Closed georgevreilly closed 10 years ago
jorgenschaefer
commented
11 years ago Just like good old usenet! :-D Thank you for the bug report.
The traceback says that Elpy is trying to get a calltip, and then the backend hangs.
Which backend are you using? (Rope, Jedi, native) Also, which version of Python? (2.x, 3.x)
georgevreilly
commented
11 years ago I have
jedi==0.5b5
rope==0.9.4
elpy==1.0.1according to pip freeze. I have no explicit elpy initialization in ~/.emacs.d/init.el beyond (package-initialize) and (elpy-enable)
I'm running Python 2.7.4, as installed by brew
jorgenschaefer
commented
11 years ago Ok, if you haven't changed anything, then M-: (elpy-rpc-get-backend) should return "rope".
What's your project root set to (M-: (elpy-project-root))? Is the the root directory of the git repository of pip2pi? Or are there other files/directories/anything in there?
I tried to reproduce this problem, but it seems to simply work for me.
georgevreilly
commented
11 years ago "rope""~/src/pip2pi"Now I can't reproduce it either, though it repro'd several times last night. I even restarted Emacs a couple of times last night, since I thought at first it might be Evil but I disabled that, and it continued to happen. I reverted commands.py to the version that caused me problems last night, but it still doesn't repro.
If it happens again, I'll report it. If it does happen, is there other data that I should gather?
jorgenschaefer
commented
11 years ago Do check the project root again to make sure it didn't get confused about that.
You can check with ps xar or top which process it is that causes CPU consumption (should be the python process).
You can use strace -p <pid> to see what the python process is doing (most likely, nothing, as that only shows syscalls and it will probably be in an infinite loop in Rope somewhere; but maybe it got confused about the project root and is indexing your home directory for possible Python files …)
If it's not doing anything in that trace, you can also use strace -p <pid> -o foo.txt before reproducing the bug, and then post the trace output here. That might give some hint on what it was doing.
That's all very much stabbing in the dark, though. I wish there was a simple way to hook into a running Python process to get some idea on what it was doing when it did hang. :-|
georgevreilly
commented
11 years ago It repro'd a couple of times in the last half hour. https://gist.github.com/georgevreilly/5745407 has samples from both Emacs.app and the child Python process.
jorgenschaefer
commented
11 years ago That's strace output on OS X? Fascinating :-D
As Python is stuck in PyEval_EvalFrameEx, I suspect it's indeed stuck evaluating some Rope code.
Any idea what you are doing to reproduce this?
georgevreilly
commented
11 years ago Actually, that's the output from the built-in Activity Monitor. There's no strace, but there is DTrace. Maybe I'll finally learn how to use it.
Every time this has happened, I've been doing something innocuous like moving up and down lines. Obviously, moving around in buffers is something I do thousands of times a day. This CPU utilization problem only started in the last few days. I can't think of anything's that's changed.
jorgenschaefer
commented
11 years ago My current idea would be that it's a bug in Rope that we can't reproduce, so difficult to debug or report upstream. :-|
As a workaround, you can use the Jedi backend (M-x elpy-set-backend RET jedi RET). If that also causes those hangs, something else is seriously broken, but I do not know what.
Sorry I can't be of more help here. :-(
Nikratio
commented
11 years ago If there's a possibility to upgrade to Python 3.3, then you could use the faulthandler module to get a (Python level) stack trace by sending a signal to the Python process when the freeze occurs. That would tell you where exactly it is hanging.
jorgenschaefer
commented
11 years ago Any new ideas on what has been causing this?
georgevreilly
commented
11 years ago I had forgotten about this, as I have not noticed it for several weeks.
On Fri, Jul 19, 2013 at 11:21 AM, Jorgen Schäfer notifications@github.comwrote:
Any new ideas on what has been causing this?
— Reply to this email directly or view it on GitHubhttps://github.com/jorgenschaefer/elpy/issues/96#issuecomment-21267696 .
jorgenschaefer
commented
11 years ago I'll close this, then – if it shows up again, do not hesitate to re-open the issue! And thanks for the extensive reports.
glasserc
commented
10 years ago Hi, I've seen this issue as well. Emacs is definitely the process that spikes the CPU. I can't definitively assert that it's due to elpy, but I thought I'd keep you informed. The "proximate cause" that causes the CPU to spike is usually cursor movement or hitting C-x C-f (I use ido). The emacs stack traces tend to just show "redisplay_internal" as the top stack frame. I have my machine set up to use the jedi backend by default, but I've explicitly changed it to Rope; I feel like that works sometimes, but eventually I end up back in a slow state. I'm mostly working on the codebase for my job, which is about 6000 lines of Python (so not that big), but pulls in some thirty or forty packages in a virtualenv. I've tried both using and not using virtualenv-workon to specify the virtualenv. Another quirk of our codebase is that we append the project root to PYTHONPATH, which I have had to do manually when working with virtualenv.
