d0e0393 | Steven G. Johnson | 31 January 2020, 17:34:08 UTC | :ᵀ no longer parses | 31 January 2020, 17:34:08 UTC |
f5b45ff | Steven G. Johnson | 28 January 2020, 15:52:27 UTC | broken test | 28 January 2020, 15:52:27 UTC |
326a960 | Steven G. Johnson | 28 January 2020, 14:02:58 UTC | don't allow identifiers to start with Lm (Letter, modifier) | 28 January 2020, 14:02:58 UTC |
fba188c | Steven G. Johnson | 28 January 2020, 03:57:18 UTC | allow @foo{...} macro calls (#34505) | 28 January 2020, 03:57:18 UTC |
a5c422f | Elliot Saba | 27 January 2020, 19:24:12 UTC | tests: Use `realpath()` before `==` path comparisons (#34506) These should probably be using `samefile`, if they were real code instead of just tests. Though it's unclear why real code would be doing this. Maybe just don't put your paths in hash-tables and you'll normally be fine. | 27 January 2020, 19:24:12 UTC |
91a118e | Jameson Nash | 27 January 2020, 19:09:48 UTC | track-allocation: charge entry cost to caller (#34391) Don't include one-time costs (JIT compilation) so that warm-up isn't generally required. And adjust codegen emission to charge call entry costs to the caller. fixes #11753 fixes #19981 fixes #21871 fixes #34054 close #18595 | 27 January 2020, 19:09:48 UTC |
c5da9de | Jameson Nash | 27 January 2020, 19:09:00 UTC | build: remove mingw32 support macro (#34471) We have not supported this compiler in many years (though its successor project, mingw-w64, also defined it). And it's a weird combination of compiler and platform specifics: the present uses were generally either using it instead of either the more appropriate GCC or WINDOWS. | 27 January 2020, 19:09:00 UTC |
ed177d1 | bramtayl | 27 January 2020, 16:45:39 UTC | Missing generators.jl test (#34450) | 27 January 2020, 16:45:39 UTC |
5bbf37d | AzamatB | 27 January 2020, 13:02:48 UTC | Fix wrong issue reference typo (#34532) | 27 January 2020, 13:02:48 UTC |
54c1667 | Takafumi Arakaki | 27 January 2020, 13:00:46 UTC | Always use Symbol as _group value of log records (#34509) | 27 January 2020, 13:00:46 UTC |
4eab1bf | Benoît Legat | 27 January 2020, 10:14:21 UTC | [ci skip] Gustavsen -> Gustavson (#34530) | 27 January 2020, 10:14:21 UTC |
3595c7c | Mohit Nain | 27 January 2020, 08:07:13 UTC | Typo correction in line 64 (#34529) is should be replaced with are , as we are talking about 2 different contents | 27 January 2020, 08:07:13 UTC |
57c6786 | Sascha Timme | 26 January 2020, 22:36:39 UTC | Fix condskeel definition (#34512) * Fix condskeel definition Skeel's condition number is a relative condition number, however the definition of `condskeel(A,x)` missed to divide by the norm of `x`. The definition is for example in Higham "Accuracy and stability of numerical algorithms" eq. (7.13) (p. 123). * Test condskeel definition fix | 26 January 2020, 22:36:39 UTC |
c4a0c83 | kamalojasv181 | 26 January 2020, 14:50:04 UTC | Remove incorrect type info from `range` documentation (#34475) | 26 January 2020, 14:50:04 UTC |
9790a8d | Tim Holy | 25 January 2020, 20:28:58 UTC | Support brief and extended docs (closes #25930) (#34226) | 25 January 2020, 20:28:58 UTC |
0234e00 | simeonschaub | 25 January 2020, 18:06:36 UTC | fix printing of LineNumberNode in block within ref (#34504) | 25 January 2020, 18:06:36 UTC |
5157d08 | simeonschaub | 24 January 2020, 23:30:22 UTC | fix bug in printing of block expressions in ref (#34503) | 24 January 2020, 23:30:22 UTC |
e5e094c | Jeff Bezanson | 24 January 2020, 21:11:17 UTC | fix method ambiguities in testhelpers/Furlongs.jl (#34480) | 24 January 2020, 21:11:17 UTC |
