damnwidget / anaconda Goto Github PK

Sorry to pelt you with Issues, but there's a keybindings collision under OS X, where Super+F is normally mapped to "Find." It's easily corrected, except that---this again---Package Control will restore the defaults if it encounters a change. It'd be great if you could add a Default (OSX).sublime-keymap file that doesn't use Super; personally, I've changed it to:

[
    {"command": "anaconda_goto", "keys": ["ctrl+alt+g"]},
    {"command": "anaconda_find_usages", "keys": ["ctrl+alt+f"]},
    // {"command": "sublime_jedi_params_autocomplete", "keys": ["("]}
]

Thanks again! I've had a chance to use the plugin a bit, and it's just fantastic so far.

Anaconda/Sublime still autocompletes as I type, despite the following in my settings file:

{
    "suppress_word_completions": true,
    "suppress_explicit_completions": true,
}

Hi,

I'm running ST3 with Anaconda downloaded via Package Control on Ubuntu 13.04 64-bit machine.

I was installing some plugins to fresh ST3, after I had installed Anaconda I felt my machine slowing down.

Here's what happens after I run ST3:

(At least it doesn't grow exponentially)

Process responsible for eating my resources is plugin_host.
I use default Anaconda settings.

I have no idea how to provide you with more useful information. Any hints on that?

Here's my ST3 user configuration: https://gist.github.com/pbanaszkiewicz/6261690

I have noticed that no suggestions are displayed after a from - import statement.

For example if I type "from os." a list of possibility gets displayed so for example I can choose "path".
However if I type "from os import " nothing is shown until I type a letter. For example, in this case, if I type "from os import p" the the a list of options gets displayed.

The problem here is that not always I know what kind of things I can import from a module and if Anaconda could display them, it would be a massive help for the user.

I installed anaconda using git clone, and there are no menu entries for the Default or User configs

Hi all.

Anaconda had been growing a lot in just one month and is already big enough and messy enough to need a complete refactor. I am not happy with the actual structure of the plugin and its directories hierarchy.

We have a jedi, linting and asynconda packages and then those anaconda_whatever in the root of the plugin, thats obviously wrong and hard to maintain its lack of logic just ashame me with the strength of one million burning suns.

The anaconda.py file has too many responsabilities, anaconda_linter.py needs love. I am moving the code to a modular structure.

The plugin is just entering in a lock state until the refactor is complete. After the refactor it will be updated to v1.0 and it should be considered the first total stable release. That means we will have a new development branch and the stable master branch after we reach that point.

Regards.

Occasionally, Anaconda seems to lose it's mind and just continuously register new handlers. Not sure what triggers it, but my ST console looks a bit like this:

