vulture and pylint failing #73
Comments
|
@hschilling use the prospector argument |
|
Hi Jeff, Thanks for the suggestion. Maybe our project size is stretching the limits ? Thanks for the help Traceback (most recent call last): On Dec 4, 2014, at 6:42 PM, Jeff Quast <notifications@github.commailto:notifications@github.com> wrote: @hschillinghttps://github.com/hschilling use the prospector argument --die-on-tool-error and it may help discover the root issue — Herb Schilling |
|
Suggest this issue be closed, prospector can't help you much here. However I filed a bug on your behalf with astroid: https://bitbucket.org/logilab/astroid/issue/65/oserror-errno-24-too-many-open-files you may watch this issue for any resolution or response. In the meantime, I suggest finding a way to increase the number of open files for your user account. |
|
Hi, I see. Thanks for filing that bug for me. I did try ( I am using Mac OS X 10.8.5 ) sudo launchctl limit maxfiles 1000000 1000000 and got the same error. I monitored the process using Activity Monitor's Open Files pane and I saw that it just kept growing as it opened up source files but it sure didn't look like 1000000 files in there ! On Dec 6, 2014, at 6:34 PM, Jeff Quast <notifications@github.commailto:notifications@github.com> Suggest this issue be closed, prospector can't help you much here. However I filed a bug on your behalf with astroid: https://bitbucket.org/logilab/astroid/issue/65/oserror-errno-24-too-many-open-files you may watch this issue for any resolution or response. In the meantime, I suggest finding a way to increase the number of open files for your user account. — Herb Schilling |
|
That doesn't work for some reason. Just type: $ ulimit -n 16384 on your shell. 1000000 is also too high, 16384 is the actual upperbound limit? |
|
This was fixed upstream, Pylint 1.4.1 should resolve the issue. prospector is currently bound to |
|
If you are interested, I have a cheap workaround for installing and running prospector using tox, removing the upperbound, and upgrading pylint and astroid: tox.ini excerpt: requirements-analysis.txt |
|
Hi Jeff, Thanks for this and also for letting me know about the fix to Pylint ! On Dec 15, 2014, at 11:53 AM, Jeff Quast <notifications@github.commailto:notifications@github.com> wrote: If you are interested, I have a cheap workaround for installing and running prospector using tox, removing the upperbound, and upgrading pylint and astroid: tox.ini excerpt: [testenv:prospector] cheap workaround until prospector's next release#75commands = pip install --upgrade astroid pylint requirements-analysis.txt prospector[with_dodgy,with_frosted,with_mccabe,with_pep257,with_pep8,with_pylint,with_pyroma,with_vulture] — Herb Schilling |
|
Finally got around to making a 0.8 release, sorry about the delay. #10 and #74 turned out to be a bit more fiddly than expected. Let me know if this fixes these issues - it should do as @jquast by the way you can use |
|
@hschilling if you upgrade to prospector 0.8, this should hopefully fix your issue. |
|
Thanks ! Will give it a try over the weekend |
|
I tried it. I did install everything using: pip install prospector[with_everything] When I just do prospector in a directory with a super simple Python script, it works. If I do the same in our OpenMDAO project directory, I get this. Traceback (most recent call last): The value that gets returned from autodetect_libraries(self.workdir) is: Let me know if I should try something else. Sorry! |
|
I'm submitting a PR for this shortly, this broke for me too |
|
submitted new issue #82 for tracking |
|
Fixed in PR #85 |
|
Thanks! I tried it again and things are working much better. The only problem left is I see this in the output at the top of the output file. I ran this with --with-tool pyroma --with-tool vulture. There is nothing in the output that tells me more about the exception. I can seem to run pylint and pyroma without errors Messages. |
|
Don't forget, use argument |
|
@hschilling I tried this myself on the OpenMDAO-Framework repository and pylint worked fine, so if you could use Pyroma fails because one of the |
|
Sorry, I DID forget about --die-on-tool-error :-) Pylinit dies with "RuntimeError: maximum recursion depth exceeded" Here is most of the stack trace. I cut out the part where it has a LOT of these lines repeated over and over again. File "/Users/herbertschilling/Documents/OpenMDAO/dev/latest/devenv/lib/python2.7/site-packages/astroid/bases.py", line 341, in wrapped Traceback (most recent call last): |
|
BTW, the usage: output for prospector looks like it could be tweaked. Missing --die-on-tool-error and the alternate ways of using --with-tool,etc... usage: prospector [-h] [-0] [-A] [-u USES] [-B] [--no-common-plugin] [-D] [-T] |
|
I do have numpy installed. This is why pyroma dies for me. Looks like you just need to use if '_setuptools' in data: instead of if data['_setuptools']: In fact, if I make that code change, everything runs for me if I --without-tool pylint Traceback (most recent call last): and data is equal to (Pdb) data |
|
@hschilling I've seen the recursion depth thing before but I think in an older version of astroid. Can you do a 'pip freeze' and show the versions of astroid, logilab-common and pylint? For pyroma, it's the same stacktrace as the one I got, but the problem is in pyroma itself, so I'll need to make a PR on that. |
|
@carlio Thanks. Here is the output of pip freeze: Cython==0.20.1 |
|
@hschilling thanks, unfortunately those are exactly the versions I was using so that's not the issue. I tried it again on my Mac this time, and I get a similar problem - I don't get a stacktrace but prospector is taking forever and when I Ctrl-C I see a very deep stacktrace at the same point you did. This doesn't happen on my Ubuntu machine. Very odd. Out of interest, how long does it take to get the stacktrace? |
|
47 seconds |
|
It just finished for me, and no stack trace or problem. I'm 100% baffled why it's not working for you. |
|
Wait, ignore that! Of course, if I don't have all the dependencies installed, then the pylint inference will be less intense. I'll try installing everything and see what happens. |
|
I installed as much as I could, but some things were not available - for example, Traits 4.3.0 is no longer available for download. I think in general though that the problem is simply 'too much code'. Pylint builds up a map of all code in the project but also follows inference chains for libraries used, so there will be so much going on. This is basically a limitation of pylint/astroid rather than a bug in prospector. I'm not sure exactly what could be done either, without severe restructuring to avoid recursing when inferring types. The only solution I have would be to run prospector several times on different parts of the code. |
|
Makes sense to me. BTW, we are trying to get rid of Traits as a dependency. Good to know you can't even download it any more! |
|
(i have a love/hate relationship with traits too) the reason it doesn't find it is because enthought does a very poor job of publishing it, this is what pip sees: https://pypi.python.org/simple/traits/ -- as you see, 4.3.0 just redirects to a website where no download is found, you can probably safely upgrade to latest traits. |
|
I'll close this as there's not much I can do to fix it and #93 covers what can be fixed. |
|
That's fine. Thanks for the help ! On Jan 10, 2015, at 11:16 PM, Carl Crowder <notifications@github.commailto:notifications@github.com> I'll close this as there's not much I can do to fix it and #93#93 covers what can be fixed. — Herb Schilling |
Prospector already solves the issues in landscapeio/prospector#73
Prospector solved the issues in landscapeio/prospector#73
For a very large code base ( https://github.com/OpenMDAO/OpenMDAO-Framework ) , I get this in the output
.:
L-:- None: pylint - failure
Tool pylint failed to run (exception was raised)
.:
L-:- None: vulture - failure
Tool vulture failed to run (exception was raised)
There are no errors given by either tool as a result.
I tried using pylint and vulture with some smaller codes and it worked fine.
Great tool though!!!
The text was updated successfully, but these errors were encountered: