Thu 2020-04-23 à 09:09, Dima Pasechnik:
>
> On Thu, 23 Apr 2020, 14:55 Samuel Lelievre:
>>
>> > Subject: Sage 9.1.rc1 released
>>
>> On macOS 10.14.6 Mojave with lots of Homebrew
>> packages installed, failed to build Pillow. Logs:
>>
>> https://www.math.u-psud.fr/~lelievre/t/sage3logs.zip
>
> What Python3 it is with?
> Could you check whether the pillow problem is related to #29019?
> There are there links to Pillow bug tracker...
Thanks Dima for the excellent pointers once more.
The discussion at
https://trac.sagemath.org/ticket/29019
indeed had a link to
https://github.com/python-pillow/Pillow/issues/3438
with tricks related to
macOS_SDK_headers_for_macOS_10.14.pkg
I went for a serious spring cleaning, uninstalled
Apple's "Command-Line Tools for Developers" and
reinstalled them as well as the headers, uninstalled
and reinstalled Homebrew and MacTeX. Now more
up-to-date and with less accumulated cruft around.
With that, the build of SageMath 9.1.rc1 completed.
```
$ source ./.homebrew-build-env
$ make configure
[...]
$ CC=clang CXX=clang++ FC=gfortran \
OBJC=clang OBJCXX=clang++
./configure
[...]
$ make -s V=0
[...]
Sage build/upgrade complete!
```
And I can now happily use Sage 9.1.rc1 based on Python 3.7.7.
```
$ ./sage
SageMath version 9.1.rc1, Release Date: 2020-04-22
Using Python 3.7.7. Type "help()" for help.
Warning: this is a prerelease version, and it may be unstable.
sage:
```
Everything seems fine with tests:
```
$ make ptestlong
[...]
----------------------------------------------------------------------
All tests passed!
----------------------------------------------------------------------
Total time for all tests: 9656.0 seconds
cpu time: 23962.4 seconds
cumulative wall time: 37234.1 seconds
```
A few things maybe worth reporting.
1. The 'configure' step suggested to run
$ brew install ppl
even though I already had Homebrew's ppl. Intrigued,
I took a look at the part of 'config.log' checking ppl:
```
## ---------------------------------------------------- ##
## Checking whether SageMath should install SPKG ppl... ##
## ---------------------------------------------------- ##
configure:27878: checking whether any of glpk gmp mpir
is installed as or will be installed as SPKG
configure:27882: result: yes; install ppl as well
configure:28277: no suitable system package found for SPKG ppl
```
Since at least one of (glpk, gmp, mpir) either
was or would be installed as an SPKG, the script
concluded ppl should be installed as an SPKG too.
So the decision to install ppl was based on that,
not on the absence of ppl. In such a case, maybe
the script should not recommend installing ppl
(at least not without checking if it is there).
2. During `make ptestlong`, on three occasions
an alert that "Python quit unexpectedly" popped
up and offered to either (a) Ignore, (b) Inspect,
(c) Report to Apple (or something like that).
Does this happen because some doctests crash
Python on purpose and, because of now using
Homebrew's Python instead of building Sage's
own, this is now caught by macOS which offers
to report the crash to Apple?
--
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 [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/sage-release/CAEcArF11p%3Di%3D2%3D3b3xw2nLNgNBSCW-JXLKFSCqzSrrpvRSoT2Q%40mail.gmail.com.