Writing file /home/dblanchard/contextual_feature_generation/code/features/char_count.py with encoding UTF-8 (atomic)
Writing file /home/dblanchard/contextual_feature_generation/code/features/char_count.py with encoding UTF-8 (atomic)
Registering handler with address ('localhost', 34908)
Registering handler with address ('localhost', 54091)
Registering handler with address ('localhost', 37572)
Registering handler with address ('localhost', 33339)
Registering handler with address ('localhost', 34997)
Registering handler with address ('localhost', 44606)
Registering handler with address ('localhost', 53905)
Registering handler with address ('localhost', 53905)
Registering handler with address ('localhost', 33110)
Registering handler with address ('localhost', 39281)
Registering handler with address ('localhost', 57144)
Registering handler with address ('localhost', 38357)
Registering handler with address ('localhost', 57975)
Registering handler with address ('localhost', 42316)
Registering handler with address ('localhost', 42562)
Registering handler with address ('localhost', 36738)
Registering handler with address ('localhost', 57380)
Registering handler with address ('localhost', 46412)
Registering handler with address ('localhost', 43626)
Registering handler with address ('localhost', 33005)
Registering handler with address ('localhost', 46580)
[FindKeyConflicts] ERROR [Errno 111] Connection refused
[FindKeyConflicts] ERROR Traceback (most recent call last):
  File "/home/dblanchard/.config/sublime-text-3/Packages/Anaconda/worker.py", line 77, in start
    worker['client'] = AsynClient(worker['port'])
  File "/home/dblanchard/.config/sublime-text-3/Packages/Anaconda/anaconda_client/jsonclient.py", line 34, in __init__
    EventHandler.__init__(self, ('localhost', port))
  File "/home/dblanchard/.config/sublime-text-3/Packages/Anaconda/asynconda/ioloop.py", line 105, in __init__
    self.sock.connect(address)
ConnectionRefusedError: [Errno 111] Connection refused
Registering handler with address ('localhost', 52198)
Registering handler with address ('localhost', 42297)
Registering handler with address ('localhost', 37509)
Registering handler with address ('localhost', 56559)
Registering handler with address ('localhost', 47862)
Registering handler with address ('localhost', 51919)
Registering handler with address ('localhost', 56524)
Registering handler with address ('localhost', 35307)
Registering handler with address ('localhost', 35214)
Registering handler with address ('localhost', 45863)
Writing file /home/dblanchard/contextual_feature_generation/code/features/__init__.py with encoding UTF-8 (atomic)
Registering handler with address ('localhost', 50726)
Registering handler with address ('localhost', 35685)
Registering handler with address ('localhost', 54345)
Registering handler with address ('localhost', 45350)
Registering handler with address ('localhost', 50899)
Registering handler with address ('localhost', 41594)
Registering handler with address ('localhost', 48021)
Registering handler with address ('localhost', 49774)
Registering handler with address ('localhost', 49774)
Writing file /home/dblanchard/contextual_feature_generation/code/features/__init__.py with encoding UTF-8 (atomic)
Writing file /home/dblanchard/contextual_feature_generation/code/features/__init__.py with encoding UTF-8 (atomic)
Registering handler with address ('localhost', 49516)
Registering handler with address ('localhost', 34121)
Registering handler with address ('localhost', 58553)
Registering handler with address ('localhost', 37590)
Registering handler with address ('localhost', 51802)
Registering handler with address ('localhost', 48151)
Registering handler with address ('localhost', 50663)
Registering handler with address ('localhost', 41187)
Registering handler with address ('localhost', 51685)
Registering handler with address ('localhost', 51685)
Registering handler with address ('localhost', 48064)
Registering handler with address ('localhost', 57503)
Registering handler with address ('localhost', 54914)
Registering handler with address ('localhost', 56900)
Registering handler with address ('localhost', 34748)
Registering handler with address ('localhost', 46947)
Registering handler with address ('localhost', 38202)
Registering handler with address ('localhost', 58752)
Writing file /home/dblanchard/contextual_feature_generation/code/features/__init__.py with encoding UTF-8 (atomic)
Registering handler with address ('localhost', 54521)
Registering handler with address ('localhost', 46548)
Registering handler with address ('localhost', 45259)
Registering handler with address ('localhost', 55291)
Registering handler with address ('localhost', 51306)
Registering handler with address ('localhost', 46559)
Writing file /home/dblanchard/contextual_feature_generation/code/features/__init__.py with encoding UTF-8 (atomic)
Registering handler with address ('localhost', 34246)
Registering handler with address ('localhost', 55651)
Registering handler with address ('localhost', 38246)
Registering handler with address ('localhost', 42679)
Registering handler with address ('localhost', 53920)
Registering handler with address ('localhost', 36629)
Registering handler with address ('localhost', 44083)
Registering handler with address ('localhost', 51277)
Registering handler with address ('localhost', 36073)
Registering handler with address ('localhost', 47453)
Registering handler with address ('localhost', 60459)
Registering handler with address ('localhost', 49095)
Registering handler with address ('localhost', 40981)
Registering handler with address ('localhost', 56924)Writing file /home/dblanchard/contextual_feature_generation/code/features/__init__.py with encoding UTF-8 (atomic)

