[release-0.6] backports for 0.6.0-rc2 #21878

tkelman · 2017-05-14T18:47:50Z

pkgeval
@nanosoldier runbenchmarks(ALL, vs = "@6bdb3950bdf64152") vs 0.6.0-rc1

* clearing the meaning of comment * correct back file name. (cherry picked from commit 1571b62)

… closure (#21700) * fix serialization of multiple global bindings to the same object in a closure (cherry picked from commit 3930eef)

(cherry picked from commit 700bc5f)

(cherry picked from commit f3e4256) ref #21703

most notably in Core.Inference, this may cause the addition of bad edges and suboptimal results (cherry picked from commit 49c1910) ref #21690 inference: improve function call signatures ensures that abstract_iteration gets better information, and that specialization can compile these methods ahead-of-time (cherry picked from commit 0250e4f)

(cherry picked from commit c4ed32e) ref #21721

(cherry picked from commit 3cfb989) ref #21706

(cherry picked from commit d106662) ref #21716 clarify some quasi-ambiguous or unnecessary method signatures Without the fix to issue #21710, these are likely sources of ambiguities. (cherry picked from commit 79605e8)

This was caused by incorrectly using a TypeVar as the value of a static parameter in the case of a compiled specialization whose signature is not a leaf type. In this case, one slot was specialized on `Function`, but the slot with the static parameter had a leaf type. (cherry picked from commit a581292) ref #21729

(cherry picked from commit 4d08b54)

(cherry picked from commit 14da9d3) ref #21707

* Improve docstrings for norm, vecnorm, normalize and normalize! Add equations defining the norms, reorganize the docstrings by argument type, and mention what "normalizing" means. * Remove docstring for the norm() function (cherry picked from commit e257ef4)

(cherry picked from commit b1d0897) ref #21702

(cherry picked from commit 310b6dd)

(cherry picked from commit 15fd965) ref #21756

fix #21763 (cherry picked from commit 57a8f0a) ref #21765

(cherry picked from commit d34c596) ref #21749

(cherry picked from commit d63f72d) ref #21739

(cherry picked from commit 37f20f1) ref #21761

(cherry picked from commit ff73d57) ref #21699 Replace visually confusing asterisks (cherry picked from commit c21e77a)

* Optimize remotecall pattern used in distributed loading (cherry picked from commit 9694de7)

(cherry picked from commit ba87e99) ref #21776

(cherry picked from commit 1129de3)

(cherry picked from commit a2a6d18) ref #21772

Avoid a library call into GMP. (cherry picked from commit 05708f8) ref #21231 speed-up prevpow2 & nextpow2 for BigInt (cherry picked from commit 42fd2a7) GMP: add flipsign[!] functions and use more iszero/isneg/ispos (cherry picked from commit f63bdfd) add ZERO & ONE global BigInt constants for optimization (cherry picked from commit b521acd)

(cherry picked from commit 6771421) ref #21732

… calls (cherry picked from commit 66b8036) ref #21718 don't insert toplevel expression unless there are imports (cherry picked from commit 18291b0) eliminate unnecessary array allocations from isexpr and extract_imports functions (cherry picked from commit 305037a) test fixes (cherry picked from commit 24a74a5)

(cherry picked from commit 4c54935) ref #21799

…g. (#21806) (cherry picked from commit 9ac9071)

…8444). (#21817) (cherry picked from commit d052e88)

GregPlowman · 2017-05-15T21:39:31Z

I guess it's too late for rc2, but would it be possible for PR #21695 to be included before 0.6 final?
And also included in 0.5.3?

…1880) (cherry picked from commit 85a4783)

…18012). (#21881) (cherry picked from commit 379ae49)

(cherry picked from commit af92634)

(cherry picked from commit c7e4fe7)

). (#21883) (cherry picked from commit 66402ac)

…ods (#17607). (#21882) (cherry picked from commit 38cfb73)

(cherry picked from commit 149e589)

tkelman · 2017-05-17T06:09:26Z

d818788 was in pretty good shape on pkgeval https://gist.github.com/df43d5af59dd34a3520e59a57c7ff3d0 with one spurious new failure in LPVSpectral, and Feather is segfaulting on this branch as opposed to freezing on rc1 (it was working up until April 6th though https://github.com/JuliaCI/pkg.julialang.org/blob/88a41b735886f267c8e71ad63be5b3e7eba38449/logs/Feather_0.6.log)

But I think there have been enough new bugfixes on master that it'd be better to get them in here and re-run pkgeval again.

