Sage 8.2.rc3 released

[Volker Braun]

As always, you can get the latest beta version from the "develop" git branch. Alternatively, the self-contained source tarball is at http://www.sagemath.org/download-latest.html

a9d274b933 (tag: 8.2.rc3, trac/develop) Updated SageMath version to 8.2.rc3

d85d03cfc3 Trac #25051: Add DESTDIR support to additional Python packages; upgrade pip to latest patch release

ab932bd4d3 Trac #23969: Downgrade cryptominisat to experimental

9bac572eb4 Trac #25107: Ignored OSErrors in test output on Cygwin

3d4ffb3923 Trac #25092: sage --gdb does not start due to SIGFPE

f150587302 Trac #25137: Temporarily skip the test that causes #24986

303efaa461 (tag: 8.2.rc2) Updated SageMath version to 8.2.rc2

[Ralf Stephan]

Builds fine from scratch with clang-5.0.1. sage -gdb will now abort here however.

https://trac.sagemath.org/ticket/25196

[Erik Bray]

Thanks for getting the other two Cygwin issues fixed! It looks good
to me. However, there is still a problem
https://trac.sagemath.org/ticket/25150 still needs to be fixed--I'd
be perfectly open to a different solution if the one I've suggested is
bad.

> --
> You received this message because you are subscribed to the Google Groups
> "sage-release" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to sage-release...@googlegroups.com.
> To post to this group, send email to sage-r...@googlegroups.com.
> Visit this group at https://groups.google.com/group/sage-release.
> For more options, visit https://groups.google.com/d/optout.

[Justin C. Walker]

> On Apr 16, 2018, at 15:34 , Volker Braun <vbrau...@gmail.com> wrote:
>
> As always, you can get the latest beta version from the "develop" git branch. Alternatively, the self-contained source tarball is at http://www.sagemath.org/download-latest.html

Built from a fresh clone/update of the develop branch on macOS 10.11.6 (Quad-core Core i7) w/o problems, and all tests (‘ptestlong’) passed!

Justin


--
Justin C. Walker
Curmudgeon-at-large
Director
Institute for the Absorption of Federal Funds
----
186,000 Miles per Second
Not just a good idea:
it's the law!
----

[John H Palmieri]

OS X, latest Xcode. Built from scratch, also separately built by upgrading an earlier version. Both give doctest failures:

sage -t --long src/sage/doctest/external.py  # Killed due to abort
sage -t --long src/sage/combinat/designs/ext_rep.py  # Killed due to abort

[John H Palmieri]

This has now happened on a second machine, OS X 10.13.4, latest Xcode. Fails in parallel, fails on its own, fails when upgrading an old Sage, fails when building from a new rc3 tarball. I don't think I built rc2 but I think I did build rc1 and didn't have this problem.

sage -t --long src/sage/doctest/external.py
    Killed due to abort
**********************************************************************
Tests run before process (pid=46054) failed:
sage: from sage.doctest.external import has_internet ## line 38 ##
sage: has_internet() # random ## line 39 ##
objc[46054]: +[__NSPlaceholderDate initialize] may have been in progress in another thread when fork() was called.
objc[46054]: +[__NSPlaceholderDate initialize] may have been in progress in another thread when fork() was called. We cannot safely call it or ignore it in the fork() child process. Crashing instead. Set a breakpoint on objc_initializeAfterForkError to debug.
------------------------------------------------------------------------
0   signals.so                          0x00000001019a2548 print_backtrace + 40
1   ???                                 0x6c70704118105f6b 0x0 + 7813868778366721899
------------------------------------------------------------------------
Unhandled SIGABRT: An abort() occurred.
This probably occurred because a *compiled* module has a bug
in it and is not properly wrapped with sig_on(), sig_off().
Python will now terminate.
------------------------------------------------------------------------

**********************************************************************
----------------------------------------------------------------------

sage -t --long src/sage/doctest/external.py  # Killed due to abort

----------------------------------------------------------------------

[Dima Pasechnik]

"objc_" error message might be a sign of a gcc runtime incompatibility, perhaps.

[John H Palmieri]

On Thursday, April 19, 2018 at 4:33:12 PM UTC-7, Dima Pasechnik wrote:

"objc_" error message might be a sign of a gcc runtime incompatibility, perhaps.