Registering handler with address ('localhost', 45194)
Registering handler with address ('localhost', 56058)
Registering handler with address ('localhost', 38981)
Registering handler with address ('localhost', 39336)
Registering handler with address ('localhost', 42889)
Registering handler with address ('localhost', 52803)
Registering handler with address ('localhost', 44408)
Registering handler with address ('localhost', 37437)
Registering handler with address ('localhost', 55848)
Writing file /home/dblanchard/contextual_feature_generation/code/features/__init__.py with encoding UTF-8 (atomic)
Registering handler with address ('localhost', 57807)
Registering handler with address ('localhost', 41355)
Registering handler with address ('localhost', 38685)
Registering handler with address ('localhost', 54679)
Registering handler with address ('localhost', 41917)
Registering handler with address ('localhost', 54073)
Registering handler with address ('localhost', 56764)
Registering handler with address ('localhost', 50480)
Registering handler with address ('localhost', 46454)
Registering handler with address ('localhost', 34342)
Registering handler with address ('localhost', 36023)
Registering handler with address ('localhost', 41870)
Writing file /home/dblanchard/contextual_feature_generation/code/features/__init__.py with encoding UTF-8 (atomic)
Registering handler with address ('localhost', 55665)
Registering handler with address ('localhost', 50915)
Registering handler with address ('localhost', 44020)
Registering handler with address ('localhost', 44020)
Registering handler with address ('localhost', 44268)
Registering handler with address ('localhost', 35989)
Registering handler with address ('localhost', 35989)
Registering handler with address ('localhost', 50635)
Registering handler with address ('localhost', 40523)
Registering handler with address ('localhost', 37222)
Registering handler with address ('localhost', 38680)
Registering handler with address ('localhost', 48674)
Registering handler with address ('localhost', 55849)
Writing file /home/dblanchard/contextual_feature_generation/code/features/__init__.py with encoding UTF-8 (atomic)
Registering handler with address ('localhost', 50662)

I just noticed the bit about "FindKeyConflicts" in the error log, so it might be the combination that makes things go awry. I'll try removing that and see if it helps.

I was using this package just fine yesterday, but today I noticed it wasn't working. I saw this error in the log:

error: uncaptured python exception, closing channel <Anaconda.anaconda_client.jsonclient.AsynClient connected localhost:58040 at 0x1047eef90> (<class 'BlockingIOError'>:[Errno 35] Resource temporarily unavailable [X/asynchat.py|handle_read|116] [X/asyncore.py|recv|379])

I then quit ST3 and restarted and got a popup saying package control could not download Anaconda. This was the error in the log:

Package Control: Error downloading package. HTTP error 500 downloading https://codeload.github.com/DamnWidget/anaconda/zip/master.
error: Package Control

I tried removing and reinstalling the package but get the same error.

I'm using Sublime Text Build 3047 on OS X 10.7.5

As the title says - the package doesn't come up in the list, for me and others.

Would like to try it out!

how to set max-line-length for pep8?

It's been a while I've been coding in Sublime and since that time I've been looking for a nice autocompletion plugin. I tried SublimeJedi for some time but it behaves so annoyingly sometimes that after a short time, I disabled it and never looked back. I was using just SublimeLinter for linting.

After a friend pointed me here, it was it some distrust that I tried Anaconda. Fortunately, it revealed itself to be a fine project with consistent behavior that works just fine for my needs. And I don't even need the SublimeLinter anymore, who was causing me some slowdowns in the recent times.

Well, this is just a big thank you for the author(s) of this project. Well done!

If you try to use a Python 3 interpreter jsonserver just crash as reported here #10 (comment)

It would be nice to have the same JediSublime autocompletion for classes and functions arguments, also with the same settings behaviour.

Knowing what arguments a function or a class accepts and what are their default values (if any) is quite important.
For sure the possibility to look at the documentation that Anaconda offers, make this one almost a non-issue, but as improvement I think it will be appreciated by many users

Would be nice if there is an option to show a popup window with list of error on linting and go to the line by just selecting from the list

I'm not sure when it happened, but I'm getting errors when trying to plug into a python3 interpreter. The following shows up on the ST3 console:

