dbf0b45 | Simon Byrne | 06 January 2018, 23:14:28 UTC | allow use of begin blocks inside enum definitions (#25424) | 06 January 2018, 23:14:28 UTC |
886900c | Stefan Karpinski | 06 January 2018, 23:13:09 UTC | Merge pull request #25425 from JuliaLang/mb/reshapeundeferr Fix undefined var error in reshape error | 06 January 2018, 23:13:09 UTC |
7b82e50 | Jameson Nash | 06 January 2018, 20:03:43 UTC | README: show coveralls status in coverall badge (#25419) I assume that this is more helpful than seeing the same number twice. | 06 January 2018, 20:03:43 UTC |
2b0f219 | Jay Weisskopf | 06 January 2018, 20:01:54 UTC | Replace "labelled" with "labeled" in documentation (#25426) "labeled" is the more common spelling, and less likely to be interpreted as a misspelling at first glance. https://books.google.com/ngrams/graph?content=labeled%2Flabelled&corpus=15 | 06 January 2018, 20:01:54 UTC |
2cfa6a5 | Timo Kluck | 06 January 2018, 18:44:50 UTC | Add fast implementation of find_next and find_prev for sparse vectors and matrices. * Sparse vector/matrix: add fast implementation of find_next and find_prev Before this commit, find_next() will just use the default implementation of looping over each element. When find_next is called without a function filter as first argument, we *know* that semantics are to find elements x satisfying x != 0, so for sparse matrices/vectors, we may only loop over the stored elements. Some care must be taken for stored zero values; that's the reason for the indirection of _sparse_find_next (which only finds the next stored element) and the actual find_next (which does actual non-zero checks). * Optimized findnext() for sparse: update now that predicate needs to be explicit Since we now need explicit predicates [1], this optimization only works if we know that the predicate is a function that is false for zero values. As suggested in that pull request, we could find out by calling `f(zero(eltype(array)))` and hoping that `f` is pure, but I like being a bit more conservative and only applying this optimization only to the case where we *know* `f` is equal to `!iszero`. For clarity, this commit also renames the helper method _sparse_findnext() to _sparse_findnextnz(), because now that the predicate-less version doesn't exist anymore, the `nz` part isn't implicit anymore either. | 06 January 2018, 18:44:50 UTC |
08620e5 | Sacha Verweij | 06 January 2018, 08:09:55 UTC | Slightly relax LAPACK gelsy test tolerance. (#25418) | 06 January 2018, 08:09:55 UTC |
daa0262 | Matt Bauman | 06 January 2018, 00:40:42 UTC | fix undefined var error in reshape error | 06 January 2018, 00:47:23 UTC |
fb52730 | Kristoffer Carlsson | 05 January 2018, 19:41:16 UTC | error on missing testfile (#25408) | 05 January 2018, 19:41:16 UTC |
f41a8f8 | Jeff Bezanson | 05 January 2018, 19:38:16 UTC | expand `min`/`max` with `missing` to all types (#25403) | 05 January 2018, 19:38:16 UTC |
9dd18b1 | Klaus Crusius | 05 January 2018, 19:32:37 UTC | Replace without stack overflow # 25384 (#25386) avoid stack overflow for unsupported argument types in replace! | 05 January 2018, 19:32:37 UTC |
06c66d2 | Kristoffer Carlsson | 05 January 2018, 19:25:14 UTC | fix return value of write(::IO, ::IO) (#25406) | 05 January 2018, 19:25:14 UTC |
b8b4fb3 | Jeff Bezanson | 05 January 2018, 19:21:59 UTC | preserve sign of NaN in conversions to and from BigFloat (#25378) | 05 January 2018, 19:21:59 UTC |
e8c7469 | Fredrik Ekre | 05 January 2018, 17:39:05 UTC | fix some NEWS entries that should be in the 0.7 section (#25405) | 05 January 2018, 17:39:05 UTC |
