC |
35ce0fd | Alex Arslan | 22 February 2017, 00:52:40 UTC | Remove duplicate ctranspose definition for real vectors (#20685) | 22 February 2017, 00:52:40 UTC |
d168287 | Katharine Hyatt | 22 February 2017, 00:17:46 UTC | Merge pull request #20705 from JuliaLang/ksh/diffshow Add show methods for DiffDelta, DiffFile and test | 22 February 2017, 00:17:46 UTC |
eac043a | Jameson Nash | 21 February 2017, 18:22:00 UTC | add union-splitting commit comment to devdocs starting a document describing the overall code-generator structure | 21 February 2017, 19:57:17 UTC |
e8d2d95 | Jeff Bezanson | 21 February 2017, 19:11:51 UTC | tolerate non-types as Tuple parameters better part of #19998 | 21 February 2017, 19:11:51 UTC |
6c34f8f | Katie Hyatt | 21 February 2017, 19:02:39 UTC | Test of show method for QR | 21 February 2017, 19:02:39 UTC |
6e317a4 | Katie Hyatt | 21 February 2017, 19:02:26 UTC | Test of show method for Cholesky | 21 February 2017, 19:02:26 UTC |
1b30ff8 | Katie Hyatt | 21 February 2017, 19:02:14 UTC | Show method and test for LQ | 21 February 2017, 19:02:14 UTC |
3f7de0f | kshyatt | 21 February 2017, 00:26:14 UTC | Add show methods for DiffDelta, DiffFile and test | 21 February 2017, 18:24:59 UTC |
474d05d | Yichao Yu | 21 February 2017, 18:13:14 UTC | Fix misc test to not hard code test module name (#20550) Use randomized test module name to prevent this from happening again. | 21 February 2017, 18:13:14 UTC |
33f464e | Katharine Hyatt | 21 February 2017, 17:48:16 UTC | Tests for getindex on diff and move error throw (#20683) | 21 February 2017, 17:48:16 UTC |
6131a0f | Jameson Nash | 21 February 2017, 02:41:58 UTC | better optimize code for copying a value to an unboxed union | 21 February 2017, 17:46:07 UTC |
cc43c0b | Jameson Nash | 16 February 2017, 22:20:37 UTC | ensure that constants are treated as values rather than opaque pointers | 21 February 2017, 17:46:07 UTC |
cad430c | Jameson Nash | 16 February 2017, 20:59:53 UTC | ensure the value in the gcroot is transferred during union conversion convert_julia_type is only ever called with needsgcroot=false, so this is mostly dead-code right now | 21 February 2017, 17:46:06 UTC |
93fcde0 | Jameson Nash | 16 February 2017, 04:43:42 UTC | fix over-pessimistic emit_isa for Type | 21 February 2017, 17:46:06 UTC |
a915e0b | Jameson Nash | 15 February 2017, 21:42:41 UTC | change union tindex representation to have the high bit indicate boxed this allows (tindex & 0x7f) to always be exact, saving the need for any dynamic loads in many cases | 21 February 2017, 17:46:05 UTC |
bbe8e40 | Jameson Nash | 15 February 2017, 17:37:21 UTC | distinguish between "must" and "may" when remarking julia type information expr-type and slot-type may improve type information (but try not to destroy it) while slot-assignment and return must convert the type information (even if it loses information / forces boxing) | 21 February 2017, 17:46:05 UTC |
de6ad69 | Jameson Nash | 08 February 2017, 04:19:26 UTC | implement a union-return-type calling convention | 21 February 2017, 17:46:04 UTC |
00b30a8 | Jameson Nash | 23 January 2017, 23:32:43 UTC | codegen: represent Union types via tagged unions this covers the code-generator support for representing types on the stack via a tagged type representation The function `remark_julia_type` now can transform between any two types. Instead of asserting when asked to drop type information, it'll instead cast the object to the requested representation. It'll make boxes, allocate stack copies, and compute tagged unions as needed to perform the request. This means that we can use `remark_julia_type` whenever we want type information to forcibly change. For example, storing to a slot requires dropping the intermediate type information and generating a representation suitable for copying into the memory location. Loading from a slot has the opposite requirement: it needs to be able to load the slot via it's declared