…1877 (#21889) Make sure the methods are getting called from A'\b and A.'\b. Don't promote Factorizations in general Promote Bunch-Kaufman since it doesn't have a generic fallback solver (cherry picked from commit 57ebd82)

Ref https://github.com/JuliaLang/julia/issues/21453#issuecomment-296036653 (cherry picked from commit 6f7c763) ref #21891

* Add NEWS.md entry for three-argument rem and the new rem2pi (#10946). * Clarify precise behaviour (cherry picked from commit 8f9a948)

(cherry picked from commit a34fd51)

Also test for `fd != -1` instead since that is a defined invalid value. (cherry picked from commit 84eeb40) ref #21893

(cherry picked from commit 10c0746) ref #21695

This would widen e.g. `Complex{T} where T` to `_ <: Complex`, which is valid in some sense but yields a type that's not a supertype of the input, which is bad. (cherry picked from commit f46ba0a) ref #21907

(cherry picked from commit 1db6e10) ref #21902

(cherry picked from commit 18562e0) ref #21911

tkelman · 2017-05-17T07:33:52Z

@nanosoldier runbenchmarks(ALL, vs = "@6bdb3950bdf64152") vs 0.6.0-rc1

tkelman · 2017-05-17T08:23:52Z

@jlbuild !nuke

Base.runtests()
Base.runtests("pkg libgit2-online download")

jlbuild · 2017-05-17T08:23:59Z

Status of `be004c1` builds:

Builder Name	Nuke	Build	Download	Code Output
linux32	COMPLETE	COMPLETE	Download	ERRORED
linux64	COMPLETE	COMPLETE	Download	COMPLETE
linuxaarch64	CANCELED	N/A	N/A	N/A
linuxarmv7l	COMPLETE	ERRORED	N/A	N/A
linuxppc64le	COMPLETE	COMPLETE	Download	ERRORED
osx64	COMPLETE	ERRORED	N/A	N/A
win32	CANCELED	N/A	N/A	N/A
win64	ERRORED	N/A	N/A	N/A

nanosoldier · 2017-05-17T10:39:47Z

Your benchmark job has completed - possible performance regressions were detected. A full report can be found here. cc @jrevels

tkelman · 2017-05-17T10:48:45Z

ouch, what happened to date parsing?

tkelman · 2017-05-17T17:46:35Z

hm. flaky false alarm then? @nanosoldier runbenchmarks(ALL, vs = "@6bdb3950bdf64152")

nanosoldier · 2017-05-17T20:51:37Z

Your benchmark job has completed - possible performance regressions were detected. A full report can be found here. cc @jrevels

tkelman · 2017-05-18T02:25:21Z

https://gist.github.com/4804ae73c6d5cec89d78890617c558dc same Feather issue (#21920) and MatrixNetworks had a tolerance too tight

merging and tagging this as rc2

tkelman · 2017-05-18T02:34:01Z

@staticfloat @yuyichao can you check on and reconnect the old arm buildbot? the new buildbots aren't quite ready to be making releases off of just yet

DrKrar and others added 30 commits May 13, 2017 17:50

clearing the meaning of comment (#21741)

7593289

* clearing the meaning of comment * correct back file name. (cherry picked from commit 1571b62)

fix serialization of multiple global bindings to the same object in a…

e585dc8

… closure (#21700) * fix serialization of multiple global bindings to the same object in a closure (cherry picked from commit 3930eef)

Add a word missing from the README (#21711)

dbf2a1e

(cherry picked from commit 700bc5f)

Do not hard code temp file path in the test.

b05b8c9

(cherry picked from commit f3e4256) ref #21703

Add examples for clone

d3a25b9

(cherry picked from commit c4ed32e) ref #21721

Add example for authors

9a47376

(cherry picked from commit 3cfb989) ref #21706

fix #21710, specificity of Array vs. AbstractVector

886f20f

(cherry picked from commit d106662) ref #21716 clarify some quasi-ambiguous or unnecessary method signatures Without the fix to issue #21710, these are likely sources of ambiguities. (cherry picked from commit 79605e8)

Move testenv specific consts and addprocs to a separate file. (#21713)

de4ccaf

(cherry picked from commit 4d08b54)

Add example for checkout

7001598

(cherry picked from commit 14da9d3) ref #21707

Word wrap at 92-characters

ecae3a7

(cherry picked from commit b1d0897) ref #21702

Add example for iscommit and reset! (#21734)

19ed349

(cherry picked from commit 310b6dd)

Fix a promotion failure with linspace

e2e21c5

(cherry picked from commit 15fd965) ref #21756

add typeof_tfunc for UnionAll types

205bf77

fix #21763 (cherry picked from commit 57a8f0a) ref #21765

Docs for blob content and example for addblob

2cfe3b1

(cherry picked from commit d34c596) ref #21749

improve error reporting for invalid type arguments to ccall

7b28753

(cherry picked from commit d63f72d) ref #21739

fix method specificity issue causing problems for pr #21750

54d68e8

(cherry picked from commit 37f20f1) ref #21761

Switch to dash for list default

c15639a

(cherry picked from commit ff73d57) ref #21699 Replace visually confusing asterisks (cherry picked from commit c21e77a)

Optimize remotecall pattern used in distributed loading (#21745)

e83527e

* Optimize remotecall pattern used in distributed loading (cherry picked from commit 9694de7)

fix error in showerror with backtraces from CapturedExceptions

3cb69ea

(cherry picked from commit ba87e99) ref #21776

Fix stack overflow in SSHCredentials constructor (#21779)