0ddf9ed | Tim Holy | 05 January 2018, 12:22:47 UTC | Merge pull request #25321 from dlfivefifty/dl/stridedarrayinterface Fix #17812, fix #25247 | 05 January 2018, 12:22:47 UTC |
b10f30b | Tim Holy | 05 January 2018, 12:18:10 UTC | Merge pull request #25393 from JuliaLang/teh/fix_25286 In high-precision arithmetic tests, don't outsmart the obvious | 05 January 2018, 12:18:10 UTC |
5953286 | Jeff Bezanson | 05 January 2018, 02:58:31 UTC | update doc string for `cmp` after #25380 (#25390) | 05 January 2018, 02:58:31 UTC |
b5d4f3c | Fredrik Ekre | 04 January 2018, 22:02:18 UTC | rewrite all ones(...) calls in base, stdlib, test and documentation (#25087) | 04 January 2018, 22:02:18 UTC |
346605b | Sheehan Olver | 04 January 2018, 20:18:15 UTC | test_deprecated for strides before implementing | 04 January 2018, 20:18:15 UTC |
7ce3036 | Sheehan Olver | 04 January 2018, 20:07:02 UTC | merge master | 04 January 2018, 20:07:02 UTC |
b3fa4ef | Tim Holy | 04 January 2018, 19:46:39 UTC | In high-precision arithmetic tests, don't outsmart the obvious Fixes #25286 | 04 January 2018, 19:46:39 UTC |
1950086 | Milan Bouchet-Valat | 04 January 2018, 19:38:28 UTC | Merge pull request #24673 from JuliaLang/nl/search Clean up search and find API | 04 January 2018, 19:38:28 UTC |
236d190 | Kristoffer Carlsson | 04 January 2018, 18:53:30 UTC | make code_warntype and co work with unlowered form of getproperty and setproperty! (#25282) | 04 January 2018, 18:53:30 UTC |
93454e2 | Simon Byrne | 04 January 2018, 16:44:16 UTC | bump MPFR to 3.1.6 (#24924) We can also remove the patch from #22857. | 04 January 2018, 16:44:16 UTC |
1b061e1 | Andrew Keller | 04 January 2018, 16:33:29 UTC | Improve consistency of formatting and wording of deprecation warning messages. (#25366) | 04 January 2018, 16:33:29 UTC |
a621739 | Fredrik Ekre | 04 January 2018, 16:31:28 UTC | remove old Dates exports from base/exports.jl (#25387) | 04 January 2018, 16:31:28 UTC |
89c815b | Kristoffer Carlsson | 04 January 2018, 16:26:11 UTC | remove Markdown's own color printing functionality, use Base's instead (#25293) * remove Markdown's own color printing functionality, use Base's instead * update line number in test | 04 January 2018, 16:26:11 UTC |
e422547 | Simon Byrne | 04 January 2018, 16:21:58 UTC | Add mapreduce_single function (#25051) Since the demise of `r_promote` in #22825, there is now a type-instability in `mapreduce` if the operator does not give an element of the same type as the input. This arose during my implementation of Kahan summation using a reduction operator, see: https://github.com/JuliaMath/KahanSummation.jl/pull/7 This adds a `mapreduce_single` function which defines what the result should be in these cases. | 04 January 2018, 16:21:58 UTC |
a558d76 | Sheehan Olver | 04 January 2018, 16:12:46 UTC | state argument type in substrides so inference is correct | 04 January 2018, 16:12:46 UTC |
e3b7b18 | Jeff Bezanson | 04 January 2018, 16:08:39 UTC | Merge pull request #25319 from JuliaLang/rf/rand/unleash-remains two small changes from #24912 | 04 January 2018, 16:08:39 UTC |
6a46b10 | Jeff Bezanson | 04 January 2018, 16:01:24 UTC | remove `using DelimitedFiles` from default juliarc (#25374) | 04 January 2018, 16:01:24 UTC |
edeb002 | Jeff Bezanson | 04 January 2018, 15:57:43 UTC | fix fallback floating-point `isless` for `-NaN, NaN` (#25376) | 04 January 2018, 15:57:43 UTC |
068f94b | Jeff Bezanson | 04 January 2018, 15:56:42 UTC | Merge pull request #25362 from JuliaLang/jb/pointer_from_objref disallow `pointer_from_objref` on immutable values. fixes #15857 | 04 January 2018, 15:56:42 UTC |