type, then possibly to cast the result to add type information that may be used later (to emit optimizations of other intrinsic function calls). The primitive routines that need to be able to handle tagged unions are: - remark-type - load-local - store-local - isa - is - emit_typeof - emit_sizeof - boxed - unbox - specsig (not implemented here yet) Everything else should be possible to handle in inference by using these primitives to implement union-splitting. Additionally, the following will be supported once the type layout changes: - getfield / setfield The representation of the tagged-union is as a pair of < union*, selector > memory addresses. The selector is fixed as a byte, and will union-tag the first 126 isbits. It records the one-based depth-first count into the type-union of the isbits objects inside. An index of zero indicates that the union* is actually a tagged heap-allocated jl_value_t*, and needs to treated as normal for a boxed object rather than as a tagged union. The `union*` memory region may be allocated at *any* size. The only constraint is that it is big enough to contain the data currently specified by `selector`. It might not be big enough to contain the union of all types that could be stored there according to the associated Union type field. Use appropriate care when copying. this also includes improvements for codegen of isa with mixed types | 21 February 2017, 17:46:03 UTC |
ae77740 | Keno Fischer | 21 February 2017, 17:11:50 UTC | Merge pull request #20700 from JuliaLang/kf/safepoint Fix missing root across safe-point | 21 February 2017, 17:11:50 UTC |
9d96056 | Katharine Hyatt | 21 February 2017, 16:59:08 UTC | Fix fetchheads callback and add tests (#20669) | 21 February 2017, 16:59:08 UTC |
201737a | Tony Kelman | 21 February 2017, 16:45:45 UTC | Add imports and module qualifiers for Distributed (#20710) * Add imports and module qualifiers for Distributed helps more of `JULIA_TESTFULL=1 make testall` pass * some more imports from Base into Distributed for use * fix rmprocs test to wait for removal of added workers | 21 February 2017, 16:45:45 UTC |
0e33baa | Jeff Bezanson | 21 February 2017, 15:37:38 UTC | Merge pull request #20706 from JuliaLang/jb/fixcxx fix specificity of Tuples with non-types in them | 21 February 2017, 15:37:38 UTC |
e1b88f7 | Stefan Karpinski | 21 February 2017, 14:59:25 UTC | Merge branch 'master' into yyc/codegen/llvm-stall | 21 February 2017, 14:59:25 UTC |
bb1f704 | Tim Holy | 21 February 2017, 12:02:15 UTC | Add docstrings for CartesianIndex and CartesianRange | 21 February 2017, 12:02:15 UTC |
bb83c0e | Jeff Bezanson | 21 February 2017, 06:18:00 UTC | fix specificity of Tuples with non-types in them | 21 February 2017, 06:18:00 UTC |
e21caee | Tim Holy | 21 February 2017, 00:47:05 UTC | Merge pull request #20702 from JuliaLang/teh/fix_extra_manual Delete accidental revival of methods.rst [ci skip] | 21 February 2017, 00:47:05 UTC |
4a60ad2 | Tim Holy | 20 February 2017, 23:51:49 UTC | Delete accidental revival of methods.rst [ci skip] | 20 February 2017, 23:51:49 UTC |
30a215d | Tim Holy | 20 February 2017, 23:35:09 UTC | Merge pull request #20686 from JuliaLang/teh/linear_indices_amb Minor fixes to checkbounds_linear_indices and getindex for ranges | 20 February 2017, 23:35:09 UTC |
75cfcb9 | Tim Holy | 20 February 2017, 23:34:27 UTC | Merge pull request #18111 from JuliaLang/teh/docs_methods_ambiguities Add manual section on method design and more advanced ambiguity resolution | 20 February 2017, 23:34:27 UTC |
3181500 | Keno Fischer | 20 February 2017, 22:31:37 UTC | Fix missing root across safe-point | 20 February 2017, 23:09:49 UTC |
fe2a3b5 | Jeff Bezanson | 20 February 2017, 22:46:52 UTC | Merge pull request #20698 from JuliaLang/jb/20643 fix #20643, stack overflow in type intersection | 20 February 2017, 22:46:52 UTC |
baf1900 | Alex Arslan | 20 February 2017, 22:33:17 UTC | Merge pull request #20675 from iblis17/fbsd-libdl Libdl: implement dllist on FreeBSD | 20 February 2017, 22:33:17 UTC |
