Does a deprecation warning prevent start-up of jupyter server of Double clickable application SageMath-X.X?

50 views
Skip to first unread message

rana-aere

unread,
Mar 28, 2020, 10:38:49 AM3/28/20
to sage-support
I am testing binary-pkg.

This question is relevant to the previous one on deprecation warning on `\QQ`.
But, the focus is different.

I made SageMath-9.1beta8 by binary-pkg.
Then, I  copied SageMath-9.1beta8 out of the disk-image-file mounted of MacOS.
I ejected the disk-image.
And double clicked SageMath-9.1beta8.
There was an error message (dialogue) on failure on invoking jupyter server.

So, I invoked command-line version SageMath-9.1beta8/sage on Terminal (bash).
There were messages on patching and then there was a deprecation warning.
For command line, this was just one deprecation warning one can ignore.
I found the source code which is responsible for this warning.
I made a correction in the relocated folder tree of SageMath-9.1beta9.
Then, I reinvoked sage through command line and verified that the warning does not show up again.

I quitted sage and double clicked the Mac application.
This time, jupyter server started up normally.

Is it possible Mac application is designed to stop invocation even for a deprecation warning?

I am currently investigating trouble of Mac application with error messages.
(Trouble like adding three extra diagnosis message prevents Mac application of SageMath-9.0 from normal start-up.)
So, any information on the question 4 lines above is helpful.

rana-aere

unread,
Mar 29, 2020, 3:40:51 AM3/29/20
to sage-support
I edited the code back to original and tested how SageMath starts up.
Deprecation warning does not block start-up.
There remains some more mystery in start-up of double clickable SageMath.app of MacOS.

For possible users who encounter the same symptom of SageMath, I leave some record.

Step 1: I copied SageMath9.1beta8.app in /Users/Shared/SagemathTestground.
Step 2: I double clicked the copy  /Users/Shared/SagemathTestground/ SageMath9.1beta8.app
--------> Mac dialogue appeared an reported failure of jupyter server to start.
Step 3: I used "view log" to read the log. The log was incomplete. It seemed patching was on the way.
Step 4: I started  /Users/Shared/SagemathTestground/ SageMath9.1beta8.app/sage from the command line.
--------> Plenty of messages on patching were displayed.
--------> Then, a deprecation warning on \Q in the definition of _macaulay2_init_ appeared.
--------> sage prompt appeared.
Step 5: I quitted sage and edited the relevant file and silenced that warning.
Step 6: I invoked sage from command line and verified warning message is silenced.
Step 7: I double clicked  /Users/Shared/SagemathTestground/ SageMath9.1beta8.app.
--------> Jupyter server started up and Sagemath was OK.
Step 8: I put the code back in original for testing the software for sage-support-group.
Step 9: I quitted /Users/Shared/SagemathTestground/ SageMath9.1beta8.app and double clicked it again.
--------> Jupyter server started up and Sagemath was OK.

Lesson?   Patching at the initial invocation might fail occasionally. And, retrying initial invocation of SageMath several times might help. Deprecation warning seems not the real obstruction.

rana-aere

unread,
Mar 29, 2020, 3:42:35 AM3/29/20
to sage-support
I forgot to mention one thing.

Patching in SageMath-9.1beta8 was on libraries not on source codes.
This might be another difference from the distribution of SageMath-9.0.

rana-aere

unread,
Mar 29, 2020, 10:44:58 AM3/29/20
to sage-support
I finally made it to make a disk-image-file of SageMath-9.1.app with updates of https://trac.sagemath.org/ticket/29371.

The issue of failing start-up with partial record of patching files is solved.
I explain what was happening.
Mac application SageMath-9.0.app set-up bash command using pipes:
Roughly speaking it is
"sage starting jupyter server | tee for recording log | grep in 2 stages for ticket of jupyter".
There were some problem in the last two stages so that grep terminated by error.
Hence, the pipe got stuck.
This prevented tee from writing logs beyond some amount of bytes.
Another effect was probably tee command also terminated and sage command also terminated.
This problem will be solved in SageMath-9.1.

Current work-around is here:
Start /Applications/SageMath-9.0.app/sage by command line.
Then patching will be done.
Exit sage after sage-prompt comes up.

Then, double click Mac Application SageMath-9.0.app.
This solution has a problem of sage terminates unexpectedly when huge amount of logging takes place
for one invocation of jupyter-server.
I admit this.
If you are not in a hurry, waiting for SageMath-9.1 will be an easier solution.

rana-aere

unread,
Mar 29, 2020, 10:47:08 AM3/29/20
to sage-support
I think another trouble of SageMath-9.0.app was about failure of getcwd.
I have an idea on this problem.
I will check it.
Reply all
Reply to author
Forward
0 new messages