Here's the base64-encoded profiler-report of doing C-x C-f and then aborting with C-g a few times.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I have python3.3 on my machine so I could try to run elpy using it (even though our codebase is definitely not py3k safe). Is there an environment variable I could change?
jorgenschaefer
commented
10 years ago Thank you!
Which version of Emacs are you using?
glasserc
commented
10 years ago I'm running:
jorgenschaefer
commented
10 years ago Pity. They recently broke something in Emacs trunk that might be causing this. But 24.3 should unaffected.
I'm pretty much at a loss here. The profile output doesn't contain any elpy code. Hanging in redisplay should be a C problem, too. I'd love to say "can't reproduce, unlikely to be elpy's fault", but that would be cheating a bit. I'll keep the issue open in case someone can find more information here.
glasserc
commented
10 years ago Is there some elpy-minor-mode that is suspect? Somethiing I could try in order to rule something out?
jorgenschaefer
commented
10 years ago The only thing I can imagine would be the RPC interface doing something really weird. Disabling both auto-complete and eldoc-mode would mean Elpy does not use RPC while moving in a buffer. If you can semi-reliably reproduce this, disabling both of those would be an idea.
glasserc
commented
10 years ago Hmm. I haven't hit the problem since disabling those two minor modes. Is there some way to narrow down a cause, or maybe get a log or something?
jorgenschaefer
commented
10 years ago Hm. I have a suspicion.
I added one (remotely) possible bugfix and some debug output to the issue-96 branch. Could you use that? (It'll be a bit slower and spam your message log, though; might make more sense to log to a separate buffer)
If it still hangs, the debug output should show whether it's running some kind of endless loop with the backend there somehow.
glasserc
commented
10 years ago Hmm, took me a little while to figure out how to co-opt package to install elpy@issue-96, but I think I've done it -- I moved elpy-1.2.1 out of the way, symlinked both elpy-1.2.1 and elpy-1.2.50 to elpy-issue96, and copied elpy-autoloads.el from elpy-1.2.1 to elpy-issue96. I'll let you know how it goes!
jorgenschaefer
commented
10 years ago Oh. You should be able to just add the elpy repository to your load-path and do a (require 'elpy) (and add2virtualenv for the repository to get the python module, too).
glasserc
commented
10 years ago Hmm. I've been using it all day and at least it hasn't spiked the CPU. But I just got "Flymake: Configuration error has occurred while running (flake8 /home/ethan/..../filename_flymake.py). Flymake will be switched OFF." I get it every time I open a Python file (i.e. C-x C-f to a file that I haven't already C-x C-f'd to in this session).
From having looked at the changes you introduced, I can't imagine that it's related. Just for fun, I checked *Messages* and didn't see anything from Flymake (though I do see the elpy-rpc--call messages).
glasserc
commented
10 years ago Sorry, this is even weirder than that -- it only shows up in files in certain directories. It looks like the _flymake files are being created correctly with the correct contents, too. It's almost certainly unrelated.
glasserc
commented
10 years ago Ah, this is definitely unrelated, it's something to do with my codebase (I have a file called stat.py, which seems to conflict somehow with the stdlib's stat.py).
glasserc
commented
10 years ago Actually, just to clarify a little more in case this turns out to be related in some way to the larger problem -- for some reason I have PYTHONPATH=:/home/ethan/some-directory, and flake8 is getting run from the directory with stat.py. flake8, when run in exactly this way, imports os, which imports posixpath, which imports stat, which finds stat because it's first on PYTHONPATH.
jorgenschaefer
commented
10 years ago That's a different bug definitely. I'm currently on vacation and won't be able look any closer at the moment, could you open a separate issue so I won't forget?
eN-Joy
commented
10 years ago Well, I kind of having the same problem, however it occurs randomly and I just can't reproduce it. Basically emacs would work most of the time, especially when it just started. Then after a while, when it tries to auto-complete, it eats up 100% CPU, you'll have to wait a long time before it becomes responsive to key stroke, and you'll end up with this cycle until emacs crashes.