e0740fe | Erik Schnetter | 24 January 2020, 14:13:51 UTC | jl_special_vector_alignment: Remove work-around for old LLVM version (<=3.8) (#34490) LLVM <= 3.8 segfaults when LLVM SIMD vectors are not powers of 2 (or a sum of 2 different powers of 2). This LLVM bug was corrected in 2017 (see <https://bugs.llvm.org/show_bug.cgi?id=27708>). This change removes this work-around, finally allowing arbitrary SIMD vector lengths in Julia. This changes Julia's ABI for SIMD types. All SIMD types are now represented as SIMD vectors in LLVM. The work-around represented some SIMD types as LLVM arrays instead to avoid the LLVM segfault. I don't think that SIMD vector lengths that are not a power of 2 were in widespread use, if at all. | 24 January 2020, 14:13:51 UTC |
4db419b | Fredrik Ekre | 24 January 2020, 08:11:58 UTC | Remove URLs with /latest and use correct, relative, (#34491) documenter links for the manual, fixes #34488. | 24 January 2020, 08:11:58 UTC |
6af5f03 | Kristoffer Carlsson | 23 January 2020, 23:09:11 UTC | Update Documenter version to 0.24.4. (#34487) | 23 January 2020, 23:09:11 UTC |
551947b | bramtayl | 23 January 2020, 19:21:05 UTC | Missing enum tests (#34468) | 23 January 2020, 19:21:05 UTC |
f579d14 | Matt Bauman | 23 January 2020, 19:13:22 UTC | More efficient indexing by CartesianIndices (#31214) Treat CartesianIndices as its tuple of ranges when indexing. This allows views that use CartesianIndices as indices to be strided and changes the complexity of boundschecking from O(n^d) to O(d) Co-authored-by: Kristoffer Carlsson <kristoffer.carlsson@chalmers.se> | 23 January 2020, 19:13:22 UTC |
6d86384 | Valentin Churavy | 23 January 2020, 18:34:14 UTC | Merge pull request #34489 from JuliaLang/vc/llvm8_only Remove support for LLVM6 and LLVM7 | 23 January 2020, 18:34:14 UTC |
643c931 | Valentin Churavy | 23 January 2020, 13:28:59 UTC | remove LLVM patches for v6,v7 | 23 January 2020, 14:34:51 UTC |
5393e4c | Valentin Churavy | 23 January 2020, 12:59:51 UTC | remove support for LLVM prior to v8 | 23 January 2020, 14:34:51 UTC |
c2abaee | Stefan Karpinski | 22 January 2020, 06:54:13 UTC | Distributed, SHA tests: fix soft scope shadowing (#34472) | 22 January 2020, 06:54:13 UTC |
5c635ad | Valentin Churavy | 21 January 2020, 22:14:22 UTC | Merge pull request #34313 from JuliaLang/vc/llvm10 Support LLVM 10 changes | 21 January 2020, 22:14:22 UTC |
38675f8 | Valentin Churavy | 21 January 2020, 20:47:53 UTC | LLVM10 is going to hide target specific symbols and CUDAnative no longer uses this | 21 January 2020, 20:47:53 UTC |
ebce82e | Valentin Churavy | 21 January 2020, 19:04:55 UTC | [LLVM10] remove unused stream argument | 21 January 2020, 20:04:19 UTC |
c9940ed | bramtayl | 21 January 2020, 18:51:52 UTC | Missing pair tests (#34451) | 21 January 2020, 18:51:52 UTC |
b7efa97 | Valentin Churavy | 08 January 2020, 20:14:41 UTC | [LLVM10] use M->global_object().{begin|end}() | 21 January 2020, 18:40:52 UTC |
b81321d | Valentin Churavy | 08 January 2020, 17:07:43 UTC | [LLVM10] Memset requires MaybeAlign | 21 January 2020, 18:40:52 UTC |
3aeb3ec | Rafael Fourquet | 21 January 2020, 17:54:12 UTC | isdisjoint: add NEWS and include in the docs (#34460) | 21 January 2020, 17:54:12 UTC |
46a98ad | Kenta Sato | 21 January 2020, 16:50:53 UTC | replace mantissa with significand in docstring (#34453) | 21 January 2020, 16:50:53 UTC |
0549bf1 | Elliot Saba | 21 January 2020, 05:17:25 UTC | `fieldnames()` can throw for types such as `typeof(+)` (#32840) Fixes #32558 | 21 January 2020, 05:17:25 UTC |
b28bb24 | 0x005c | 21 January 2020, 03:41:48 UTC | Improve error message for character literal (#34435) | 21 January 2020, 03:41:48 UTC |
159de94 | bramtayl | 20 January 2020, 21:42:18 UTC | Missing numbers tests (#34448) | 20 January 2020, 21:42:18 UTC |
