No serverpy #324

pjz · 2014-04-04T13:23:29Z

Fix #283 - remove server.py, fix code that relied on it, and remove execution.py also.

Also fix up the testing client to instantiate Website instead of Server

chadwhitacre · 2014-04-22T15:42:30Z

I tried serving the docs so I could review the docs:

[aspen]$ python build.py docs
Traceback (most recent call last):
  File "env/bin/aspen", line 9, in <module>
    load_entry_point('aspen==0.30-dev', 'console_scripts', 'aspen')()
  File "/Users/whit537/personal/gittip/aspen-python/env/lib/python2.7/site-packages/pkg_resources.py", line 353, in load_entry_point
    return get_distribution(dist).load_entry_point(group, name)
  File "/Users/whit537/personal/gittip/aspen-python/env/lib/python2.7/site-packages/pkg_resources.py", line 2321, in load_entry_point
    return ep.load()
  File "/Users/whit537/personal/gittip/aspen-python/env/lib/python2.7/site-packages/pkg_resources.py", line 2048, in load
    entry = __import__(self.module_name, globals(),globals(), ['__name__'])
ImportError: No module named server
fabricate: 'aspen' exited with status 1
[aspen]$

chadwhitacre · 2014-04-22T16:00:03Z

It's weird that we still use argv to configure even though we aren't a cli anymore. I guess that gets ripped out after this lands.

pjz · 2014-04-22T16:04:39Z

Yeah, it's currently ignored, though I guess it doesn't have to be - other modules take args from the cmdline. As long as we're also configurable from the environment and/or a config file, it's fine, I think.

chadwhitacre · 2014-04-24T19:42:28Z

pid-37703 thread-140735299662208 (MainThread) project_root set to /Users/whit537/personal/gittip/aspen-python/doc/.aspen.
Traceback (most recent call last):
  File "/Users/whit537/lib/python2.7/runpy.py", line 162, in _run_module_as_main
    "__main__", fname, loader, pkg_name)
  File "/Users/whit537/lib/python2.7/runpy.py", line 72, in _run_code
    exec code in run_globals
  File "/Users/whit537/personal/gittip/aspen-python/aspen/__main__.py", line 26, in <module>
    make_server('0.0.0.0', 8080, Website()).serve_forever()
  File "aspen/website.py", line 35, in __init__
  File "aspen/configuration/__init__.py", line 275, in configure
  File "/Users/whit537/personal/gittip/aspen-python/env/lib/python2.7/site-packages/pkg_resources.py", line 2048, in load
    entry = __import__(self.module_name, globals(),globals(), ['__name__'])
  File "/Users/whit537/personal/gittip/aspen-python/env/lib/python2.7/site-packages/aspen_tornado_renderer.py", line 2, in <module>
    from aspen import renderers
ImportError: cannot import name renderers
fabricate: 'python' exited with status 1

pjz · 2014-04-25T01:15:45Z

Replicable from the commandline. Do python build.py clean docs, watch it fail, then to manually see it fail again do ./env/bin/python -m aspen -a:5370 -wdoc -pdoc/.aspen.

There's a hint in the fact that a simple ./env/bin/python -m aspen -a:5370 does NOT fail, but I'm too out of it to dig it out right now.

chadwhitacre · 2014-04-25T01:19:05Z

Good find!

pjz · 2014-04-25T01:21:00Z

Also, putting an absolute path as an arg to -w doesn't help either.

chadwhitacre · 2014-04-30T17:52:11Z

Confirmed that ./env/bin/python -m aspen -a:5370 doesn't fail while ./env/bin/python -m aspen -a:5370 -wdoc -pdoc/.aspen does.

chadwhitacre · 2014-04-30T17:52:34Z

python -m aspen -a:5370 -wdoc fails!