I have tried to elliminate some minor modes, such as eldoc-mode, flymake-mode, auto-complete-mode, etc, no help. I've set toggle-debug-on-error, or toggle-debug-on-quit, but emacs quits and creates a dumps, I can't see logs.
Here's my environment: Python 2.7.5 (default, Oct 2 2013, 22:34:09) [GCC 4.8.1] on cygwin
dot emacs (I have a very simple setup) (add-to-list 'load-path "~/.emacs.d")
;; Python mode in elpa is not the newest (add-to-list 'load-path "~/.emacs.d/python-mode.el-6.1.2") (load-theme 'wombat t)
(require 'package) (setq package-archives '(("gnu" . "http://elpa.gnu.org/packages/") ("marmalade" . "http://marmalade-repo.org/packages/") ("melpa" . "http://melpa.milkbox.net/packages/"))) (package-initialize)
(require 'python-mode) (add-hook 'python-mode-hook 'show-paren-mode) (add-hook 'python-mode-hook 'electric-pair-mode)
(when (require 'elpy nil t) (elpy-enable) (elpy-clean-modeline)) (elpy-use-ipython)
Django==1.6.1 Scrapy==0.20.2 Twisted==13.2.0 beautifulsoup4==4.3.2 cssselect==0.9.1 django-debug-toolbar==1.0 elpy==1.2.1 epc==0.0.5 flake8==2.1.0 ipython==1.1.0 jedi==0.7.0 lxml==3.2.4 mccabe==0.2.1 pep8==1.4.6 pyOpenSSL==0.13.1 pyflakes==0.7.3 queuelib==1.1.1 requests==2.1.0 sexpdata==0.0.3 simplejson==3.3.1 six==1.4.1 sqlparse==0.1.10 w3lib==1.5 wsgiref==0.1.2 zope.interface==4.0.5
Sample emacs dump
Stack trace: Frame Function Args 00000427030 0018006F943 (001802DFEA0, 001802DFEA0, 00000427368, 001802E4D28) 00000000006 00180070E0A (00100004000, 00000000000, 00000000234, 00000000000) 00000427210 00180119138 (5E86A32000000006, 00000000000, 00100935481, 0060002D568) 000000000C1 001801161EE (00000427780, 00000000000, 00000000000, 00100488200) 001004881A0 001801166BB (00000000000, 00000000000, 00000000000, 00000000006) 001004881A0 0018011688C (00000000079, 0000042D440, 001801750AB, 001004881A0) 001004881A0 0018011227B (0000042D440, 001801750AB, 001004881A0, 00000000020) 001004881A0 000004275DC (001801750AB, 001004881A0, 00000000020, 00100488200) 001004881A0 00000000079 (001004881A0, 00000000020, 00100488200, 001008C49E0) 001004881A0 0000042D440 (00000000020, 00100488200, 001008C49E0, 00000000001) 001004881A0 001801750AB (00000000001, 000004275DC, 00180137010, 00600E3A400) 001004881A0 0010048819A (00000006D3F, 1FFFFFFFFFFFFFFF, 0000000000C, 001008FB5E6) 001004881A0 0010049FFB3 (001004DE720, 00600BDE460, 0010083A640, 00100FAEE58) 00000000002 00180070E0A (001008DD432, 001008C2AA9, 001004F72A3, 00000000000) 00000427A00 00180112391 (001008C49E0, 001008C49E0, 00600C9ADD8, 00100D92612) 00100C818F2 00000427A48 (001008C49E0, 00600C9ADD8, 00100D92612, 00600F648A6) End of stack trace (more stack frames may be present)
jorgenschaefer
commented
10 years ago (require 'python-mode)
Hm. Elpy doesn't work with python-mode.el (and actually should throw an error when used in this configuration). Could you re-try without?