fca037a | Jeff Bezanson | 20 January 2020, 21:30:58 UTC | fix #33882, `step(::StepRangeLen)` should not usually convert (#34412) | 20 January 2020, 21:30:58 UTC |
6959a0c | bramtayl | 20 January 2020, 20:52:05 UTC | Missing complex tests (#34446) | 20 January 2020, 20:52:05 UTC |
046dfb0 | bramtayl | 20 January 2020, 20:51:51 UTC | missing NamedTuple test (#34447) | 20 January 2020, 20:51:51 UTC |
bc0e5d1 | Keno Fischer | 20 January 2020, 20:40:32 UTC | Add `isdisjoint` (#34427) This was added in #13192, but the PR became stale. We've also changed a bit how we write these kinds of algorithms, so make use of the tools we have (e.g. `fastin`). | 20 January 2020, 20:40:32 UTC |
f68753c | Jeffrey Lin | 20 January 2020, 19:57:58 UTC | base/libc.jl: fix crash when run before epoch (#34056) (#34442) | 20 January 2020, 19:57:58 UTC |
a3ee6fe | Jeff Fessler | 20 January 2020, 16:27:56 UTC | Document the preferred use of show over display (#34438) One-line addition to documentation to help prevent future questions like this: https://discourse.julialang.org/t/what-function-display-does-ijulia-call-at-the-end-of-each-cell/33554 | 20 January 2020, 16:27:56 UTC |
5928786 | Jeff Bezanson | 20 January 2020, 08:34:34 UTC | fix #34415, remove affinity setting code (#34433) This was removed before and added back apparently by mistake. | 20 January 2020, 08:34:34 UTC |
0d99f4c | bramtayl | 20 January 2020, 08:33:05 UTC | Missing rational tests (#34443) | 20 January 2020, 08:33:05 UTC |
bcd5882 | Kristoffer Carlsson | 20 January 2020, 08:26:43 UTC | Merge pull request #34129 from sostock/gcd Make it easier to extend lcm/gcd/gcdx to custom types | 20 January 2020, 08:26:43 UTC |
4baf0f0 | Katharine Hyatt | 20 January 2020, 06:52:21 UTC | Missing spzeros test (#34440) | 20 January 2020, 06:52:21 UTC |
34fd6f0 | Fernando Chorney | 20 January 2020, 04:39:18 UTC | Force "uUeE" dateformats to return Int64 values (#31448) | 20 January 2020, 04:39:18 UTC |
2ba1d11 | Elliot Saba | 19 January 2020, 20:34:47 UTC | BinaryBuilder dependencies bump (#34242) BinaryBuilder dependencies bump | 19 January 2020, 20:34:47 UTC |
248bc46 | Stefan Karpinski | 19 January 2020, 02:41:37 UTC | src/jloptions.c: fix buggy use of strchr (#34432) | 19 January 2020, 02:41:37 UTC |
f120989 | Keno Fischer | 19 January 2020, 02:41:16 UTC | Check return value of SystemFunction036 (#34429) In both places we call it. Also tabs-to-spaces for the new code to match the formatting of the rest of the code. | 19 January 2020, 02:41:16 UTC |
5bed770 | Keno Fischer | 19 January 2020, 02:40:10 UTC | Add clamp(x, T) (#34426) This is #22235, but with the reversed argument order. ``` julia> clamp( 260, UInt8 ) 0xff ``` Co-authored-by: "arghhhh <david.hossack@gmail.com>" | 19 January 2020, 02:40:10 UTC |
2b4b3d5 | Elliot Saba | 03 January 2020, 00:13:24 UTC | Remove OpenBLAS `skylakexdgemm` patch now that we're on 0.3.7 | 18 January 2020, 22:24:21 UTC |
4916063 | Keno Fischer | 31 December 2019, 14:51:02 UTC | Fix CI hangs on win32 After much debugging, it turns out that OpenBlas is sometimes hanging on exit due to bad assumptions in its thread management. The upstream discussion is at https://github.com/xianyi/OpenBLAS/pull/2350, but this should fix our frequent win32 CI failures in the meantime. | 18 January 2020, 22:24:21 UTC |
9e154b9 | Elliot Saba | 02 January 2020, 22:29:21 UTC | Rename LLVM checksums to match new naming convention | 18 January 2020, 22:24:21 UTC |
a9635cc | Elliot Saba | 02 January 2020, 22:27:04 UTC | Upgrade `OpenBLAS` to `v0.3.7` | 18 January 2020, 22:24:21 UTC |