ff8b17d | Milan Bouchet-Valat | 16 December 2017, 17:44:02 UTC | Deprecate findin(a, b) in favor of find(occursin(b), a) | 04 January 2018, 15:52:54 UTC |
6ca43fc | Milan Bouchet-Valat | 16 December 2017, 14:08:21 UTC | Deprecate ismatch(r, s) in favor of contains(s, r) | 04 January 2018, 15:52:54 UTC |
d5f74cd | Milan Bouchet-Valat | 15 December 2017, 21:22:28 UTC | Deprecate (r)searchindex(...) in favor of first(findnext/findprev(...)) | 04 January 2018, 15:52:53 UTC |
93eef2d | Milan Bouchet-Valat | 20 November 2017, 21:02:53 UTC | Merge search into findfirst/findnext and rsearch into findlast/findprev | 04 January 2018, 15:52:53 UTC |
69f9c2e | Stefan Karpinski | 04 January 2018, 15:41:26 UTC | Merge pull request #25380 from JuliaLang/jb/lexcmp deprecate `lexcmp`, `lexless`; define `cmp` in terms of `isless` | 04 January 2018, 15:41:26 UTC |
3e4ab51 | Fredrik Ekre | 04 January 2018, 11:43:28 UTC | fix deprecation warning for findn(::AbstractVector) (#25365) adjust documentation for findn (fix #25343) remove special special cases for findn(::AbstractArray{T,(1|2)}) | 04 January 2018, 11:43:28 UTC |
caf31eb | Kenta Sato (佐藤 建太) | 04 January 2018, 05:46:56 UTC | remove deprecated usage of float [ci skip] (#25379) | 04 January 2018, 05:46:56 UTC |
d1d48b0 | Jeff Bezanson | 04 January 2018, 02:48:41 UTC | base `cmp` for floats on `isless` instead of throwing errors part of #5290 | 04 January 2018, 05:39:18 UTC |
c3a76c0 | Jeff Bezanson | 03 January 2018, 21:30:30 UTC | deprecate `lexcmp` and `lexless` to `cmp` and `isless` part of #5290 | 04 January 2018, 04:05:42 UTC |
2138144 | Jeff Bezanson | 04 January 2018, 02:34:27 UTC | widen `empty(::Set)` to `empty(::AbstractSet)` (#25363) this allows more operations e.g. on `KeySet` to work. | 04 January 2018, 02:34:27 UTC |
c2bdb99 | Jeff Bezanson | 03 January 2018, 20:45:55 UTC | don't double-count data shared by multiple arrays in `summarysize` (#25372) also count Array object header size fixes #25367 | 03 January 2018, 20:45:55 UTC |
2043060 | Jeff Bezanson | 03 January 2018, 18:48:05 UTC | Merge pull request #25241 from JuliaLang/jb/vectorstring safer vector<->string conversions, fixing #24388 | 03 January 2018, 18:48:05 UTC |
1502bd0 | Jeff Bezanson | 03 January 2018, 00:01:17 UTC | disallow `pointer_from_objref` on immutable values. fixes #15857 | 03 January 2018, 18:06:09 UTC |
1750a83 | Jeff Bezanson | 03 January 2018, 00:03:06 UTC | remove redundant `data_pointer_from_objref` | 03 January 2018, 18:06:09 UTC |
5dcd469 | Jeff Bezanson | 03 January 2018, 18:05:24 UTC | improve inference of `T.mutable` field access | 03 January 2018, 18:06:09 UTC |
3d1886f | Milan Bouchet-Valat | 03 January 2018, 16:04:19 UTC | Fix hash(::AbstractArray) failure in when some types support - and others do not (#25250) The existing code only checked that subtraction was supported for the first two elements, but not for subsequent elements, which is necessary for heterogeneous arrays. Also fix a StackOverflow error due to the widen() fallback calling itself recursively by throwing a MethodError manually. | 03 January 2018, 16:04:19 UTC |
d7a5569 | Jeff Bezanson | 03 January 2018, 15:08:29 UTC | add even more arrow operators (#25360) | 03 January 2018, 15:08:29 UTC |
a70504f | Andreas Noack | 03 January 2018, 08:28:22 UTC | A few fixes to the bkfact docstring (#25316) Adjust signature of `bkfact!` Clarify meaning of ' for complex symmetric matrices Pivoting matrix needs transposition | 03 January 2018, 08:28:22 UTC |