55bb927

(cherry picked from commit 1129de3)

Implement tryparse(Float16, str) via Float32 conversion

a461fb5

(cherry picked from commit a2a6d18) ref #21772

Clean up testenv and old test isolation code

844ef37

(cherry picked from commit 6771421) ref #21732

fix #21793, regression in serializing nested closure

0f1aa84

(cherry picked from commit 4c54935) ref #21799

Add link to libuv's uv_os_homedir documentation to homedir's docstrin…

0222855

…g. (#21806) (cherry picked from commit 9ac9071)

Add NEWS.md entry for deprecation of invoke's tuple-of-types form (#1…

2a92944

…8444). (#21817) (cherry picked from commit d052e88)

Sacha0 and others added 7 commits May 16, 2017 06:11

Add NEWS.md entry for single-argument write deprecation (#17654). (#2…

3b4a1a3

…1880) (cherry picked from commit 85a4783)

Add NEWS.md entry for delete!(::EnvHash, ::AbstractString, def) dep (#…

d19cd5e

…18012). (#21881) (cherry picked from commit 379ae49)

Add NEWS.md entry for at-textmime deprecation (#18441). (#21884)

92c95c0

(cherry picked from commit af92634)

Add NEWS.md entry for ipermutedims deprecation (#18891). (#21885)

be5689f

(cherry picked from commit c7e4fe7)

Add NEWS.md entry for three-arg SubArray constructor deprecation (#19259

94f73eb

). (#21883) (cherry picked from commit 66402ac)

Add NEWS.md entry for dep of ambiguous UniformScaling-Number +/- meth…

8654afd

…ods (#17607). (#21882) (cherry picked from commit 38cfb73)

Use AbstractVector in _cpu_summary (#21876)

4feba43

(cherry picked from commit 149e589)

andreasnoack and others added 9 commits May 16, 2017 23:34

Mention inv in docstrings of factorizations that implement it

e6f1f34

Ref https://github.com/JuliaLang/julia/issues/21453#issuecomment-296036653 (cherry picked from commit 6f7c763) ref #21891

add NEWS.md entry for three-argument rem and rem2pi (#21812)

5f9a9bb

* Add NEWS.md entry for three-argument rem and the new rem2pi (#10946). * Clarify precise behaviour (cherry picked from commit 8f9a948)

Make expm1(::Complex64) type stable. (#21896)

f445c6a

(cherry picked from commit a34fd51)

Make CoreIO thread-safer

a42d4d9

Also test for `fd != -1` instead since that is a defined invalid value. (cherry picked from commit 84eeb40) ref #21893

Issue 19960 - loading module on remote workers

9564404

(cherry picked from commit 10c0746) ref #21695

fix #21848, bug in widening done by limit_type_depth

09d3b26

This would widen e.g. `Complex{T} where T` to `_ <: Complex`, which is valid in some sense but yields a type that's not a supertype of the input, which is bad. (cherry picked from commit f46ba0a) ref #21907

Only populate Julia-specific LDFLAGS with ASAN options.

a9d418e

(cherry picked from commit 1db6e10) ref #21902

fix #21906, ccall causes unnecessary variable Box

be004c1

(cherry picked from commit 18562e0) ref #21911

tkelman merged commit 419992a into release-0.6 May 18, 2017

tkelman deleted the tk/backports-0.6.0-rc2 branch May 18, 2017 02:29

odow mentioned this pull request May 18, 2017

Deserialization Bug odow/SDDP.jl#1

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[release-0.6] backports for 0.6.0-rc2 #21878

[release-0.6] backports for 0.6.0-rc2 #21878

tkelman commented May 14, 2017 •

edited

Loading

GregPlowman commented May 15, 2017

tkelman commented May 17, 2017 •

edited

Loading

tkelman commented May 17, 2017

tkelman commented May 17, 2017

jlbuild commented May 17, 2017 •

edited

Loading

nanosoldier commented May 17, 2017

tkelman commented May 17, 2017

tkelman commented May 17, 2017 •

edited

Loading

nanosoldier commented May 17, 2017

tkelman commented May 18, 2017

tkelman commented May 18, 2017

[release-0.6] backports for 0.6.0-rc2 #21878

[release-0.6] backports for 0.6.0-rc2 #21878

Conversation

tkelman commented May 14, 2017 • edited Loading

GregPlowman commented May 15, 2017

tkelman commented May 17, 2017 • edited Loading

tkelman commented May 17, 2017

tkelman commented May 17, 2017

jlbuild commented May 17, 2017 • edited Loading

Status of be004c1 builds:

nanosoldier commented May 17, 2017

tkelman commented May 17, 2017

tkelman commented May 17, 2017 • edited Loading

nanosoldier commented May 17, 2017

tkelman commented May 18, 2017

tkelman commented May 18, 2017

tkelman commented May 14, 2017 •

edited

Loading

tkelman commented May 17, 2017 •

edited

Loading

jlbuild commented May 17, 2017 •

edited

Loading

Status of `be004c1` builds:

tkelman commented May 17, 2017 •

edited

Loading