chadwhitacre · 2014-04-30T17:53:29Z

[aspen]$ python -m aspen -a:5370 
> /Users/whit537/personal/gittip/aspen-python/aspen/__main__.py(27)<module>()
-> make_server('0.0.0.0', 8080, Website()).serve_forever()
(Pdb) import sys
(Pdb) pp sys.path
['',
 '/Users/whit537/personal/gittip/aspen-python',
 '/Users/whit537/lib/python',
 '/Users/whit537/personal/gittip/aspen-python/lib/python',
 '/Users/whit537/personal/gittip/aspen-python/env/lib/python27.zip',
 '/Users/whit537/personal/gittip/aspen-python/env/lib/python2.7',
 '/Users/whit537/personal/gittip/aspen-python/env/lib/python2.7/plat-darwin',
 '/Users/whit537/personal/gittip/aspen-python/env/lib/python2.7/plat-mac',
 '/Users/whit537/personal/gittip/aspen-python/env/lib/python2.7/plat-mac/lib-scriptpackages',
 '/Users/whit537/personal/gittip/aspen-python/env/lib/python2.7/lib-tk',
 '/Users/whit537/personal/gittip/aspen-python/env/lib/python2.7/lib-old',
 '/Users/whit537/personal/gittip/aspen-python/env/lib/python2.7/lib-dynload',
 '/Users/whit537/lib/python2.7',
 '/Users/whit537/lib/python2.7/plat-darwin',
 '/Users/whit537/lib/python2.7/lib-tk',
 '/Users/whit537/lib/python2.7/plat-mac',
 '/Users/whit537/lib/python2.7/plat-mac/lib-scriptpackages',
 '/Users/whit537/personal/gittip/aspen-python/env/lib/python2.7/site-packages']
(Pdb)

chadwhitacre · 2014-04-30T17:54:21Z

IRC, btw.

chadwhitacre · 2014-04-30T17:54:46Z

[aspen]$ python -m aspen -a:5370 -wdoc
> /Users/whit537/personal/gittip/aspen-python/aspen/__main__.py(27)<module>()
-> make_server('0.0.0.0', 8080, Website()).serve_forever()
(Pdb) import sys
(Pdb) pp sys.path
['',
 '/Users/whit537/personal/gittip/aspen-python',
 '/Users/whit537/lib/python',
 '/Users/whit537/personal/gittip/aspen-python/lib/python',
 '/Users/whit537/personal/gittip/aspen-python/env/lib/python27.zip',
 '/Users/whit537/personal/gittip/aspen-python/env/lib/python2.7',
 '/Users/whit537/personal/gittip/aspen-python/env/lib/python2.7/plat-darwin',
 '/Users/whit537/personal/gittip/aspen-python/env/lib/python2.7/plat-mac',
 '/Users/whit537/personal/gittip/aspen-python/env/lib/python2.7/plat-mac/lib-scriptpackages',
 '/Users/whit537/personal/gittip/aspen-python/env/lib/python2.7/lib-tk',
 '/Users/whit537/personal/gittip/aspen-python/env/lib/python2.7/lib-old',
 '/Users/whit537/personal/gittip/aspen-python/env/lib/python2.7/lib-dynload',
 '/Users/whit537/lib/python2.7',
 '/Users/whit537/lib/python2.7/plat-darwin',
 '/Users/whit537/lib/python2.7/lib-tk',
 '/Users/whit537/lib/python2.7/plat-mac',
 '/Users/whit537/lib/python2.7/plat-mac/lib-scriptpackages',
 '/Users/whit537/personal/gittip/aspen-python/env/lib/python2.7/site-packages']
(Pdb)

chadwhitacre · 2014-04-30T17:55:57Z

So it looks like sys.path is the same for both. os.getcwd() is also the same.

chadwhitacre · 2014-04-30T18:02:14Z

My hunch is that this is somehow related to the fact that we have a docs/aspen directory.