9c4286c | Elliot Saba | 02 January 2020, 22:24:59 UTC | Upgrade BB triplet naming from `-gccX` tags to `-libgfortranX` tags. This will require us to change our LLVM and OpenBLAS binaries, as they are the only libgfortran-specific libraries that we ship with base Julia. | 18 January 2020, 22:24:21 UTC |
e4d0808 | Elliot Saba | 02 January 2020, 22:22:19 UTC | Update `openlibm` to `v0.7.0`. It's also no longer libgfortran-dependent, so flip that flag and reduce the number of separate files by 2/3rds. | 18 January 2020, 22:24:21 UTC |
372a7a7 | Keno Fischer | 18 January 2020, 18:35:35 UTC | Merge pull request #34389 from JuliaLang/tb/asan ASAN fixes | 18 January 2020, 18:35:35 UTC |
85c9c66 | Keno Fischer | 18 January 2020, 18:31:06 UTC | Merge pull request #34424 from JuliaLang/kf/docrollup Roll up stalled out doc PRs | 18 January 2020, 18:31:06 UTC |
d759b5b | Mustafa M | 18 January 2020, 18:21:19 UTC | Use UUID's to generate random tempname on Windows (#33785) * Use UIUD to create random tempname on Windows * Use underscores and remove extension * Truncate to 10 chars the UUID * Generate the random name from a random byte array * Update file.jl | 18 January 2020, 18:21:19 UTC |
ca5642d | Tom Short | 18 January 2020, 18:08:17 UTC | Dev-doc updates for the SSAIR section (#30622) * Dev-doc updates for the SSAIR section This mainly fixes typos. Note that I think the `foo` function on line 114 is wrong. It errors when run (`y` not defined), and it doesn't match the IR shown below it (there's no `bar` function for example). I don't have a fix for that. * Fix foo example Co-authored-by: Keno Fischer <keno@alumni.harvard.edu> | 18 January 2020, 18:08:17 UTC |
cab645d | Cédric St-Jean | 18 January 2020, 18:07:55 UTC | Fix docs typo (bytes -> bits) (#30452) Co-authored-by: Keno Fischer <keno@alumni.harvard.edu> | 18 January 2020, 18:07:55 UTC |
5c75bec | Katharine Hyatt | 18 January 2020, 11:56:38 UTC | Test copy for ReshapedArray of SparseMatrixCSC (#34420) | 18 January 2020, 11:56:38 UTC |
870b7fd | Keno Fischer | 18 January 2020, 04:44:50 UTC | Add tests for mixed-integer-type lcm/gcd (#34425) As pointed out in #30512 | 18 January 2020, 04:44:50 UTC |
bf25dc5 | Keno Fischer | 18 January 2020, 00:10:30 UTC | Small tweak to embedding.md (#34422) Adds -rpath to the link line which was suggested in #28982 and removes the line about JULIA_BINDIR, which isn't necessary anymore, since we autodetect the directory. Also remove the reference to libstdc++. If you built julia yourself, it's definitely compatible with the system one. If not, we now build julia binary distributions with wide compatibility with different libstdc++ versions, so the system one should be fine. | 18 January 2020, 00:10:30 UTC |
250aecc | Markus Kuhn | 17 January 2020, 23:01:53 UTC | Sockets: fix getipaddr/getipaddrs documentation, add islinklocaladdr function (#34300) * fix getipaddr/getipaddrs documentation - resolved IPv4/IPv6 confusion, e.g. getipaddrs() does _not_ return only IPv4 - more realistic example addresses - example addresses taken from address ranges reserved for documentation https://en.wikipedia.org/wiki/Reserved_IP_addresses * add Sockets.islinklocaladdr IPv6 interfaces always have a link-local address configured, which is also reported by getipaddrs(). However, typical application servers are not reachable under their link-local address, therefore this convenience function helps users of getipaddrs() to filter out such addresses with `filter(!islinklocaladdr, getipaddrs())`. This commits also adds a "See also" hint that in applications where a server process has been started remotely via ssh that `split(ENV["SSH_CONNECTION"], ' ')[3]` can be a preferable alternative source for the IP address under which this server is reachable (namely the address that ssh has already used successfully to get here). | 17 January 2020, 23:01:53 UTC |
