No such file or directory: '/tmp/ipython-....py'

Am 19.04.2015 um 17:59 schrieb Zeitlinie:
> Public bug reported:
>
> Hi,
>
> with the present python-mode revno: 1842 and vanilla Emacs 24.3.1 I am
> completely stuck with frequent and erratic occurrence error messages of
> type
>
> ---------------------------------------------------------------------------
> IOError Traceback (most recent call last)
> <ipython-input-30-f3cf7bacfa68> in <module>()
> ----> 1 exec(compile(open('/tmp/ipython-5828X6t.py').read(), '/tmp/ipython-5828X6t.py', 'exec')) # PYTHON-MODE
>
> IOError: [Errno 2] No such file or directory: '/tmp/ipython-5828X6t.py'
>
> These errors show up irregularly and very frequently on any set of commands submitted to the interpreter by C-c |
> Using emacs and ipython this way is impossible.
>

Hmm, may you send some example code, which triggers the bug?
All tests pass here so far.

Might be related to virtualenv resp. write-permissions.

There is no 'particular' code which generates this.
It happens on any code.
It happens totally random. The code may work for some time - and then it generates this error in large amounts.
What do you mean by virtualenv?
And where should I look for potential write-permission issues?

Just to add one more regarding write permissions.
Since it seems as if some files of type '/tmp/ipython-xyz.py' are missing:
my /tmp dir contains some such files ... some of which are also of recent date.
So it seems as if such files can be written into that location.
However it also seems as if not all of the necessary files are there ...
This is beyond me..

Am 22.04.2015 um 18:44 schrieb Zeitlinie:
> There is no 'particular' code which generates this.
> It happens on any code.
> It happens totally random.
That would be great. Until now computers couldn't provide randomness :-)

> The code may work for some time - and then it generates this error in large amounts.
> What do you mean by virtualenv?
https://pypi.python.org/pypi/virtualenv

> And where should I look for potential write-permission issues?
>

If not using a virtualenv, we can drop this question and look for at
other places.
I'm going to write a couple of tests aimed at the description, so maybe
it shows up.

Am 22.04.2015 um 18:55 schrieb Zeitlinie:
> Just to add one more regarding write permissions.
> Since it seems as if some files of type '/tmp/ipython-xyz.py' are missing:
> my /tmp dir contains some such files ... some of which are also of recent date.
> So it seems as if such files can be written into that location.
> However it also seems as if not all of the necessary files are there ...
> This is beyond me..
>

python-mode.el by defaul uses temporary filest.
However these should be deleted afterwards. Should you see many of them,
it might indicate an other bug.

BTW there is a way to avoid use of temp-files at environments without
write-permission. Which seems not the case here.

> python-mode.el by defaul uses temporary filest.
> However these should be deleted afterwards. Should you see many of them,
> it might indicate an other bug.

Yes, you are right, they get deleted automatically, and my /tmp dir does not contain many, but only 1,2, or lately even 3 of them.

> BTW there is a way to avoid use of temp-files at environments without
> write-permission. Which seems not the case here.

How would I do that anyway?

Am 24.04.2015 um 15:55 schrieb Zeitlinie:

> BTW there is a way to avoid use of temp-files at environments without
> write-permission. Which seems not the case here.
> How would I do that anyway?
>

set py-execute-no-temp-p to non-nil

However, it's not recommended.

I'm suffering from this exact problem, and it's happening all the time.

Forgot to mention that I've tried (setq py-execute-no-temp-p t), but it doesn't help at all.

After pulling the latest development bzr version, I no longer have this problem. Thanks a lot!

I have the same problem as Zeitlinie, and it occurs SPORADICALLY. That is, e.g., when I try executing the same region two times, it may succeed on one attempt and not on the other. Obviously the state may somehow have changed between the two attempts and so it may not be 'random' as such. I have verified that the file in question gets written, but it seems as though it is erased (by python-mode) before it is executed!

Using python-mode 6.2.1

IOError Traceback (most recent call last)
<ipython-input-31-b539482515e8> in <module>()
----> 1 exec(compile(open('/tmp/ipython-18915s8X.py').read(), '/tmp/ipython-18915s8X.py', 'exec')) # PYTHON-MODE

IOError: [Errno 2] No such file or directory: '/tmp/ipython-18915s8X.py'

With an error happening randomly as told, fixing will be difficult. Need some glue WRT circumstances. Anyway - will have another look ASAP.

Dear Andreas,
I'd be very willing to have a chat with you on the phone if that can help you identify and fix this. I'd very much appreciate the fix, because I use emacs and ipython a lot and have been getting frustrated with trying to fix this by using different combinations of versions of emacs, ipython, and python-mode etc... and trying to hack python-mode.el
Thanks

  On 12.11.2015 22:33, Balu Nadiga wrote:
> Dear Andreas,
> I'd be very willing to have a chat with you on the phone if that can help you identify and fix this. I'd very much appreciate the fix, because I use emacs and ipython a lot and have been getting frustrated with trying to fix this by using different combinations of versions of emacs, ipython, and python-mode etc... and trying to hack python-mode.el
> Thanks
>

Thanks back. Unfortunately, without having some description WRT to
circumstances that would not help.
Helpful would be valid but failing tests.