5af191e | Keno Fischer | 02 April 2018, 21:42:56 UTC | WIP for XLA demo | 02 April 2018, 21:42:56 UTC |
bb1386a | Jameson Nash | 30 March 2018, 16:24:11 UTC | ssair: another bug ``` julia> iterate(t::Tuple, i=1) = (i > length(t) ? nothing : (t[i], i+1)) julia> function any_iterate(f, y) st = iterate(y) while !(st === nothing) f(st[1]) && return true st = iterate(y, st[2]) end return false end julia> code_typed(any_iterate, typeof((x -> :c == x, (:a, :b)))) ``` | 30 March 2018, 16:24:15 UTC |
f5cd2f5 | Jameson Nash | 30 March 2018, 15:13:08 UTC | ssair: fix misc bugs, improve printer | 30 March 2018, 15:53:09 UTC |
2b4df97 | Jameson Nash | 30 March 2018, 15:07:56 UTC | codegen: always preserve code with -g2 | 30 March 2018, 15:24:29 UTC |
cd695e2 | Jameson Nash | 29 March 2018, 22:22:28 UTC | fix Conditional inference after linear IR broke it | 30 March 2018, 15:24:29 UTC |
a6726f1 | Keno Fischer | 28 March 2018, 00:26:10 UTC | Add reverse affinity flag for compaction | 28 March 2018, 00:26:10 UTC |
97210bd | Keno Fischer | 24 March 2018, 20:28:28 UTC | WIP: Handle exceptions in new ir | 26 March 2018, 21:57:35 UTC |
bb9d70b | Keno Fischer | 20 March 2018, 22:13:51 UTC | WIP: Teach codegen about new IR directly: | 21 March 2018, 19:30:48 UTC |
ba085f7 | Keno Fischer | 19 March 2018, 21:38:41 UTC | Add back _apply optimization | 20 March 2018, 20:22:34 UTC |
1e3b771 | Keno Fischer | 18 March 2018, 08:36:42 UTC | Enable new optimizer | 19 March 2018, 18:25:27 UTC |
aa120c0 | Keno Fischer | 14 March 2018, 17:59:09 UTC | WIP: [NewOptimizer] Rewrite inlining pass for new optimizer Now that the optimizer is complete, we need to start looking at latency. The biggest problem at the moment is that we keep converting back and forth to the legacy IR format. In order to cease doing so, we need to move the other parts of the optimizer that still operate on the old data structure to operate on the new ones instead. This PR does so for inlining. Current status: Boostraps. | 19 March 2018, 18:25:27 UTC |
b4af81d | Jeff Bezanson | 19 March 2018, 17:27:36 UTC | remove `promote_rule` for strings. fixes #26200. (#26495) | 19 March 2018, 17:27:36 UTC |
bd0123d | Jeff Bezanson | 19 March 2018, 16:58:42 UTC | use host byte order consistently in precompiled module files (#26493) | 19 March 2018, 16:58:42 UTC |
0fffb36 | Jeff Bezanson | 19 March 2018, 16:56:56 UTC | fresh variables on each iteration in nested loop and generator syntax (#26483) fixes #330 | 19 March 2018, 16:56:56 UTC |
8812682 | Keno Fischer | 19 March 2018, 14:16:08 UTC | Merge pull request #26514 from JuliaLang/kf/invalidageupdate Fix incorrect world age range expansion | 19 March 2018, 14:16:08 UTC |
9a55c8f | Alex Arslan | 19 March 2018, 03:59:05 UTC | Add deprecations for equalto and occursin (#26480) Also deprecate the unexported EqualTo and OccursIn aliases (which were previously types before 26436). | 19 March 2018, 03:59:05 UTC |
90a2162 | Keno Fischer | 19 March 2018, 02:57:36 UTC | Fix incorrect world age range expansion This is a bit of a tricky one, so let me start at the beginning. In PR #25506, I pushed my WIP of a new inlining passes. However, one of the things it temporarily doesn't do is turn call sites to :invoke, causing significantly more jl_apply_generic calls to be codegened. On CI for that PR (as well as locally), we non-deterministically see errors like: ``` LoadError("sysimg.jl", 213, KeyError(LibGit2 [76f85450-5226-5b5a-8eaa-529ad045b433])) ``` Some investigation revealed that