Repository navigation
ipython_directive causing OperationalError: attempt to write a readonly database #8850
Description
I have a minimal sphinx project generated by sphinx_quickstart and with the conf.py edited only to add:
extensions = ['IPython.sphinxext.ipython_console_highlighting',
'IPython.sphinxext.ipython_directive']
I added a file ipy_eg.rst containing:
.. ipython::
In [1]: x = 1
Running make html builds a correct-looking html page, but always exits with this error:
build succeeded, 1 warning.
Error in atexit._run_exitfuncs:
Traceback (most recent call last):
File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/atexit.py", line 24, in _run_exitfuncs
func(*targs, **kargs)
File "/Users/mb312/.virtualenvs/regreg/lib/python2.7/site-packages/IPython/core/interactiveshell.py", line 3394, in atexit_operations
self.history_manager.end_session()
File "/Users/mb312/.virtualenvs/regreg/lib/python2.7/site-packages/IPython/core/history.py", line 547, in end_session
len(self.input_hist_parsed)-1, self.session_number))
OperationalError: attempt to write a readonly database
Error in sys.exitfunc:
Traceback (most recent call last):
File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/atexit.py", line 24, in _run_exitfuncs
func(*targs, **kargs)
File "/Users/mb312/.virtualenvs/regreg/lib/python2.7/site-packages/IPython/core/interactiveshell.py", line 3394, in atexit_operations
self.history_manager.end_session()
File "/Users/mb312/.virtualenvs/regreg/lib/python2.7/site-packages/IPython/core/history.py", line 547, in end_session
len(self.input_hist_parsed)-1, self.session_number))
sqlite3.OperationalError: attempt to write a readonly database
I'm running ipython current master, and my history file is writeable.
$ ipython --version
4.1.0-dev
$ ls -al $(ipython locate)/profile_default/history.sqlite
-rw-r--r-- 1 mb312 staff 15002624 Sep 29 21:57 /Users/mb312/.ipython/profile_default/history.sqlite
Activity
jenshnielsen commentedon Sep 30, 2015
I have seen this too building the matplotlib docs. I believe it happened when upgrading to 4.0. Perhaps fixed by #8846?
minrk commentedon Sep 30, 2015
It's probable that the directive should be using
:memory:by default, anyway. I doubt people really want sphinx builds to be populating their command history.This wouldn't be fixed by #8846, since that only catches the creation of the session, whereas the reported traceback doesn't seem to fail until the last operation.
minrk commentedon Oct 9, 2015
I think this should be fixed by #8888, which forces the history to be in-memory when using the directive.
tacaswell commentedon Nov 20, 2015
#8405 was also an attempt to fix this.