I'll try with 8.2.rc2 (which also builds gcc) and with 8.2.rc3 + #25118 (which should use clang instead).

--
John

[John H Palmieri]

According to 'git bisect', the problem is #25092, and in particular this commit:

8b3a7c56bec2f70f96b9eee5fa8fba029bc53587 is the first bad commit
commit 8b3a7c56bec2f70f96b9eee5fa8fba029bc53587
Author: Jeroen Demeyer <jdemeyer@...>
Date:   Fri Apr 6 15:11:18 2018 +0200

    Upgrade cysignals to version 1.7.0

[Samuel Lelièvre]

Under macOS 10.10.5, while Sage 8.2.rc2 builds fine,

building Sage 8.2.rc3 fails while building cysignals 1.7.0.

​

[Jeroen Demeyer]

On 2018-04-20 01:23, John H Palmieri wrote:
> objc[46054]: +[__NSPlaceholderDate initialize] may have been in progress
> in another thread when fork() was called.

Googling for this error messages shows that we are not alone:

https://blog.phusion.nl/2017/10/13/why-ruby-app-servers-break-on-macos-high-sierra-and-what-can-be-done-about-it/

[Dima Pasechnik]

Does what they suggest,

export OBJC_DISABLE_INITIALIZE_FORK_SAFETY=YES

before starting anything, help in Sage's case?

Dima

[Dima Pasechnik]

I can't reproduce this on OSX 10.13 Xeon machine, with 8.2.rc3 + #23353 and + #24969

built from scratch with clang/clang++

Perhaps #24969 would fix your error?

Or maybe it's because of 

$ ./sage 
SageMath version 8.2.rc3, Release Date: 2018-04-16
sage: from sage.doctest.external import has_internet
sage: has_internet()
False

What does this has_internet() do, really?

I get this on the machine at LRI, where I log in via ssh...

[John H Palmieri]

On Friday, April 20, 2018 at 8:43:58 AM UTC-7, Dima Pasechnik wrote:

I can't reproduce this on OSX 10.13 Xeon machine, with 8.2.rc3 + #23353 and + #24969

built from scratch with clang/clang++

Perhaps #24969 would fix your error?

Why would upgrading R fix the error? I can try it, but I don't see why, since cysignals seems to be the problem.
 

Or maybe it's because of 

$ ./sage 
SageMath version 8.2.rc3, Release Date: 2018-04-16
sage: from sage.doctest.external import has_internet
sage: has_internet()
False

What does this has_internet() do, really?

I get this on the machine at LRI, where I log in via ssh...

I get this, too. has_internet() does this:

    from six.moves import urllib
    urllib.request.urlopen("http://www.sagemath.org",timeout=1)

which, when I run interactively, gives

    HTTPError: HTTP Error 403: Forbidden

and this causes the function returns False. 

[John H Palmieri]

On Monday, April 16, 2018 at 3:34:24 PM UTC-7, Volker Braun wrote:

As always, you can get the latest beta version from the "develop" git branch. Alternatively, the self-contained source tarball is at http://www.sagemath.org/download-latest.html

By the way, the tarball is not self-contained: it is missing gcc.

[John H Palmieri]

On Friday, April 20, 2018 at 3:17:28 AM UTC-7, Dima Pasechnik wrote:

Does what they suggest,

export OBJC_DISABLE_INITIALIZE_FORK_SAFETY=YES

before starting anything, help in Sage's case?

I did this and then built from a clean tarball. Those tests now pass.

--
John

[John H Palmieri]

On Monday, April 16, 2018 at 3:34:24 PM UTC-7, Volker Braun wrote:

As always, you can get the latest beta version from the "develop" git branch. Alternatively, the self-contained source tarball is at http://www.sagemath.org/download-latest.html

I'm OS X. I'm getting weird rebuilding behavior. For example: I took a clean tarball, built Sage (using "make"), then ran a few tests. Those passed, so I ran "make ptestlong" to run all tests. At this point, many packages were rebuilt, starting with patch. Why?

[John H Palmieri]

I opened https://trac.sagemath.org/ticket/25222 for the has_internet issue.
 

[John H Palmieri]

In fact just doing this before doctesting is good enough.

--
John