e24e50a | Rafael Fourquet | 03 January 2018, 04:18:37 UTC | make union(::Tuple,...) work again (fix #25338) (#25354) | 03 January 2018, 04:18:37 UTC |
4808eda | Jeff Bezanson | 02 January 2018, 23:57:23 UTC | simplify definition of `isimmutable` This seems to date from before the representation of tuple types was changed in v0.4. | 02 January 2018, 23:57:23 UTC |
d481b1b | Stefan Karpinski | 02 January 2018, 23:32:46 UTC | Merge pull request #25358 from JuliaLang/jn/loading-pre-error require: refactor error handling | 02 January 2018, 23:32:46 UTC |
f532093 | Jameson Nash | 02 January 2018, 20:24:12 UTC | require: refactor error handling it was a redundant to repeat the name->file mapping in require_from_serialized, factor it out to manage more of the error handling at one time (even though this approach requires the duplication of slightly more code) | 02 January 2018, 20:55:43 UTC |
7bce3b1 | Jeff Bezanson | 23 December 2017, 06:38:38 UTC | prevent modifying vector after in-place conversion to string part of #24388 | 02 January 2018, 18:21:35 UTC |
7a44c04 | Jeff Bezanson | 22 December 2017, 00:54:09 UTC | deprecate Vector<->String conversion in favor of something safer add `CodeUnits` and `codeunits` fixes #24388 | 02 January 2018, 18:21:01 UTC |
2cc82d2 | Curtis Vogt | 02 January 2018, 11:44:13 UTC | Correct return type for `std` along region (#25304) * Correct return type for std along region * Use isa instead of typeof | 02 January 2018, 11:44:13 UTC |
337785d | Sheehan Olver | 02 January 2018, 09:22:55 UTC | work around type inference failure in substrides | 02 January 2018, 09:22:55 UTC |
c3c0964 | Andrew Keller | 02 January 2018, 08:33:22 UTC | Some adjustments for dimensionful numbers (#25035) * Replace `0` and `1` with `zero(T)` and `one(T)` following PR 24656. * Revise NEWS entry. * More informative depwarns. * Tweak the deprecations for `zeros(D::Diagonal[, opts...])`. | 02 January 2018, 08:33:22 UTC |
6729e4f | Stefan Karpinski | 02 January 2018, 00:25:04 UTC | Merge pull request #25350 from jayschwa/fix-news-links Fix NEWS.md links for grouped #22064 and #24278 | 02 January 2018, 00:25:04 UTC |
973e548 | Jay Weisskopf | 01 January 2018, 20:16:18 UTC | Fix NEWS.md links for grouped #22064 and #24278 Ran NEWS-update.jl to refresh link references. | 01 January 2018, 20:16:18 UTC |
a655e66 | Bogumił Kamiński | 01 January 2018, 19:49:17 UTC | Fix typo in supported types by Atomic (#25346) Change `Float128` to `Float64`. | 01 January 2018, 19:49:17 UTC |
a4757db | cormullion | 01 January 2018, 19:48:40 UTC | Update typos in stdlib/Dates/docs/src/index.md (#25347) | 01 January 2018, 19:48:40 UTC |
22d9421 | Iblis Lin | 01 January 2018, 17:06:43 UTC | Bump :copyright: (#25345) | 01 January 2018, 17:06:43 UTC |
a3e24cb | Jeff Bezanson | 31 December 2017, 22:09:36 UTC | add leftsquigarrow and rightsquigarrow operators in parser (#25342) | 31 December 2017, 22:09:36 UTC |
fae72aa | Jameson Nash | 31 December 2017, 20:29:02 UTC | Merge pull request #25290 from JuliaLang/jn/kwarg_pairs decouple kwargs performance improvement from API changes | 31 December 2017, 20:29:02 UTC |
fc4fc14 | Kristoffer Carlsson | 31 December 2017, 18:12:58 UTC | Revert "REPL: change ^C to "Clear" input area if not already empty" (#25064) This reverts commit 2f3abe9fc391167ad418b31894d0f491284b0ad8. | 31 December 2017, 18:12:58 UTC |
579ceec | Sheehan Olver | 31 December 2017, 13:55:39 UTC | Deprecate old substrides | 31 December 2017, 13:55:39 UTC |
9018379 | Fredrik Ekre | 31 December 2017, 05:08:29 UTC | fix deprecation of spones from #25037 (#25330) | 31 December 2017, 05:08:29 UTC |