what happened here is that a `PkgId` got placed in a dictionary, but was later failed to be retrieved because its hash had supposedly changes. Further investigation reveals that while the hash function used to place the item in the dictionary was `hash(x::Union{String,SubString{String}}, h::UInt64)` in hashing2.jl, the hash function being used to retrieve it is `hash(@nospecialize(x), h::UInt64)` in hashing.jl. The former is loaded later than the latter and should thus invalidate all specializations of the latter, making them ineligible for selection by jl_apply_generic. However, looking at the appropriate age ranges showed that `typeof(hash).name.mt.cache` had entries for both hash functions with overlapping age ranges (which is obviously incorrect). Tracking age range updates, it turns out the world age range for the specialization of `hash(@nospecialize(x), h::UInt64)` was expanded by invalidate_backedges called upon the definition of a hash function inside `LibGit2`. This seems wrong. I believe invalidate_backedges should only ever truncate world age ranges rather than expanding them. This patch simply does that. The non-determinism of the original issue appears to have to do with which of the specializations happen to be in the `jl_lookup_generic_` `call_cache` fast-path. This issue is fairly hard to reproduce because it requires a very specific confluence of circumstances. Since the range of the captured specialization only gets extended to before the min_world age of the new definition, it is never visible in the latest world (e.g. at the REPL). The included test case demonstrates the issue by capturing the world age with a task. Commenting out the `f(x::Float64)` definition makes the test pass, because it is that definition that causes the expansion of the original specialization of `f`. | 19 March 2018, 03:16:55 UTC |
cb5fee8 | Kenta Sato (佐藤 建太) | 19 March 2018, 02:54:24 UTC | remove unused Lisp procedures (#26510) | 19 March 2018, 02:54:24 UTC |
ed37984 | Valentin Churavy | 18 March 2018, 19:34:54 UTC | Merge pull request #26500 from JuliaLang/aa/revert-remove-repackage Revert "Remove repackaging SuiteSparse script. It is no longer needed." | 18 March 2018, 19:34:54 UTC |
dc67f12 | Stefan Karpinski | 18 March 2018, 18:08:42 UTC | Merge pull request #26491 from bicycle1885/mod-name prohibit reserved-word module names | 18 March 2018, 18:08:42 UTC |
57cbdf9 | Kristoffer Carlsson | 18 March 2018, 09:35:53 UTC | make the REPL start faster (#26502) | 18 March 2018, 09:35:53 UTC |
86c9d9c | Keno Fischer | 18 March 2018, 08:32:12 UTC | Merge pull request #26420 from JuliaLang/kf/heapsroa WIP: [NewOptimizer] Get to feature parity | 18 March 2018, 08:32:12 UTC |
fb292c1 | Keno Fischer | 13 March 2018, 21:09:44 UTC | Fix the last couple test failures | 18 March 2018, 04:41:33 UTC |
1fbdd0d | Jeff Bezanson | 17 March 2018, 22:55:24 UTC | more documentation linking `Dict`, `==`, and `hash` (#26498) | 17 March 2018, 22:55:24 UTC |
69ee505 | Kristoffer Carlsson | 17 March 2018, 20:45:13 UTC | Merge pull request #26489 from JuliaLang/kc/bump_pkg3_2 Update Pkg3 | 17 March 2018, 20:45:13 UTC |
4d9c4a9 | Kristoffer Carlsson | 17 March 2018, 09:09:53 UTC | set user for generate in test | 17 March 2018, 18:44:49 UTC |
a81c5b7 | Kristoffer Carlsson | 16 March 2018, 21:35:43 UTC | improve building for project (#210) (cherry picked from commit ef37aa7d4f68e47f74346da40daa671528385755) | 17 March 2018, 18:44:49 UTC |