File "/Users//Library/Application Support/Sublime Text 3/Packages/Anaconda/anaconda_lib/worker.py", line 71, in start
self.start_json_server(worker['port'])
File "/Users//Library/Application Support/Sublime Text 3/Packages/Anaconda/anaconda_lib/worker.py", line 93, in start_json_server
self.build_server(port)
File "/Users//Library/Application Support/Sublime Text 3/Packages/Anaconda/anaconda_lib/worker.py", line 188, in build_server
self.json_server = subprocess.Popen(args, **kwargs)
File "X/subprocess.py", line 818, in init
File "X/subprocess.py", line 1416, in execute_child
FileNotFoundError: [Errno 2] No such file or directory: 'python3'
Package Control: Skipping automatic upgrade, last run at 2013-09-03 21:46:56, next run at 2013-09-03 22:46:56 or after
Traceback (most recent call last):
File "/Applications/Sublime Text.app/Contents/MacOS/sublime_plugin.py", line 526, in run
return self.run()
File "/Users//Library/Application Support/Sublime Text 3/Packages/Anaconda/commands/get_lines.py", line 17, in run
self._harvest_errors(errors, 'ERRORS')
File "/Users//Library/Application Support/Sublime Text 3/Packages/Anaconda/commands/get_lines.py", line 44, in _harvest_errors
for line, error_strings in ANACONDA[error_type][vid].items():
KeyError: 31

The Jedi log contains the following:

2013-09-03 21:52:08,369: DEBUG : bind: address=('localhost', 60120)
2013-09-03 21:52:08,369: DEBUG : listen: backlog=5
2013-09-03 21:52:08,369: INFO : Anaconda Server started in port 60120 for PID 23197 with cache dir /Users//Library/Caches/Jedi/grumpy_dropbox and extra paths /usr/local/lib/python3.3/site-packages/numpy,/usr/local/lib/python3.3/site-packages/scipy,/usr/local/lib/python3.3/site-packages/pandas,/Users//Dropbox/Code/Python3/grumpy
2013-09-03 21:52:08,374: INFO : Incomming connection from ('127.0.0.1', 60299)
2013-09-03 21:52:08,375: INFO : Incomming connection from ('127.0.0.1', 60300)
2013-09-03 21:52:08,378: INFO : client requests: run_linter
2013-09-03 21:52:08,482: INFO : client requests: run_linter
2013-09-03 21:52:13,314: INFO : client requests: run_linter
2013-09-03 21:52:13,901: INFO : client requests: run_linter
2013-09-03 21:52:28,036: INFO : client requests: run_linter
2013-09-03 21:52:28,953: INFO : client requests: run_linter
2013-09-03 21:52:29,930: INFO : client requests: run_linter
2013-09-03 21:52:39,138: INFO : client requests: run_linter
2013-09-03 21:52:39,381: INFO : client requests: run_linter
2013-09-03 21:52:51,990: INFO : client requests: run_linter
2013-09-03 21:52:52,020: INFO : client requests: run_linter
2013-09-03 21:52:53,439: INFO : client requests: run_linter
2013-09-03 21:52:56,661: INFO : client requests: run_linter
2013-09-03 21:52:56,688: INFO : client requests: run_linter
2013-09-03 21:53:02,918: INFO : client requests: run_linter
2013-09-03 21:53:03,527: INFO : client requests: run_linter
2013-09-03 21:53:04,234: INFO : client requests: run_linter
2013-09-03 21:53:06,555: INFO : client requests: run_linter
2013-09-03 21:53:12,014: INFO : client requests: run_linter
2013-09-03 21:53:16,933: INFO : client requests: run_linter
2013-09-03 21:53:17,649: INFO : client requests: run_linter
2013-09-03 21:53:18,444: INFO : client requests: run_linter
2013-09-03 21:53:18,570: INFO : client requests: run_linter
2013-09-03 21:53:42,785: INFO : client requests: run_linter
2013-09-03 21:54:29,992: INFO : client requests: run_linter
2013-09-03 21:54:33,797: INFO : client requests: run_linter
2013-09-03 21:54:34,509: INFO : client requests: run_linter
2013-09-03 21:54:34,676: INFO : client requests: run_linter
2013-09-03 21:54:41,299: INFO : client requests: autocomplete
2013-09-03 21:54:47,928: INFO : client requests: run_linter
2013-09-03 21:54:47,938: INFO : client requests: run_linter
2013-09-03 21:54:47,948: INFO : client requests: run_linter
2013-09-03 21:54:47,958: INFO : client requests: run_linter
2013-09-03 21:54:50,363: INFO : client requests: run_linter
2013-09-03 21:54:57,381: INFO : client requests: run_linter
2013-09-03 21:54:57,711: INFO : client requests: run_linter
2013-09-03 21:55:26,074: INFO : client requests: run_linter
2013-09-03 21:55:28,293: INFO : client requests: autocomplete
2013-09-03 21:55:28,836: INFO : client requests: run_linter
2013-09-03 21:55:29,689: INFO : client requests: autocomplete
2013-09-03 21:55:30,156: INFO : client requests: run_linter
2013-09-03 21:55:30,332: INFO : client requests: run_linter
2013-09-03 21:55:34,496: INFO : client requests: run_linter
2013-09-03 21:55:35,197: INFO : client requests: run_linter
2013-09-03 21:55:37,932: INFO : client requests: run_linter
2013-09-03 21:55:38,841: INFO : client requests: run_linter
2013-09-03 21:55:39,101: INFO : client requests: run_linter
2013-09-03 21:56:15,058: INFO : client requests: run_linter
2013-09-03 21:56:15,957: INFO : client requests: run_linter
2013-09-03 21:56:24,985: INFO : client requests: run_linter
2013-09-03 21:56:25,063: INFO : client requests: run_linter
2013-09-03 21:57:25,178: INFO : process 23197 does not exists stopping server...
2013-09-03 21:57:26,179: INFO : Closing the socket, server will be shutdown now...