5315476 | Takafumi Arakaki | 17 January 2020, 22:59:48 UTC | Document that using `invoke` on external functions is not forward-compatible (#33702) | 17 January 2020, 22:59:48 UTC |
06c4a88 | Mark | 17 January 2020, 22:13:45 UTC | Roll up #29389 | 17 January 2020, 22:13:45 UTC |
cafa097 | Keno Fischer | 17 January 2020, 21:58:27 UTC | Address problematic sentences identified in #29372 | 17 January 2020, 21:58:27 UTC |
a230280 | Keno Fischer | 17 January 2020, 21:50:25 UTC | Note about .julia/config from #29390 | 17 January 2020, 21:50:25 UTC |
4ba5647 | Keno Fischer | 17 January 2020, 21:46:02 UTC | Two changes from #29360 | 17 January 2020, 21:46:02 UTC |
8aa8e22 | Keno Fischer | 17 January 2020, 21:40:50 UTC | Fix poor formulation pointed out in #29359 | 17 January 2020, 21:40:50 UTC |
2330cb2 | Mark | 20 September 2018, 13:06:18 UTC | Minor documentation improvements [#29286] | 17 January 2020, 21:33:11 UTC |
d316024 | Keno Fischer | 17 January 2020, 21:02:13 UTC | Minor reword | 17 January 2020, 21:22:34 UTC |
f61c2e6 | Mark | 26 September 2018, 08:28:22 UTC | Minor English fixes/improvements Cherry-pick of #29370. | 17 January 2020, 21:22:34 UTC |
833e6bf | Keno Fischer | 17 January 2020, 18:45:16 UTC | Remove checked conversions in two more cases. NFC. (#34406) Like #34398, but in a different place. I also realized there was a good reason to do this beside personal preference: Once we start tracking nothrow more closely (to optimize try/catch, etc.), we we'll want to make sure not to emit error paths that the julia optimizer itself can't eliminate. | 17 January 2020, 18:45:16 UTC |
85eb20f | Sebastian Stock | 17 December 2019, 19:45:43 UTC | Make it easier to extend lcm/gcd/gcdx to custom types | 17 January 2020, 15:35:37 UTC |
3549cd1 | Sebastian Stock | 17 January 2020, 10:07:05 UTC | Fix array-argument gcd for non-integers | 17 January 2020, 14:42:59 UTC |
4254045 | Katharine Hyatt | 17 January 2020, 10:41:01 UTC | Test colon dropstored methods (#34403) | 17 January 2020, 10:41:01 UTC |
ed87ece | Jeff Bezanson | 17 January 2020, 08:18:24 UTC | remove code forcing testsets to be compiled (#34405) | 17 January 2020, 08:18:24 UTC |
40569c2 | Steven G. Johnson | 17 January 2020, 07:39:58 UTC | nonsquare (underdetermined) LQ solves (#34350) * nonsquare (underdetermined) LQ solves * fix NEWS item * fix NEWS item * narrow method signature to eliminate ambiguity * fix docs -- under/overdetermined terms were swapped * Update stdlib/LinearAlgebra/test/lq.jl Co-Authored-By: Stefan Karpinski <stefan@karpinski.org> Co-authored-by: Stefan Karpinski <stefan@karpinski.org> | 17 January 2020, 07:39:58 UTC |
ce211b9 | Travis DePrato | 16 January 2020, 21:59:03 UTC | Improve docstring for finalizer. (#34285) | 16 January 2020, 21:59:03 UTC |
d64c971 | simeonschaub | 16 January 2020, 21:56:14 UTC | nicer constructor for ImmutableDict (#34297) | 16 January 2020, 21:56:14 UTC |
270fcff | Tim Besard | 16 January 2020, 21:49:16 UTC | Clean-up the GC.gc interface. (#34401) | 16 January 2020, 21:49:16 UTC |
60b60f1 | Steven G. Johnson | 16 January 2020, 15:44:42 UTC | some basic views documentation (#34390) Added a little basic discussion at the beginning of the "views" section of the standard-library documentation, based on [this discourse post](https://discourse.julialang.org/t/could-you-explain-what-are-views/17535/3?u=stevengj) which people seem to have found helpful. | 16 January 2020, 15:44:42 UTC |