56b59ec | Jeff Bezanson | 31 December 2017, 02:56:01 UTC | some small fixes to `jl_log` (#25335) - synchronize printing format of fallback logging - make new logging work under `--lisp` | 31 December 2017, 02:56:01 UTC |
7fc268f | Jeff Bezanson | 31 December 2017, 02:55:28 UTC | remove 0.6 deprecations in the front end (#25336) | 31 December 2017, 02:55:28 UTC |
d1086e0 | Sheehan Olver | 30 December 2017, 21:09:57 UTC | O(N) substrides implementation | 30 December 2017, 21:09:57 UTC |
2116b4c | Chris Foster | 30 December 2017, 18:42:07 UTC | Send all parser and lowering depwarns to the logging system (#25257) Forward all frontend depwarn messages through to the julia logging system for consistency of formatting and dispatch. Notable changes: * Depwarn messages go to the logging system, rather than STDERR * Meta.parse() is given a `depwarn` keyword for convenience in controlling deprecation warnings. * Tests for all parser and lowering deprecation warnings Detail: * Ensure that syntax-deprecation, deprecation-message forwards to the logging system. Split these into distinct functions for depwarns coming from the parser vs lowering, as these extract line number information in a different way. * Include file and line number as metadata everywhere that the frontend knows about it. * Remove jl_parse_depwarn(), replace flisp *depwarn* / *deperror* with simplified *depwarn-opt* handled in one place. * Replace Base.syntax_deprecation_warnings() with Meta.parse(..., depwarn=false) * Internal C functions jl_log() and jl_logf() for use when communicating log messages from C code. These will need to be augmented with an async jl_safe_log() or something similar when printing log messages from the runtime internals. Todo: * Figure out a decent `id` for the syntax and lowering depwarn messages | 30 December 2017, 18:42:07 UTC |
9c8fb44 | Sheehan Olver | 30 December 2017, 17:34:33 UTC | Make strides(A) the default override, make sterner warning in deprecation message and docs | 30 December 2017, 17:34:33 UTC |
31d4ce6 | Sheehan Olver | 30 December 2017, 16:27:22 UTC | Remove restriction to StridedArray in LAPACK | 30 December 2017, 16:27:22 UTC |
4d1d723 | Stefan Karpinski | 30 December 2017, 16:14:26 UTC | Merge pull request #25276 from JuliaLang/sk/ironic remove ironic invisible Unicode junk | 30 December 2017, 16:14:26 UTC |
d353ab0 | Sheehan Olver | 30 December 2017, 09:57:33 UTC | Remove _defaultstride since it's only used one place | 30 December 2017, 09:57:33 UTC |
17ba03a | Jacob Quinn | 30 December 2017, 09:42:17 UTC | Merge pull request #25227 from invenia/rf/abstractdatetime Introduce an `AbstractDateTime` type | 30 December 2017, 09:42:17 UTC |
224b40f | Sacha Verweij | 30 December 2017, 07:34:28 UTC | Better condition LAPACK least squares solver test matrices to avoid CI failures. (#25287) | 30 December 2017, 07:34:28 UTC |
dad96fb | Andreas Noack | 30 December 2017, 05:49:46 UTC | Use -ffc2 when compiling Fortran libraries with gfortran and MKL (#25298) Fixes #17946 | 30 December 2017, 05:49:46 UTC |
3bcc952 | Katharine Hyatt | 29 December 2017, 23:26:33 UTC | Some xrefs and grammar fixes for manual (#25309) | 29 December 2017, 23:26:33 UTC |
41697f9 | Jeff Bezanson | 02 November 2017, 21:59:43 UTC | remove some uses of `Vector{UInt8}(::String)` | 29 December 2017, 20:45:37 UTC |
2a56a37 | Jeff Bezanson | 29 December 2017, 20:38:15 UTC | make `using A.B` only for modules, `using A: B` only for single bindings (#25306) part of #8000 | 29 December 2017, 20:38:15 UTC |
bbc7b83 | Jameson Nash | 29 December 2017, 20:37:01 UTC | fix replace deprecation in julia-config (#25288) | 29 December 2017, 20:37:01 UTC |