By the way, thanks for maintaining this package. It's great.

Hi,

StringIO is no longer included in Python 3. anaconda_pylint.py has:

try:
    from cStringIO import StringIO
except ImportError:
    from StringIO import StringIO
    assert StringIO

Would it be possible to use io.StringIO instead?

I've got multiple files and multiple windows open. Some of the files are linting correctly, but one is not (and it's not the last one that I opened). From my log:

Registering handler with address ('localhost', 53253)
Unhandled exception in poll, restarting the poll request
[Errno 35] Resource temporarily unavailable


ERROR:root:[Errno 61] Connection refused
ERROR:root:Traceback (most recent call last):
  File "~/Library/Application Support/Sublime Text 3/Packages/Anaconda/worker.py", line 88, in start
  File "~/Library/Application Support/Sublime Text 3/Packages/Anaconda/anaconda_client/jsonclient.py", line 34, in __init__
EventHandler.__init__(self, ('localhost', port))
  File "~/Library/Application Support/Sublime Text 3/Packages/Anaconda/asynconda/ioloop.py", line 104, in __init__
ConnectionRefusedError: [Errno 61] Connection refused

Add to anaconda support to wrapper virtualenv and virtualenvwrapper tools and commands. Add new features like virtualenv replication, and automatic creation of virtualenvs with dependencies using configuration templates.

The main idea is be able to configure python interpreters that are installed and running in vagrant boxes using channeled SSH connections.

I came to this idea through the issue #51

pylint 1.0 recently came out and it is far and away the most thorough linter, so it would be really nice to have support for that in anaconda.

There was an excellent ST2 project called Pylinter that supported it, but the author is no longer responding to pull requests, so I think it's dead.

It takes seconds for the auto-complete pop-up to show. This is probably due to the following error on console:

[Errno 10061] No connection could be made because the target machine actively refused it

Hi, at first thank for adding autopep to you plugin.
Can you consider to add option for running auto format before save? I have trying this myself but fail ;( I've bound auto format to shortcut but it is annoying press extra shortcut before save all time :)

Add minimal debugging tools so the developer can place breakpoints and inspect variables (for example) using anaconda builder.

Investigating about the recent reports about problems with anaconda in Windows and Mac I discover that there is some problem that is not being written into the logs with python3 configured interpreters.

This is a critic bug because it is making the plugin completely unusable to those people using Python3.

SublimeLinter has a quick option for disabling linting on the current view:

Disable Linting - Disables linting mode for the current view and clears all lint errors.

Is there any way Anaconda can support a similar option? This is very helpful when opening others code temporarily without having to fix all the issues just to be able to read the file.

Thanks!