7b27ae0 | Kristoffer Carlsson | 13 March 2018, 08:49:51 UTC | fix using relative paths w.r.t the project file (cherry picked from commit 799d49741cd5ab0edc28185a4fed82eb0f974b58) | 17 March 2018, 18:44:49 UTC |
f4cf681 | Kristoffer Carlsson | 16 March 2018, 14:22:56 UTC | update precompile | 17 March 2018, 18:44:46 UTC |
7dfa6ec | Steven G. Johnson | 17 March 2018, 03:02:30 UTC | additional tests for take/drop iterators (#12029) test that take and drop arguments are nonnegative | 17 March 2018, 03:02:30 UTC |
21058ef | Alex Arslan | 16 March 2018, 23:59:46 UTC | Revert "Remove repackaging SuiteSparse script. It is no longer needed." This reverts commit 1b099846879a02cd4c9884f33f690e6813f1415f. This script is used by Travis. | 16 March 2018, 23:59:46 UTC |
d3fc586 | Keno Fischer | 13 March 2018, 02:29:05 UTC | [NewOptimizer] Do proper liveness analysis before idf | 16 March 2018, 23:48:58 UTC |
464f29f | Keno Fischer | 13 March 2018, 00:08:31 UTC | Put a bandaid over sroa in loops The full solution requires the liveness analysis in idf I've been punting on so far. | 16 March 2018, 23:48:58 UTC |
87927b9 | Keno Fischer | 12 March 2018, 01:05:50 UTC | [NewOptimizer] Teach getfield elim about ccall preserves The old optimizer knows how to decay preserves of a mutable struct that is eligible for alloc elimination to its constituent parts. Teach the new optimizer the same. Fixes the remaining codegen tests. | 16 March 2018, 23:48:58 UTC |
8af40ea | Keno Fischer | 11 March 2018, 22:31:24 UTC | [NewOptimizer] Perform getfield elim for mutable structs The algorithm works essentially the same as SSA renaming and reuses idf. | 16 March 2018, 23:48:58 UTC |
cf65ee1 | Matt Bauman | 16 March 2018, 20:45:04 UTC | Fix SparseVector broadcasting oddities (#26474) * Fix SparseVector broadcasting oddities Lots of SparseVector oddities fixed here, all by deleting code. Previous behaviors that are now fixed include: ```julia julia> spzeros(5) .* [1] ERROR: DimensionMismatch("") Stacktrace: [1] _binarymap(::typeof(*), ::SparseVector{Float64,Int64}, ::Array{Int64,1}, ::Int64) at /home/mbauman/julia-wip3/usr/share/julia/site/v0.7/SparseArrays/src/sparsevector.jl:1325 [2] _vmul at /home/mbauman/julia-wip3/usr/share/julia/site/v0.7/SparseArrays/src/sparsevector.jl:1369 [inlined] [3] broadcast(::typeof(*), ::SparseVector{Float64,Int64}, ::Array{Int64,1}) at /home/mbauman/julia-wip3/usr/share/julia/site/v0.7/SparseArrays/src/sparsevector.jl:1388 [4] top-level scope julia> spzeros(5) .* ones(5) 5-element Array{Float64,1}: 0.0 0.0 0.0 0.0 0.0 julia> spzeros(5) .* ones(5) .* 1 # this is as expected 5-element SparseVector{Float64,Int64} with 0 stored entries ``` Those all now behave as you'd expect, returning a SparseVector in all cases. * Remove debugging code [ci skip] | 16 March 2018, 20:45:04 UTC |
83b0231 | Matt Bauman | 16 March 2018, 17:54:04 UTC | Remove kludgy VectorStyle and MatrixStyle (#26430) * Remove kludgy VectorStyle and MatrixStyle These broadcast styles were introduced in response to https://github.com/JuliaLang/julia/pull/23939#discussion_r147042741 as a way to limit the "greediness" of Sparse's broadcasting implementation -- sparse only wanted to allow known combinations of array types (including Array but not any AbstractArray). The idea was to allow us to gradually improve the sparse broadcast implementation over 1.x in a non-breaking manner. Unfortunately, these special styles for Array make defining new styles in the heirarchy a bit of a pain (ref. https://github.com/JuliaLang/julia/pull/23939#discussion_r155953830), and it was making my life harder in getting the 1.0 breaking changes in. This commit removes these special broadcast styles in favor of just having Sparse identify the cases itself and re-dispatch back into the default implementation in the cases it doesn't know how to handle. * Add a graceful deprecation | 16 March 2018, 17:54:04 UTC |
