Anaconda mode opening new windows with errors

[badders]

Generally the autocompletion seems to be working ok and popping up with correct methods and the documentation, but it constantly opens a new emacs window with:

status: (:error (error connection-failed failed with code 10061 : host 127.0.0.1 :service 9000))

point 1

It wouldn't be so bad but focus jumps to that window every time

Emacs 25, 64 bit, latest anaconda-mode, setup is the default spacemacs configuration

[badders]

New window is titled anaconda-response if that helps

[proofit404]

I need the value of python-shell-interpreter and python-shell-virtualenv-path variables. Does minibuffer contain some error messages? What is the content of *anaconda-mode* buffer?

[badders]

python-shell-interpreter: ipython
python-shell-virtualenv-path: nil
anaconda-mode buffer contains just: anaconda_mode port 9000

no message in minibuffer

[badders]

Note I don't set either of those variables myself, they are just the defaults on this system

[proofit404]

Does this happens when you switch buffers? Does this error happens when you constantly typing within one buffer.

[proofit404]

Do you use pyenv tool?

[badders]

Dont use pyenv, the python i have installed is the anaconda distribution from continuum, version 2.5 (latest). This error happens when just opeining one python file and doing something like:

import pandas as pd
import numpy as np

pd.read_ or np.arra and waiting for completion

Usually it works for the first 1 or 2 completions, then fails.

[proofit404]

Does M-: (anaconda-mode-running-p) evaluates to nil after 1 or 2 completions?

[badders]

Nope, after i get the status error window pop up it evaluates to:

(run open listen connect stop)

[vladimir-vovk]

Same issue appears randomly when I typing...

# status: (:error (error connection-failed failed with code 35
 :host 127.0.0.1 :service 9000))
# point: 1

emacs 24.5, spacemacs 0.105.21

[proofit404]

I have no idea what can cause this problem.

[shackra]

I'm experiencing this as well:

# status: (:error (error connection-failed failed with code 111
 :host localhost :service 8118))

8118 is the port of my local Privoxy installation, I have it set as a environment variable http_proxy so the entire system knows about it.

[proofit404]

@shackra Set no_proxy variable as well to exclude localhost from proxy settings.

[shackra]

For those curious, @proofit404 talks about this https://www.gnu.org/software/emacs/manual/html_node/url/Proxies.html

[proofit404]

Exactly. Thanks for clarification!

[shackra]

The error persists even thought no_proxy is set and Emacs successfully returns the variable definition with (getenv "no_proxy")

[proofit404]

What is the value of url-proxy-services variable?

[shackra]

url-proxy-services is a variable defined in `url-vars.el'.
Its value is (("no_proxy" . "\\(localhost\\|192\\.168\\..*\\..*\\)")
 ("http" . "localhost:8118"))

Original value was nil

[proofit404]

Try to set no_proxy to localhost only and see if it works.

[shackra]

the issue is still alive and kicking, even with the change you requested me to introduce:

url-proxy-services is a variable defined in `url-vars.el'.
Its value is (("no_proxy" . "\\(localhost\\)")
 ("http" . "localhost:8118"))

Original value was nil

[proofit404]

No idea how to fix this.

[shackra]

😵

El 26 de junio de 2016 3:21:23 GMT-06:00, Malyshev Artem notifications@github.com escribió:

No idea how to fix this.

---

You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub:
#164 (comment)

Sent from my Android device with K-9 Mail. Please excuse my brevity.

[badders]

Just to add that i also had the no_proxy set for localhost when getting this issue

[churtado]

I'm also getting this issue

[aumana]

Same issue using emacs prelude

[proofit404]

Does anyone have an idea where I can find connection failure codes table?

[binjo]

FTR, try restart the emacs, which should restart the releated process, the error gone for me, I didn't look into root cause...

[proofit404]

@binjo you don't need to restart text editor. Just use M-x (anaconda-mode-stop) RET

[3rd3]

Same problem through some Spacemacs configuration.

[xarthurx]

Same problem through spacemacs configuration.
And when closing emacs, it shows there are may 127.0.0.1 buffer open that tries to connect with the command (network connection to 127.0.0.1)

[invokesus]

Having the same problem any suggestions?

[sulami]

I noticed somewhat severe lag when editing a ~3.5k lines python source file, and the error buffer popping up and stealing focus every couple of words. Both went away when I just nuked ~3k lines from the file, so it might just be a timeout issue when it's trying to lookup autocompletion suggestions or something like that.

[proofit404]

I can't find any description of the connection error codes used in Emacs. You can help if you research this question.

[sulami]

I this error code 35 was coming from the syscall, it would be EDEADALK, "Resource deadlock would occur". I don't know anything about the internal structure of anaconda-mode, but could it be the case that it's trying to open a connection before a previous one is finished or something similar, causing a conflict? Because I can very quickly reproduce this issue by typing random characters into python files with >3k lines, where everything especially autocompletion starts to take a second or two to return suggestions.

[proofit404]

I suspect it's come from parallel querying overloaded python process.

[invokesus]

#190 This is the issue I seem to be facing. It's trying to use the proxy to access localhost even though no_proxy is set. Setting http_proxy to "" works but no_proxy doesn't.

[skrech]

I have the same problem. However, I think that it's related to anaconda-eldoc-mode.
If I remove (add-hook 'python-mode-hook 'anaconda-eldoc-mode) from my init file everything seems to be working OK.

[proofit404]

Eldoc make queries as you move point. So it is still a problem with parallel queries.

[eserge]

To me it only happens when I edit large python files. So the suspicion about overloaded python processes seems reasonable.

[penn201500]

(remove-hook 'anaconda-mode-response-read-fail-hook
'anaconda-mode-show-unreadable-response)

seems like solved this problem

syl20bnr/spacemacs#3772

[skrech]

Just wanted to add that when the problem happens this is what list-processes is showing:

127.0.0.1       open    --              --           (network connection to 127.0.0.1)
127.0.0.1<1>    open     *http 127.0... --           (network connection to 127.0.0.1)
127.0.0.1<2>    open     *http 127.0... --           (network connection to 127.0.0.1)
127.0.0.1<3>    open     *http 127.0... --           (network connection to 127.0.0.1)
127.0.0.1<4>    open     *http 127.0... --           (network connection to 127.0.0.1)
127.0.0.1<5>    open     *http 127.0... --           (network connection to 127.0.0.1)
anaconda-mode   run     *anaconda-mode* --           python -c 
import sys, site
site.addsitedir('.')
import anaconda_mode
anaconda_mode.main(sys.argv[1:])

When anaconda-mode is working OK (for example without eldoc mode support) there is only one connection opened for the localhost.

[paddyhoran]

Same issue here.

Fix mentioned by @penn201500 does not work for me, still have the issue.

[LZGH]

@paddyhoran That doesn't work for me too. I sovle it by revise anaconda-mode.el source file,find "(add-hook 'anaconda-mode-response-read-fail-hook 'anaconda-mode-show-unreadable-response)" and remove it,then you need to recompile anaconda-mode.el or delete anaconda-mode.elc directly,then it worked.

[paddyhoran]

@LZGH, thanks very much. That worked.

[shivams]

Instead of localhost, try 127.0.0.1 for no_proxy. This worked for me.

[proofit404]

Error messages are not propagated to windows anymore.