Show traceback, continuing to start kernel if pylab init fails #991

Closed
wants to merge 1 commit into
from

Conversation

Projects
None yet
3 participants
@minrk
Member

minrk commented Nov 10, 2011

Error message goes directly over stderr (and traceback over stdout),
rather than as a pyerr message, which is associated with a particular
frontend, and thus ignored because startup actions were not requested
by anyone.

This better matches the terminal behavior, where ipython --pylab without matplotlib
installed shows a traceback, but continues to start.

In the qtconsole the traceback will come in below the banner, but in the notebook
the message will not arrive until after the first execution.

closes gh-990

Show traceback, continuing to start kernel if pylab init fails
Error message goes directly over stderr (and traceback over stdout),
rather than as a pyerr message, which is associated with a particular
frontend, and thus ignored because startup actions were not requested
by anyone.

This better matches the terminal behavior, where `ipython --pylab` without matplotlib
installed shows a traceback, but continues to start.

In the qtconsole the traceback will come in below the banner, but in the notebook
the message will not arrive until after the first execution.

closes gh-990
@fperez

This comment has been minimized.

Show comment
Hide comment
@fperez

fperez Nov 10, 2011

Member

Yup, looks great, thanks. Will rebase now to avoid recursive merge on just one commit and will merge.

Member

fperez commented Nov 10, 2011

Yup, looks great, thanks. Will rebase now to avoid recursive merge on just one commit and will merge.

@fperez

This comment has been minimized.

Show comment
Hide comment
@fperez

fperez Nov 10, 2011

Member

MMh, auto-close seems not to be working... Closed by ec0c00d

Member

fperez commented Nov 10, 2011

MMh, auto-close seems not to be working... Closed by ec0c00d

@fperez fperez closed this Nov 10, 2011

@minrk

This comment has been minimized.

Show comment
Hide comment
@minrk

minrk Nov 10, 2011

Member

When you rebase and merge (identical to cherry-pick), autoclose doesn't work, because the hash changes.

Member

minrk commented Nov 10, 2011

When you rebase and merge (identical to cherry-pick), autoclose doesn't work, because the hash changes.

@fperez

This comment has been minimized.

Show comment
Hide comment
@fperez

fperez Nov 10, 2011

Member

No, but I amended the commit message and manually included a closes gh-991 precisely for that reason. So it should have auto-closed because of my message.

Member

fperez commented Nov 10, 2011

No, but I amended the commit message and manually included a closes gh-991 precisely for that reason. So it should have auto-closed because of my message.

@minrk

This comment has been minimized.

Show comment
Hide comment
@minrk

minrk Nov 10, 2011

Member

ah, that's definitely strange. Maybe they don't work if you put more than one on a line? I've always been in the habit of one-line per magic directive.

Member

minrk commented Nov 10, 2011

ah, that's definitely strange. Maybe they don't work if you put more than one on a line? I've always been in the habit of one-line per magic directive.

@fperez

This comment has been minimized.

Show comment
Hide comment
@fperez

fperez Nov 10, 2011

Member

I don't think it's that, the same thing happened just now with 3830949, and that was just one on a line. I still had to close it manually, their auto-closing is definitely acting up.

Member

fperez commented Nov 10, 2011

I don't think it's that, the same thing happened just now with 3830949, and that was just one on a line. I still had to close it manually, their auto-closing is definitely acting up.

@minrk

This comment has been minimized.

Show comment
Hide comment
@minrk

minrk Nov 10, 2011

Member

Right - and there should at least be those 'mentioned in commit' backrefs, even if the magic closing doesn't happen. I guess their commit hook processing is down/backed up right now.

Member

minrk commented Nov 10, 2011

Right - and there should at least be those 'mentioned in commit' backrefs, even if the magic closing doesn't happen. I guess their commit hook processing is down/backed up right now.

@fperez

This comment has been minimized.

Show comment
Hide comment
@fperez

fperez Nov 10, 2011

Member

yup, looks like it.

Member

fperez commented Nov 10, 2011

yup, looks like it.

@thisch

This comment has been minimized.

Show comment
Hide comment
@thisch

thisch Nov 10, 2011

Great, works as described in the commit msg! however, the solution for the notebook is not the best. I think that putting more effort into it is not necessary (atm) .

thisch commented on ba9389a Nov 10, 2011

Great, works as described in the commit msg! however, the solution for the notebook is not the best. I think that putting more effort into it is not necessary (atm) .

mdboom pushed a commit to mdboom/ipython that referenced this pull request Dec 2, 2011

Show traceback, continuing to start kernel if pylab init fails
Error message goes directly over stderr (and traceback over stdout),
rather than as a pyerr message, which is associated with a particular
frontend, and thus ignored because startup actions were not requested
by anyone.

This better matches the terminal behavior, where `ipython --pylab` without matplotlib
installed shows a traceback, but continues to start.

In the qtconsole the traceback will come in below the banner, but in the notebook
the message will not arrive until after the first execution.

closes gh-990, closes gh-991.

mattvonrocketstein pushed a commit to mattvonrocketstein/ipython that referenced this pull request Nov 3, 2014

Show traceback, continuing to start kernel if pylab init fails
Error message goes directly over stderr (and traceback over stdout),
rather than as a pyerr message, which is associated with a particular
frontend, and thus ignored because startup actions were not requested
by anyone.

This better matches the terminal behavior, where `ipython --pylab` without matplotlib
installed shows a traceback, but continues to start.

In the qtconsole the traceback will come in below the banner, but in the notebook
the message will not arrive until after the first execution.

closes gh-990, closes gh-991.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment