reset does not restore "quit"

[zimmermann6]

with Sage 7.5.1:

sage: reset()
sage: quit
---------------------------------------------------------------------------
NameError                                 Traceback (most recent call last)
/localdisk/tmp/sage-7.5.1/local/lib/python2.7/site-packages/sage/all_cmdline.pyc in <module>()
----> 1 quit

NameError: name 'quit' is not defined

I believe reset() should restore exactly the initial state when starting Sage, including the quit function.

Component: user interface

Author: Steven Trogdon

Branch/Commit: cbd60fb

Reviewer: Paul Zimmermann, John Palmieri

Issue created by migration from https://trac.sagemath.org/ticket/22529

[mkoeppe]

comment:1

Near-dup of #16704.

[strogdon]

comment:2

Will something as naive as

diff --git a/src/sage/misc/reset.pyx b/src/sage/misc/reset.pyx
index bd7c94b..21e1587 100644
--- a/src/sage/misc/reset.pyx
+++ b/src/sage/misc/reset.pyx
@@ -7,7 +7,7 @@ import sys
 
 # Exclude these from the reset command.
 # DATA, base64 -- needed by the notebook
-EXCLUDE = set(['sage_mode', '__DIR__', 'DIR', 'DATA', 'base64'])
+EXCLUDE = set(['sage_mode', '__DIR__', 'DIR', 'DATA', 'base64', 'exit', 'quit'])
 
 def reset(vars=None, attached=False):
     """

do the job, or are there other ramifications that need to be addressed?

[jhpalmieri]

comment:3

I was wondering about that solution, too. It seems to work in my very superficial testing. I also wonder if we should delete 'DIR', 'DATA', and 'base64' from EXCLUDE.

[jhpalmieri]

comment:4

Another idea: can we use the reset command from IPython? Type %reset at the command-line to try it. It removes sage_mode but preserves exit and quit.

[jhpalmieri]

comment:5

Oops, no, %reset also kills Integer and other sage imports.

[jhpalmieri]

comment:6

Maybe use %reset -s, or %reset -s -f to skip the confirmation. The command %reset_selective REGEXP will reset any names matching REGEXP:

sage: a = 3
sage: a2 = 5
sage: %whos
Variable   Type       Data/Info
-------------------------------
a          Integer    3
a2         Integer    5
sage: %reset_selective a
Once deleted, variables cannot be recovered. Proceed (y/[n])?  y
sage: %whos
Interactive namespace is empty.

sage: a = 3
sage: a2 = 5
sage: %reset_selective -f ^a$
sage: %whos
Variable   Type       Data/Info
-------------------------------
a2         Integer    5

[strogdon]

comment:7

Replying to @jhpalmieri:

I was wondering about that solution, too. It seems to work in my very superficial testing. I also wonder if we should delete 'DIR', 'DATA', and 'base64' from EXCLUDE.

I don't see that these are set anywhere, even from notebook imports. What about __DIR__? I can't find it either.

[strogdon]

comment:8

Not correction above.

[jhpalmieri]

comment:9

When I search for these terms in the Sage tree, I see no hits (except for this one line in reset.pyx) for __DIR__, and nothing apparently relevant for DIR, DATA, or base64. sage_mode is used in src/sage/all_cmdline.py and src/sage/all_notebook.py.

[zimmermann6]

comment:10

the patch proposed in comment [comment:2] fixes the issue for me.

Paul

[strogdon]

Branch: u/strogdon/reset_does_not_restore__quit_

[strogdon]

Author: Steven Trogdon

[strogdon]

comment:12

I left __DIR__, DIR, DATA and base64 in place as I found some reference to these variables in the sagenb code. Feel free to alter.

[strogdon]

Commit: 8725c63

[sagetrac-git]

Changed commit from 8725c63 to cbd60fb

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1. New commits:

cbd60fb

Add exit and quit to EXCLUDE for trac #22529 and trac #16704

[zimmermann6]

comment:14

seems fine to me. However someone should check it works in the notebook too.
Paul

[jhpalmieri]

Reviewer: Paul Zimmermann, John Palmieri

[jhpalmieri]

comment:15

Looks good to me. As far as I can tell, exit and quit don't do much in the old Sage notebook. They function as they should in the Jupyter notebook. Without this branch, reset() kills them, and with this branch, it does not, and they continue to function.

[strogdon]

comment:16

So how do we deal with #16704? I don't see a dup or some other appropriate button. Perhaps positive review there too?

[zimmermann6]

comment:17

Replying to @strogdon:

So how do we deal with #16704? I don't see a dup or some other appropriate button. Perhaps positive review there too?

you should modify that other ticket as "resolved as duplicate".

Paul

[vbraun]

Changed branch from u/strogdon/reset_does_not_restore__quit_ to cbd60fb