5578a0c | Keno Fischer | 16 January 2020, 15:43:35 UTC | Use unchecked conversion in bitcounting operations. NFC. (#34398) LLVM's intrinsics for bitcounting return an integer of the same size as the input integer. On the julia side, we then trunc back to `Int` (since we're guaranteed that the answers will always be at most `log2(x)`). However, the code was calling for a checked conversion, rather than an unchecked one. This doesn't matter too much as LLVM does know about the result range of these intrinsics, but it made me sad to generate all these error checks in the first place, so use the unchecked variants instead. | 16 January 2020, 15:43:34 UTC |
a7cd97a | Jeff Bezanson | 16 January 2020, 09:22:35 UTC | mark `inbounds` and `loopinfo` as quoted during lowering (#34397) | 16 January 2020, 09:22:35 UTC |
8e1ad64 | Tim Besard | 15 January 2020, 15:49:31 UTC | Don't read trailing datatype layout bytes when first_ptr==-1. | 16 January 2020, 08:20:30 UTC |
64b9da1 | Tim Besard | 15 January 2020, 14:46:04 UTC | Don't have jl_svec statically verify the argcount with ASAN. | 16 January 2020, 08:20:30 UTC |
cb46917 | Katharine Hyatt | 16 January 2020, 07:41:40 UTC | Add test for setindex with a matrix of bools (#34388) | 16 January 2020, 07:41:40 UTC |
94249da | Jameson Nash | 16 January 2020, 00:00:03 UTC | coverage: fix name of the new code_coverage_effect Expr head (#34392) | 16 January 2020, 00:00:03 UTC |
a9f1227 | Jeff Bezanson | 15 January 2020, 22:37:56 UTC | fix some cases of diagonal subtyping (#34272) fix #26108, fix #26716, fix #31824 | 15 January 2020, 22:37:56 UTC |
15d693b | Jameson Nash | 15 January 2020, 18:13:21 UTC | revert breaking change to GC.gc() (#34336) Removes the breaking part of b0ed147c5cf0566019ff08755af71eb5649411bc from PR #34303 | 15 January 2020, 18:13:21 UTC |
08ebedc | Kristoffer Carlsson | 15 January 2020, 15:50:52 UTC | Merge pull request #34254 from JuliaLang/jn/codecov-accuracy fix codecov accuracy for optimized and dead statements | 15 January 2020, 15:50:52 UTC |
07a16d6 | Keno Fischer | 15 January 2020, 01:36:49 UTC | Change lowering of gc preserve (#34379) This fixes #34247 by changing the way gc preserve is lowered. Instead of lowering it in a macro, lower it in the frontend. This allows us to use an SSA value directly for the return token of the gc begin expression. This bypasses the slot-renaming pass of the compiler, thus preventing the compiler from trying to save and restore the token. Of course, this kind of code would generally not be legal (because it uses an SSA value outside of the regular domination relation), but since this is a julia restriction, not an LLVM restriction, we can simply exempt gc_begin tokens from this particular validation. This works fine at the LLVM level also, because it doesn't have this particular restriction. It also doesn't have the same correctness problems as doing the same for non-token values, as the tokens get lowered away by the try/catch lowering before reaching the LLVM backend. | 15 January 2020, 01:36:49 UTC |
edac6f7 | Tim Besard | 14 January 2020, 23:21:57 UTC | Execute AM/PM Dates tests in the C locale. (#34371) | 14 January 2020, 23:21:57 UTC |
0403fd5 | Liozou | 14 January 2020, 22:53:46 UTC | Fixed keyword calls to code_macros with keyword arguments (#34367) | 14 January 2020, 22:53:46 UTC |
b1e39bc | Elliot Saba | 14 January 2020, 22:43:43 UTC | Eliminate erroneous quoting on `${PATCHELF}` (#34377) * Eliminate erroneous quoting on `${PATCHELF}` This only manifests on platforms like `aarch64` and `ppc64le` because we force extra arguments to `PATCHELF` in these cases which didn't interact well with these well-meaning quotes. * patchelf doesn't parse `--arg=val` properly * `debug` output should not be captured by `$()` * Quoting, permissions, and debugging fixes. All of my favorite things in one place! | 14 January 2020, 22:43:43 UTC |