Under Mac OS X 10.8.4 and Sublime Text 3 build 3052, [Errno 61] Connection Refused errors are logged to the console consistently as I edit Python scripts. Linting is working beautifully, and method and statement names are being autocompleted, including methods from external modules, so I'm not wholly sure what's missing. I've closed all my Sublime windows except for one script, but that didn't help. No information aside from the error message is being logged to the console. Happy to help debug it, but not quite sure where to start.

It is a nice wrapper around pyflakes and pep8: https://pypi.python.org/pypi/flake8

Actually, anaconda doesn't support completion for files that resides in the project (or folder) that we are working on (if it's not installed in the PYTHONPATH). It doesn't support completion on opened buffers neither.

Should be nice to add completions for all of them.

This is a teeny-tiny little thing, but when the package is installed through Package Control, at least on OS X, the directory it creates is called Packages/Anaconda, not Packages/anaconda. This causes imports of anaconda.foo to fail under ST3, because Python apparently expects the case to match. Renaming the folder fixed the problem, but it might be worth correcting in the repo. Thanks so much for what looks like it'll be fantastic package!

I'm having trouble switching the default python interpreter, none of the IDE features or linting works once I point it to a python3 interpreter. I'm on OSX so the default settings points to the system install of python2.7

I've installed python3 with homebrew so I've tried both linking to the homebrew symlink and the actual binary with no success.

When I'm typing inside a string, it tries to autocomplete too, something that I find very annoying.

Example:

sql = """
    DELETE FROM dummy_table WHERE name = %s
"""

When I type the s in %s it gives me a list of whatever starts with 's' in the file making type a simple sql quite annoying sometimes.

I see this message in ST3's console everytime I start it:

Traceback (most recent call last):
  File "/Applications/Sublime Text.app/Contents/MacOS/sublime_plugin.py", line 157, in on_api_ready
    m.plugin_loaded()
  File "/Volumes/Home/hamid/Library/Application Support/Sublime Text 3/Packages/Anaconda/anaconda.py", line 382, in plugin_loaded
    template = Template(tplfile.read())
  File "X/encodings/ascii.py", line 26, in decode
UnicodeDecodeError: 'ascii' codec can't decode byte 0xe2 in position 1471: ordinal not in range(128)

I am running ST3 on OSX and the Python interpreter is system's default: Python 2.7.2

The anaconda JsonServer has to be rewritten following the new modular scheme.

Although jsonserver is asynchronous, it runs library calls blocking itself.

Hello,

I tried to use Anaconda but whatever I do I got this error multiple times per action:

[WinError 10061] No connection could be made because the target machine actively refused it

I have no idea what I have to do here.

Thanks,
Dan

    "python_interpreter_path": "/home/ekarlso/projects/billingstack/.venv/bin/python",
    "extra_paths": [
        "/home/ekarlso/projects/billingstack/.venv/lib/python2.7/site-packages"
    ]

Shouldn't this be correct?

Hi, i believe here is a bug :)

../Anaconda/commands/autoformat.py

line 44: 'ignore': get_settings(self.view, 'ignore', []),
should be:
line 44: 'ignore': get_settings(self.view, 'autoformat_ignore', []),

same for select property

Hi DW,

What's the expected behaviour from the auto-completer?

In my main Sublime settings I have:

{
    "auto_complete": true,
    "auto_complete_triggers": [
        {
            "selector": "source.python -string -comment -constant",
            "characters": ".",
        }
    ],
    "auto_complete_selector": "-"
}

and in the Anaconda settings, I have:

{
    "suppress_word_completions": true,
    "suppress_explicit_completions": true
}

Project settings are:

{
    "settings": {
        "python_interpreter": "/Users/michael/.virtualenv/parse/bin/python"
    }
}

I have openpyxl installed in my virtualenv. If I open a new project, and add a new python file, and type:

from openpyxl import load_workbook
wb = load_workbook(filename="file.xlsx")

and then typing wb. only gives completions for words that exist in the document (e.g. import, load_workbook, etc.)

If I type the same code into a command line python REPL, typing wb. gives a full and complete list of all methods and attributes in the wb object (i.e. get_index, get_named_range, and so on).

Thanks!