Also, which version of Emacs is that?
eN-Joy
commented
10 years ago Yes, I've tried to switch to built-in python.el, it didn't help either. Here a screen shot of the cpu usage. http://imgur.com/zqtsEIB
(However I've never received complaints from using python-mode.el)
My emacs version: GNU Emacs 24.3.1 (x86_64-unknown-cygwin) of 2013-08-20 on moufang Copyright (C) 2013 Free Software Foundation, Inc.
jorgenschaefer
commented
10 years ago I have no idea why Emacs would hang. elpy doesn't do anything to make it hang to my knowledge. It even breaks out of the busy-lock after a few seconds if it uses that, and for auto-complete/eldoc, it doesn't even busy-block. :-(
glasserc
commented
10 years ago Well, let me chime in -- using the "hacked" version of elpy with the minor fix and the logging turned on, I haven't seen elpy hang, so I think that the fix at least helps a lot. I do have one remaining complaint, but it doesn't cause emacs to hang or CPU to spike. Hmm, I think I better open another bug report for that, actually.
jorgenschaefer
commented
10 years ago The original issue might be related to #154 and #182 – though the user was reporting to have rope installed, which would by default mean they used rope, not jedi. The ticket was reopened, but that user reports the problem solved for now. So closing this.
If this resurfaces with rope, please do not hesitate to re-open this ticket.
gyst
commented
10 years ago FWIW I ran into this problem (Emacs hanging at 100% CPU w/ elpy on cursor up/down) today after I renamed a module called 'stat.py' to 'stats.py' (and was working in this module). Following the hints in this thread I renamed that module to a more innocuous name 'describe.py'. That seems to have solved the problem (for me). Using elpy with rope and flycheck.
I cannot reproduce the problem though by re-creating stat.py and/or stats.py.
jorgenschaefer
commented
10 years ago @gyst, which versions of Emacs and Elpy are you using?
gyst
commented
10 years ago On 04/05/14 12:08, Jorgen Schäfer wrote:
@gyst https://github.com/gyst, which versions of Emacs and Elpy are you using?
GNU Emacs 24.3.1 elpy-1.3.0
gyst@thor:~$ ls .emacs.d/elpa/ archives/ highlight-indentation-0.5.0/ auto-complete-1.4/ idomenu-0.1/ autofit-frame-readme.txt iedit-0.97/ color-theme-20080305.34/ less-css-mode-20140205.250/ color-theme-solarized-20140401.957/ nose-0.1.1/ dash-20140327.116/ pkg-info-20140313.513/ elpy-1.3.0/ popup-0.5/ epl-20140316.544/ pyvenv-1.1/ f-20140220.21/ s-20131223.944/ find-file-in-project-3.2/ sr-speedbar-20131207.2049/ flycheck-20140328.743/ web-mode-20140325.1631/ flycheck-color-mode-line-20131125.2138/ yasnippet-0.8.0/ fuzzy-0.1/
Guido Stevens | +31.43.3618933 | http://cosent.nl
s o c i a l k n o w l e d g e t e c h n o l o g yThank you. Really strange. And it's Emacs that is consuming CPU?
I hate these bugs being so difficult to analyze :-)
gyst
commented
10 years ago On 04/05/14 18:54, Jorgen Schäfer wrote:
Thank you. Really strange. And it's Emacs that is consuming CPU?
Yes it was. Stopped responding, basically, so I looked at top.
I hate these bugs being so difficult to analyze :-)
I felt really bad about submitting a report when I couldn't even reproduce it myself. But then I thought, the other guy had stat problems, moving that module worked for me, it might work for the next guy who's hit by this.
Guido Stevens | +31.43.3618933 | http://cosent.nl
s o c i a l k n o w l e d g e t e c h n o l o g yNever feel bad about submitting bug reports. Thank you!
gyst
commented
10 years ago On 04/05/14 19:44, Jorgen Schäfer wrote:
Never feel bad about submitting bug reports. Thank you!
Well, thank you! Elpy is a major upgrade of my dev environment :-)
Guido Stevens | +31.43.3618933 | http://cosent.nl
s o c i a l k n o w l e d g e t e c h n o l o g yI am still experiencing the cpu problem with my new Ubuntu LTS 14.04 box, it's a minimal installation, absolutely only necessary packages are installed.