569981a | Keno Fischer | 16 March 2018, 17:22:26 UTC | Merge pull request #26481 from JuliaLang/mb/showsocketerror Fix display of Sockets.DNSError | 16 March 2018, 17:22:26 UTC |
1b09984 | Viral B. Shah | 16 March 2018, 16:41:33 UTC | Remove repackaging SuiteSparse script. It is no longer needed. As per #26432. | 16 March 2018, 16:41:33 UTC |
71cb0dc | Viral B. Shah | 16 March 2018, 16:34:40 UTC | Remove mention of Intel compilers. (#26432) Add a clear message against using SYSTEM versions of libraries. Remove the SuiteSparse section, which was mainly for old versions of SuiteSparse. We also do not encourage folks to use SYSTEM SuiteSparse. Add a word of caution about SYSTEM flags to the README. | 16 March 2018, 16:34:40 UTC |
41f70ed | Kenta Sato | 16 March 2018, 15:32:57 UTC | prohibit reserved-word module names | 16 March 2018, 15:32:57 UTC |
8fe827c | Kristoffer Carlsson | 16 March 2018, 14:20:57 UTC | testing/building projects (#209) (cherry picked from commit 23af62a65487f626c7e5585d0fb4dfe7c4c9ce67) | 16 March 2018, 14:22:35 UTC |
d29de40 | Kristoffer Carlsson | 15 March 2018, 15:33:32 UTC | Autocomplete (#205) * autocompletion for pkg3 commands and options (cherry picked from commit fd0c9ed49c038728efe6c8d7dac7cc49641f5ea5) * completions for installed packages (cherry picked from commit e1260764a67d7d1991c7200e54b5d5b82983ca9a) * remote completions, sorting * add tests for autocomplete * respect -m/-p when completing * test applying completions * add some more tests (cherry picked from commit 8d3269c64cdfbaff5c86e228e731cfd4a987ff94) | 16 March 2018, 14:05:31 UTC |
540a8c5 | Kristoffer Carlsson | 15 March 2018, 13:34:12 UTC | fix typo in header when printing manifest status (cherry picked from commit 19af2115037ff936666d9dd2e7940263c2c88c99) | 16 March 2018, 14:05:29 UTC |
73b5483 | Kristoffer Carlsson | 15 March 2018, 10:01:48 UTC | make clear that `tree` is an external command (cherry picked from commit d86e683ba7dc74c56b5ceee840a5f07b2d7c8a46) | 16 March 2018, 14:04:36 UTC |
ed0e5a4 | Kristoffer Carlsson | 15 March 2018, 09:53:21 UTC | fix bug when testing packages with test-specific dependencies (cherry picked from commit 572560bdb8bf2e72d580aa8c2c5a146b0bd0e9c0) | 16 March 2018, 14:04:36 UTC |
18f6129 | Kristoffer Carlsson | 15 March 2018, 09:49:18 UTC | fix and test bug when project is a git repo (cherry picked from commit bb3e8a3b3dbb00aad2c28a58de2aa993c6767ccb) | 16 March 2018, 14:04:35 UTC |
277e6d5 | cormullion | 16 March 2018, 08:43:02 UTC | minor copy edits and formatting for Getting Started (#26439) * minor copy edits and formatting for Getting Started * fredrikekre's autojuliabannergenerator added * minor copy edits and formatting for Integers and FP * fixes * Update integers-and-floating-point-numbers.md | 16 March 2018, 08:43:02 UTC |
30318cf | Alex Arslan | 16 March 2018, 04:34:42 UTC | Fix display of pairs(::Vector) (#26482) | 16 March 2018, 04:34:42 UTC |
b16acdf | Alex Arslan | 16 March 2018, 02:36:50 UTC | Fix deprecation warning from LinearAlgebra.trace in sysimg (#26478) trace was renamed to tr in #26365 but the deprecation in sysimg was not updated. | 16 March 2018, 02:36:50 UTC |