I detected that in some files the lint process makes the server unresponsive and forces the client to reconnect up (starting a new server process in a new port) leaving the old server process stale in the system.

reloading plugin Anaconda.__init__
reloading plugin Anaconda.anaconda
cProfile doesn't seems to can be imported on ST3 + darwin, sorry.You may want to use @timeit instead, so sorry really
reloading plugin Anaconda.anaconda_linter
reloading plugin Anaconda.contexts
reloading plugin Anaconda.decorators
cProfile doesn't seems to can be imported on ST3 + darwin, sorry.You may want to use @timeit instead, so sorry really
reloading plugin Anaconda.utils
reloading plugin Anaconda.worker

I am thinking this is a Mac OSX exclusive issue?

When I have Anaconda enabled, none of my built-in python snippets are shown in autocomplete. With Anaconda disabled, they do show up.

Anyone else seeing this?

ST3 Latest, Mac OS X 10.8

Jedi library has been updated recently to fix some bugs and problems introduced in the last major update.

Actually, anaconda uses asyncore and asynchat from the standard python library to perform its asynchronous network operations. It works but is forcing the plugin to use timeouts for the underlying select system call and that's not good.

In some random situations, the anaconda sockets based IPC stop working and it's almost impossible to determine where the problme is (refer to the troubleshooting sectionin the README.md file).

In the POSIX manual pages of select you we can read:

    1.  You should always try to use select() without a timeout.  Your program should have nothing to do if there is no data available.  Code that depends on timeouts is not usually portable and is difficult to debug.

Asyncore loop really sucks, it blocks the execution (wating on select) when you try to use it from ST3 because the loop has to run in a different execution thread (or it will block the plugin_host interpreter) and it supports for threading is so bad (just a shared dictionary that is not populated in any way from outside asyncore and is almost unmaintained inside) so the read and write structures that select has to use in order to perform socket IO operations has to just wait 30 seconds (that is the default) cos asyncore-asynchat.does not support any type of data population from outside it own scope and context making select unusable until you set a really low timeout for the asyncore.loop function (that sets the timeout in the underlying select system call).

I was considering other alternatives like Facebook's tornado but use that is like kill flies with canons so I decide to create my own asynchornous sockets library just to fit anaconda's needs.

Some work is already needed to fit event listeners to the new modular scheme.

When I modify file and close it without saving, whether I select 'Yes' or 'No' in dialog Sublime fires, there is IndexError in console coming from anaconda. Only difference between choosing 'Yes' and 'No' is that if I choose 'Yes' I get same stacktrace 3 times.

This is tracktrace:

Traceback (most recent call last):
  File "C:\Program Files\Sublime Text 3\sublime_plugin.py", line 311, in on_activated_async
    callback.on_activated_async(v)
  File "C:\Users\del-boy\AppData\Roaming\Sublime Text 3\Packages\Anaconda\anaconda_lib\decorators.py", line 72, in wrapper
    if is_python(view):
  File "C:\Users\del-boy\AppData\Roaming\Sublime Text 3\Packages\Anaconda\anaconda_lib\decorators.py", line 42, in is_python
    location = view.sel()[0].begin()
  File "C:\Program Files\Sublime Text 3\sublime.py", line 450, in __getitem__
    raise IndexError()
IndexError

It looks like that if file is closing, sel() method returns no regions, so indexing failes.

Anaconda sometimes tries to lint non-python files like *.md or *.m files. There are no accompanying status-bar messages, but some of the lines will be outlined in a box. It doesn't always happen, but seems to happen fairly consistently if I restart ST3.

ST3 Build 3047, OS X 10.7.5

I just noticed running the AutoPEP8 in a really really bad file from one of my work mates that the AutoPEP8 formatting takes forever (it doesn't block the gui but is pretty annoying).

I need to do something about this, maybe make the buffer read-only and add a message with a running bar in the status bar.

Now is soooooo wrong.

consider to add autopep8 support, that would be nice :)

I installed Anaconda this morning (wow, amazing plugin!) but I work with a combination of python and SQL. Unfortunately, it looks like the linter is running on files of a type other than Python. Is there a way to prevent this?
Thanks!