c102fc2 | Kristoffer Carlsson | 20 February 2017, 22:32:56 UTC | update profile manual (#20661) * update profile manual | 20 February 2017, 22:32:56 UTC |
6eb61c6 | Jameson Nash | 20 February 2017, 22:10:13 UTC | Merge pull request #19321 from JuliaLang/jn/time-flisp add configurable flisp profiling / timing code | 20 February 2017, 22:10:13 UTC |
eb504bb | Martin Holters | 20 February 2017, 21:09:49 UTC | Allow `limit_type_depth` to introduce more than one new `TypeVar` (#20626) Allow `limit_type_depth` to introduce more than one new `TypeVar` Assert `limit_type_depth` returns a supertype of its input Fixes #20615. | 20 February 2017, 21:09:49 UTC |
37072bd | Jeff Bezanson | 20 February 2017, 20:17:07 UTC | fix #20643, stack overflow in type intersection | 20 February 2017, 20:17:07 UTC |
fec29e7 | Fredrik Ekre | 20 February 2017, 18:55:09 UTC | fix sparse setindex issue #20657 (#20677) | 20 February 2017, 18:55:09 UTC |
16a9193 | Sacha Verweij | 20 February 2017, 17:07:45 UTC | Fix signature in issorted docstring (add lt keyword) [ci skip]. (#20682) | 20 February 2017, 17:07:45 UTC |
e80782e | Iblis Lin | 20 February 2017, 14:23:57 UTC | osutils: include DragonFly in is_bsd | 20 February 2017, 14:23:57 UTC |
22790b1 | Iblis Lin | 19 February 2017, 07:47:29 UTC | libdl: implement dllist on FreeBSD DL_ITERATE_PHDR(3): https://www.freebsd.org/cgi/man.cgi?query=dl_iterate_phdr&apropos=0&sektion=3&manpath=FreeBSD+12-current&arch=default&format=html | 20 February 2017, 14:23:40 UTC |
c6b3251 | Curtis Vogt | 20 February 2017, 14:09:32 UTC | Use mbedTLS hostname verification for LibGit2 (#20443) * Use mbedTLS hostname verification for LibGit2 Corrects issues with hostname verification failing when the SSL subjectAltName is not specified and the CN should be used. * Add test for mbedTLS verification * Made mbedTLS verification a seperate patch * Disable startup-file * Remove use of which * Changes from PR review * Use localhost when hostname isn't set to loopback * Handle getaddrinfo exception * Improve warning message * Add libgit2-mbedtls-hostname.patch dependency libgit2-mbedtls-hostname patch directly depends on the libgit2-mbedtls patch. The libgit2-mbedtls-writer-fix patch also modifies the mbedtls_stream.c file seperate section of the file. To ensure that the patches are always applied in the same order we'll make the hostname patch dependent on the writer-fix patch. | 20 February 2017, 14:09:32 UTC |
06a528c | Tim Holy | 20 February 2017, 11:05:54 UTC | Add manual section on method design and more advanced ambiguity resolution | 20 February 2017, 11:05:54 UTC |
9706c8b | Katharine Hyatt | 20 February 2017, 05:27:42 UTC | Merge pull request #20659 from JuliaLang/ksh/gittree Add methods and tests for TreeEntry | 20 February 2017, 05:27:42 UTC |
faa9b15 | Tim Holy | 20 February 2017, 03:02:12 UTC | start->first in getindex This happened to work because start returns the same thing as first for our two OrdinalRange types in base, but it's not guaranteed for all possible OrdinalRange types. | 20 February 2017, 03:02:12 UTC |
0bd8e2b | Tim Holy | 20 February 2017, 02:56:47 UTC | Fix ambiguity in checkbounds_linear_indices | 20 February 2017, 02:56:47 UTC |
a88f505 | Jeff Bezanson | 20 February 2017, 01:34:53 UTC | Merge pull request #20665 from JuliaLang/jb/specificity06 fix some specificity and intersection issues | 20 February 2017, 01:34:53 UTC |
dd52712 | Pablo Zubieta | 19 February 2017, 23:37:36 UTC | Merge pull request #20680 from pabloferz/pz/zerorepeats Fix repeat for zero repetitions | 19 February 2017, 23:37:36 UTC |
208ef68 | Pablo Zubieta | 19 February 2017, 19:50:07 UTC | Fix repeat for zero repetitions | 19 February 2017, 22:10:21 UTC |
1297ede | H-225 | 19 February 2017, 17:43:55 UTC | Be equally forgiving of leading & trailing whitespace in number parsing (#20606) | 19 February 2017, 17:43:55 UTC |