857dd51 | Matt Bauman | 15 March 2018, 23:04:52 UTC | Fix display of Sockets.DNSError | 15 March 2018, 23:04:52 UTC |
29a1719 | Valentin Churavy | 15 March 2018, 22:41:26 UTC | Merge pull request #26477 from JuliaLang/vc/armv7l increase MAX_INTERP_STATE_SIZE on armv7-l to 48 | 15 March 2018, 22:41:26 UTC |
a599bed | Valentin Churavy | 15 March 2018, 18:31:07 UTC | increase MAX_INTERP_STATE_SIZE on armv7-l to 48 | 15 March 2018, 18:31:07 UTC |
71fbc70 | Jeff Bezanson | 15 March 2018, 18:28:08 UTC | replace `equalto` and `occursin` with curried `isequal`, `==`, and `in` (#26436) | 15 March 2018, 18:28:08 UTC |
66147ff | Stefan Karpinski | 15 March 2018, 18:21:27 UTC | Merge pull request #26365 from JuliaLang/jb/tr RFC: rename `trace` to `tr` | 15 March 2018, 18:21:27 UTC |
c6b2afc | Milan Bouchet-Valat | 15 March 2018, 17:44:51 UTC | Fix findall() on sparse matrices when some stored entries are false (#26182) The number of returned indices needs to be adjusted in that case. | 15 March 2018, 17:44:51 UTC |
f355935 | Milan Bouchet-Valat | 15 March 2018, 17:44:21 UTC | Ensure that all isequal methods return Bool to fix inference (#26184) Without this, the return type of isequal() would be inferred as Union{Bool,Missing} when the type of the argument is inferred as Any. | 15 March 2018, 17:44:21 UTC |
983f0f9 | Matt Bauman | 15 March 2018, 08:50:40 UTC | Support sparse broadcast with transposes and adjoints (#26331) Fixes #25331 | 15 March 2018, 08:50:40 UTC |
08445af | Felipe A. V. de Bragança Alves | 15 March 2018, 04:19:31 UTC | Change Type{Ref{T}} to Type{<:Ref{T}} (#26364) | 15 March 2018, 04:19:31 UTC |
e050ab6 | Milan Bouchet-Valat | 15 March 2018, 04:12:48 UTC | Rename isupper, islower, ucfirst and lcfirst (#26442) Spell "uppercase" and "lowercase" in full for consistency with uppercase and lowercase functions. | 15 March 2018, 04:12:48 UTC |
5ad5c12 | Matt Bauman | 15 March 2018, 04:11:50 UTC | Deprecate {Linear,Cartesian}Indices(inds...) constructors (#26448) Previously, there had been an ambiguity between `LinearIndices(-2:2)` and `LinearIndices(collect(-2:2))`, with the former treating the argument as axes directly and the latter asking the argument for its axes. This removes the constructors with varargs `(axes...)` in favor of an explicit tuple of sizes and/or axes. | 15 March 2018, 04:11:50 UTC |
f9f3843 | Liozou | 15 March 2018, 04:08:33 UTC | Corrected jl_type_infer when ENABLE_INFERENCE is not set (#26466) | 15 March 2018, 04:08:33 UTC |
8589fa7 | Valentin Churavy | 15 March 2018, 03:56:32 UTC | Merge pull request #26465 from JuliaLang/vc/skipthreadingtest Remove libatomic test | 15 March 2018, 03:56:32 UTC |
583bdea | Valentin Churavy | 14 March 2018, 19:56:06 UTC | delete invalid libatomic test. | 14 March 2018, 23:27:41 UTC |
4be51b7 | Jeff Bezanson | 14 March 2018, 22:57:12 UTC | Merge pull request #26093 from JuliaLang/nl/String Always resize input to zero length in String(::Vector{UInt8}) | 14 March 2018, 22:57:12 UTC |
4f70460 | Valentin Churavy | 14 March 2018, 22:47:16 UTC | Merge pull request #26464 from JuliaLang/vc/relaxed_test relax cholmod suitesparse test, since it fails on AArch64 and PPC64 | 14 March 2018, 22:47:16 UTC |
3fe38f8 | Stefan Karpinski | 14 March 2018, 22:23:13 UTC | docs: fix remaining uses of "by reference" for calling convention (#26447) | 14 March 2018, 22:23:13 UTC |