4c2885d | Sheehan Olver | 29 December 2017, 19:01:20 UTC | _cumsumprodsizes -> _defaultstride, addresses stevengl comments | 29 December 2017, 19:01:20 UTC |
903c42e | Sheehan Olver | 29 December 2017, 16:24:12 UTC | move strides tests to test/dense.jl | 29 December 2017, 16:24:12 UTC |
0858576 | Sheehan Olver | 29 December 2017, 15:03:51 UTC | Move stride, stride1 to dense.jl | 29 December 2017, 15:03:51 UTC |
ea5b0a3 | Rafael Fourquet | 29 December 2017, 12:42:31 UTC | faster rand(::MersenneTwister, ::Union{[U]Int{8,16,32},Bool}) (#25277) Now that generating scalar Float64 values (i.e. 52 random bits) is even faster, it becomes faster to just generate such values and make small integers out of them, rather than saving entropy bits with using the integer cache. | 29 December 2017, 12:42:31 UTC |
3788b51 | Rafael Fourquet | 04 December 2017, 17:21:50 UTC | make Sampler{E} encode the type E of elements which are generated Before, a call like `rand(rng, Sampler(rng, 1:10), 3)` generated an `Array{Any,1}`, so a way to get the `eltype` of a Sampler is necessary. Instead of changing Sampler -> Sampler{E}, implementing appropriate eltype methods would have been possible, to keep the helper Sampler subtypes more flexible, but it seemed to be simpler this way. | 29 December 2017, 12:11:50 UTC |
da29da1 | Rafael Fourquet | 05 December 2017, 08:46:02 UTC | rename CloseOpen -> CloseOpen01, Close1Open2 -> CloseOpen12 | 29 December 2017, 12:11:06 UTC |
d0ed2f2 | Rafael Fourquet | 29 December 2017, 11:54:17 UTC | add Future.copy! function (#25144) | 29 December 2017, 11:54:17 UTC |
4d166fa | Sacha Verweij | 28 December 2017, 21:38:00 UTC | Fix conjugation of UniformScaling. (#25294) | 28 December 2017, 21:38:00 UTC |
18bed58 | Sacha Verweij | 28 December 2017, 21:35:12 UTC | Merge pull request #25238 from Sacha0/higho2 optimize and fix map/broadcast over Adjoint/Transpose vectors, take 2 | 28 December 2017, 21:35:12 UTC |
eb91796 | Sacha Verweij | 28 December 2017, 20:31:26 UTC | Merge pull request #25217 from Sacha0/sunsetjazz sunset linalg jazz | 28 December 2017, 20:31:26 UTC |
93998ad | Sheehan Olver | 28 December 2017, 19:41:34 UTC | Add strides for StridedReshapedArray | 28 December 2017, 19:41:34 UTC |
58f7961 | Tamas K. Papp | 28 December 2017, 16:36:55 UTC | Update docstring of bkfact and related getindex. (#25185) * Update docstring of bkfact and related getindex. To be compatible with 88e7fbcd002c21e6a03fe18643d83838cdbb31f7. Also include the actual decomposition format. Fix missing permutation matrix in the `getindex` docstring. * Removed transpose from second permutation matrix. Also mention properties of permutation matrix in docstring. | 28 December 2017, 16:36:55 UTC |
acaf1c2 | Sheehan Olver | 28 December 2017, 14:05:54 UTC | use parents stride in substrides | 28 December 2017, 14:05:54 UTC |
13eb90f | Sheehan Olver | 28 December 2017, 10:09:44 UTC | move size_to_strides back to abstractarray.jl | 28 December 2017, 10:09:44 UTC |
b90274e | Jameson Nash | 27 December 2017, 20:17:31 UTC | undo breaking change to kwargs iteration order This starts to decouple the performance improvement of #24795 from the existence of exactly one implementation of Core.NamedTuple, in preparation for implementing NamedTuple in Julia rather than C. | 27 December 2017, 21:11:44 UTC |
1238bad | Stefan Karpinski | 27 December 2017, 20:40:20 UTC | Merge pull request #25280 from lakshayg/fix/Pkg_status update Pkg.status to throw if pkg does not exist | 27 December 2017, 20:40:20 UTC |