GNU Emacs 24.3.1, Python 2.7.6, iPython 2.0.0,
;; (autoload 'jedi:setup "jedi" nil t) ;; (setq jedi:setup-keys t) ;; (add-hook 'python-mode-hook 'jedi:setup) ;; (setq jedi:complete-on-dot t)
(when (require 'elpy nil t) (elpy-enable) (elpy-clean-modeline)) (elpy-use-ipython)
(add-hook 'python-mode-hook 'show-paren-mode) (add-hook 'python-mode-hook 'electric-pair-mode)
backports.ssl-match-hostname==3.4.0.2 beautifulsoup4==4.3.2 cffi==0.8.2 cryptography==0.4 cssselect==0.9.1 Django==1.6.4 elpy==1.4.0 epc==0.0.5 filemagic==1.6 flake8==2.1.0 ipython==2.0.0 jedi==0.8.0-final0 log-colorizer==1.6 lxml==3.3.5 mccabe==0.2.1 pep8==1.5.6 pip-tools==0.3.4 psutil==2.1.1 psycopg2==2.5.2 pycparser==2.10 pyflakes==0.8.1 pyOpenSSL==0.14 queuelib==1.1.1 Scrapy==0.22.2 sexpdata==0.0.3 six==1.6.1 South==0.8.4 tornado==3.2.1 Twisted==14.0.0 uWSGI==2.0.4 w3lib==1.5 wdb.server==2.0.0 wdb==2.0.0 zope.interface==4.1.1
archives auto-complete-20140414.2324 autopair-20140311.1211 concurrent-20140303.1635 ctable-20140304.1659 deferred-20140303.1634 elpy-20140501.744 epc-20140323.2054 ess-20140507.2352 find-file-in-project-20140221.557 flymake-cursor-20130822.1032 fuzzy-20131025.2343 highlight-indentation-20131207.835 idomenu-20111122.1048 iedit-20131021.116 jedi-20140321.1323 nginx-mode-20140321.2028 nose-20140501.1740 pastebin-20101125.1955 php-mode-20140502.1051 pony-mode-20140127.846 popup-20140207.1702 python-environment-20140321.1116 pyvenv-20140411.1319 web-mode-20140511.1206 yasnippet-20140511.947
The CPU spike is totally not predictable, I can sometimes use emacs for a while not seeing anything, yet sometimes it's just start hanging no matter what you type. The picture below shows when I was in a parenthesis.
jorgenschaefer
commented
10 years ago @eN-Joy, could you enable M-x toggle-debug-on-quit, hit C-g (possibly a few times) when Emacs hangs like this, and post the resulting backtrace? If that does not produce a backtrace, could you run strace -o foo.txt -f -p 20744 (the PID of the Emacs process), let that run for a few seconds, abort it with C-c, and post the resulting foo.txt?
eN-Joy
commented
10 years ago @jorgenschaefer sorry it took me some time to get back. M-x toggle-debug-on-quit does not always spit messages, so i've tried the strace way, the dumped file is about 500k, i don't think it can be pasted here. http://ix.io/dbU please let me know what else i can provide. (i promise i'll respond quicker;-)
jorgenschaefer
commented
10 years ago Thanks for the trace! That looks really weird.
The start of the trace is the normal loop of Emacs that I know: pselect/clock_gettime/rt_sigprocmask. Then it somehow looks like it simply forgets to call pselect, which is what it would need to do to notice input is pending.
This really looks like a bug in Emacs. What version are you using?
You probably should report this using M-x report-emacs-bug
eN-Joy
commented
10 years ago I am using GNU Emacs 24.3.1 on Ubuntu 14.04. Will try report bug later. Thanks!
gyst
commented
10 years ago On 29/06/14 18:09, eN-Joy wrote:
I am using GNU Emacs 24.3.1 on Ubuntu 14.04.
That is exactly the version I'm using as well. Hardly a coincidence.
| This is GNU Emacs 24.3.1 (x86_64-pc-linux-gnu, GTK+ Version 3.10.7) | of 2014-03-07 on lamiak, modified by Debian
Guido Stevens | +31.43.3618933 | http://cosent.nl
s o c i a l k n o w l e d g e t e c h n o l o g yTesting the latest pre-release might be an idea if you have the time. This looks a bit like a bug we traced down in Emacs a few weeks ago. Though the behavior was a bit different (only minor hiccups, no real lock-ups).
gyst
commented
10 years ago On 30/06/14 15:56, Jorgen Schäfer wrote:
Testing the latest pre-release might be an idea if you have the time.
Prerelease of elpy, or of emacs?
Guido Stevens | +31.43.3618933 | http://cosent.nl
s o c i a l k n o w l e d g e t e c h n o l o g y@gyst, oops, sorry for being unclear, I meant the Emacs pre-test.
When I move around in the
dir2pi()function inlibpip2pi/commands.pyof pip2pi, I reliably get 100% CPU utilization. Elpy 1.0 on Emacs 24.3.1 on OS X 10.7I set
debug-on-quitand got the traceback below. It's gzipped, then base64-encoded. Sorry, but when I pasted the raw text in, Github Issues failed to save it—undoubtedly because of the 29,000-character lines. And there's no way to attach an arbitrary file.