63a72d1 | Milan Bouchet-Valat | 16 February 2018, 22:36:39 UTC | Make resize! a no-op when size does not change Else we set the last element to zero for one-byte element types like UInt8, which means that resizing a vector allocated with StringVector corrupts it. Also add redundant checks in Julia code to avoid a function call. | 14 March 2018, 20:32:13 UTC |
03bd792 | Milan Bouchet-Valat | 16 February 2018, 21:34:53 UTC | Always resize input to zero length in String(::Vector{UInt8}) This makes the behavior more predictable than only resizing Vector{UInt8} inputs when they have been allocated via StringVector, as the caller may have obtained them from other code without knowing how they were created. This ensures code will not rely on the fact that a copy is made in many common cases. The behavior is also simpler to document. | 14 March 2018, 20:32:13 UTC |
b571283 | Valentin Churavy | 14 March 2018, 20:10:48 UTC | relax cholmod suitesparse test, since it fails on AArch64 and PPC64 | 14 March 2018, 20:10:48 UTC |
71ef5a0 | Valentin Churavy | 14 March 2018, 17:52:33 UTC | Merge pull request #26456 from JuliaLang/vc/fixup_ccall fixup usage of AttributeBuilder in ccall.cpp for LLVM 5+ | 14 March 2018, 17:52:33 UTC |
f354532 | Steven G. Johnson | 14 March 2018, 17:04:25 UTC | updated NEWS links (see #26454) | 14 March 2018, 17:04:25 UTC |
c4a4b06 | Valentin Churavy | 14 March 2018, 15:17:45 UTC | fixup usage of AttributeBuilder in ccall.cpp for LLVM 5+ | 14 March 2018, 15:17:45 UTC |
a0180d9 | Fredrik Ekre | 14 March 2018, 13:10:38 UTC | misc fixes that error's the doc build on --depwarn=error (#26452) | 14 March 2018, 13:10:38 UTC |
ead9eab | Kristoffer Carlsson | 14 March 2018, 08:38:25 UTC | prevent fill! from inlining (#26418) | 14 March 2018, 08:38:25 UTC |
3086c23 | Jameson Nash | 14 March 2018, 01:47:54 UTC | Merge pull request #26329 from JuliaLang/jn/cfunction-errors cfunction: improve codegen error handling | 14 March 2018, 01:47:54 UTC |
c5f1092 | Keno Fischer | 11 March 2018, 00:23:30 UTC | Add some comments to getfield elim pass | 13 March 2018, 23:30:49 UTC |
91f2f77 | Keno Fischer | 03 March 2018, 04:15:37 UTC | [NewOptimizer] Add back getfield elim pass This adds back a (fixed version of) the NewOptimizer getfield elim pass. This passes all getfield elim (and `@allocated`) tests in base, with the exception of two tests that check for getfield elim on mutables. That part is planned in the design of this pass, but will be part of a follow up PR. | 13 March 2018, 23:30:49 UTC |
3883deb | Klaus Crusius | 13 March 2018, 20:30:42 UTC | default tolerance of pinv uses `min(size(A)...)`, qrfact uses max (#26429) * default tolerance of pinv uses `max(size(A)...)` * reverted code change - changed doc. Fixed qrfact tolerance | 13 March 2018, 20:30:42 UTC |
7beb110 | Kristoffer Carlsson | 13 March 2018, 08:55:02 UTC | stop using pass by reference in noteworthy difference (#26427) | 13 March 2018, 08:55:02 UTC |
46dcb35 | Christian Kurz | 12 March 2018, 15:00:42 UTC | Deja-VU fonts link: add missing closing parenthesis (#26425) | 12 March 2018, 15:00:42 UTC |
c6a6a65 | Alex Arslan | 12 March 2018, 07:53:24 UTC | Touch pcre_h.jl during the build on Windows (#26400) This syncs the modification time, thereby avoiding issues with clock skew while running Make on Windows. Also, make sure we delete a couple of files that shouldn't be there. Fixes #26399 | 12 March 2018, 07:53:24 UTC |
e81e8f7 | Kristoffer Carlsson | 12 March 2018, 07:52:17 UTC | Merge pull request #26415 from JuliaLang/kc/Pkg3_bump Pkg3: update | 12 March 2018, 07:52:17 UTC |
0a7bb64 | Matt Bauman | 12 March 2018, 04:33:19 UTC | Deprecate fallback elsize for AbstractArray (#26379) Just like the deprecation of strides (#25321), the default implementation for `elsize` is generally not true. This deprecates the fallback, restructures it to be defined on the type (and not the value), and implements it where possible for the builtin array types and wrappers. | 12 March 2018, 04:33:19 UTC |
00ac33a | Matt Bauman | 12 March 2018, 04:07:25 UTC | Ensure fast contiguous SubArrays are actually strided (#26393) Ensure fast contiguous SubArrays are actually strided before using them in the definition of other StridedArrays. | 12 March 2018, 04:07:25 UTC |
6a02edf | Jeff Bezanson | 12 March 2018, 00:40:54 UTC | fix #26310, give `import` warnings even if source and dest are the same (#26383) | 12 March 2018, 00:40:54 UTC |
870e7e1 | Jeff Bezanson | 12 March 2018, 00:40:15 UTC | Merge pull request #26411 from JuliaLang/jb/methodroots dump.c and IR encoding tweaks | 12 March 2018, 00:40:15 UTC |
f2d5019 | Kristoffer Carlsson | 11 March 2018, 21:22:41 UTC | actually call Sys.iswindows and give a USER env in the tests | 11 March 2018, 21:22:41 UTC |
61bf935 | Christian Kurz | 11 March 2018, 19:30:43 UTC | docstringify comments, readable variable names (#26346) * docstringify comments, readable variable names, removed `m.pageoffset = 0` `m.pageoffset = 0` is the default for the implemented menus. If a menu needs to start at a different offset, it should be able to. * reverted `m.pageoffset=0` | 11 March 2018, 19:30:43 UTC |
f738de4 | Hans-Peter | 11 March 2018, 19:18:35 UTC | fix remaining `uninit`s in uninitialized PR #26316 (#26413) | 11 March 2018, 19:18:35 UTC |
c742c9a | Kristoffer Carlsson | 11 March 2018, 18:51:03 UTC | fix whitespace | 11 March 2018, 18:51:03 UTC |
b3a164b | Kristoffer Carlsson | 11 March 2018, 16:41:10 UTC | update precompile statements | 11 March 2018, 16:41:10 UTC |
f10e7ee | Kristoffer Carlsson | 11 March 2018, 15:32:48 UTC | Revert "disable optional precompile for Pkg3" This reverts commit 9c0042e5a9bbefe6818ef2a6eadd3a08e1097663. | 11 March 2018, 15:32:48 UTC |
05c24c3 | Kristoffer Carlsson | 11 March 2018, 15:03:56 UTC | put back deletion of temp dirs (cherry picked from commit 1a03bc10102c97f5646186752a75eb03549e077e) | 11 March 2018, 15:30:00 UTC |
735bb5c | Kristoffer Carlsson | 11 March 2018, 14:49:18 UTC | be more careful about closing LibGit2 objects (cherry picked from commit db39581a5011c74c5867aee8fdbf36e3ed93d0bb) | 11 March 2018, 15:30:00 UTC |
a8a89f4 | Kristoffer Carlsson | 11 March 2018, 13:09:37 UTC | make the clone path before trying to use it (cherry picked from commit 9f82c4287c5451896faaa6f596c2166d270aa60d) | 11 March 2018, 15:28:54 UTC |
3145b95 | Kristoffer Carlsson | 11 March 2018, 12:59:07 UTC | work around windows throwing on isdir (cherry picked from commit 20de9c1c740d256dd08a79af0180c286bfc68357) | 11 March 2018, 15:28:54 UTC |
fa634b9 | Kristoffer Carlsson | 11 March 2018, 12:55:00 UTC | add some more docs (cherry picked from commit df1a5c320839b179e809f83a1f07d75328440a05) | 11 March 2018, 15:28:54 UTC |
b46afe1 | Kristoffer Carlsson | 10 March 2018, 20:40:34 UTC | fixes to pinned packages tracking repo branch (cherry picked from commit 1277713735d32f7abd91d13b8688cd0acae3a519) | 11 March 2018, 15:28:54 UTC |
9c4b39d | Kristoffer Carlsson | 10 March 2018, 20:06:07 UTC | describe how to leave pkg repl mode (cherry picked from commit ec7f5e3b948769da39d458b2afa8d0b8ca5256e2) | 11 March 2018, 15:28:54 UTC |