chadwhitacre · 2014-04-30T18:02:46Z

(Pdb) l
  1     from __future__ import absolute_import
  2     import pdb; pdb.set_trace()
  3  -> from aspen import renderers
  4     from tornado.template import Loader, Template
  5     import os
  6     
  7     
  8     class Renderer(renderers.Renderer):
  9     
 10         def compile(self, filepath, raw):
 11             loader = self.meta
(Pdb) import aspen
(Pdb) from aspen import renderers
*** ImportError: cannot import name renderers
(Pdb) aspen.__file__
'aspen/__init__.py'
(Pdb) import os
(Pdb) os.getcwd()
'/Users/whit537/personal/gittip/aspen-python/doc'
(Pdb)

That's inside of the aspen_tornado_renderer.py at the bottom of the traceback.

chadwhitacre · 2014-04-30T18:03:21Z

Why is getcwd() different there than above?

pjz · 2014-04-30T18:06:57Z

Because https://github.com/gittip/aspen-python/blob/master/aspen/configuration/__init__.py#L259:

os.chdir(self.www_root)

chadwhitacre · 2014-04-30T18:40:32Z

Yeah, but the question is why the chdir call is interacting badly with the python -m machinery.

chadwhitacre · 2014-04-30T18:41:44Z

If I cd into env (could be any other directory besides the repo root) and then do python -m aspen -a:8537 -w../doc it works fine. That indicates that the presence of the aspen directory in the repo root is implicated in the confusion.

chadwhitacre · 2014-04-30T18:49:23Z

We decided to take stop calling chdir at all. That fixes the behavior we're seeing here and fits with the theme of not being an app server anymore. If someone wants their simplates to have a certain cwd then they should call chdir themselves. IRC

Calling chdir() was causing an issue serving the docs, and had the potential to cause many more. Let's just Not Do That.

We stopped calling chdir automatically as part of our inversion effort (we're a library, not a framework).

No two ways about it: if we're going to be "just a WSGI library" then we can't implement Socket.IO ourselves. Instead we'll have to figure out how to use gevent-socketio or similar. That rebuilding step is on the other side of the Great Purge of 2014.

This rips out network_address and network_engine and the rest of sockets.W

In docs and entrypoints.

chadwhitacre · 2014-05-06T19:35:23Z

IRC

Conflicts: aspen/__main__.py tests/test_sockets_transport.py

No serverpy

We've only had one ever since we removed the server in #324.

No serverpy

pjz added 2 commits April 21, 2014 14:38

Nuke server.py and the algorithm that it ran

15ea3cc

Also fix up the testing client to instantiate Website instead of Server

Nuke the execution module

10856dd

Stop using aspen command in build system

05bdcec

pjz and others added 2 commits April 30, 2014 14:51

Don't chdir() anymore

e8ca3cc

Calling chdir() was causing an issue serving the docs, and had the potential to cause many more. Let's just Not Do That.

Manually chdir for docs

92ff5e4

We stopped calling chdir automatically as part of our inversion effort (we're a library, not a framework).

This was referenced May 6, 2014

Update Cheroot dependency to new version #312

Closed

Make Cheroot dep not complain on 'pip install' #338

Closed

chadwhitacre added 3 commits May 6, 2014 14:37

Rip out Socket.IO implementation

787d234

No two ways about it: if we're going to be "just a WSGI library" then we can't implement Socket.IO ourselves. Instead we'll have to figure out how to use gevent-socketio or similar. That rebuilding step is on the other side of the Great Purge of 2014.

Watch the world burn.

67bc27e

This rips out network_address and network_engine and the rest of sockets.W

Remove some more references to aspen executable

2af5ebc

In docs and entrypoints.

chadwhitacre added 4 commits May 6, 2014 15:24

Fix quickstart in docs to use python -m aspen

87b0bec

Remove NETWORK_ENGINES from __init__

0b6367a

Remove dead link to old aspen doc

c121e4b

Clean up Quick Start a little more

c62acbe

Merge branch 'master' into no_serverpy

52dc668

Conflicts: aspen/__main__.py tests/test_sockets_transport.py

pjz added a commit that referenced this pull request May 6, 2014

Merge pull request #324 from gittip/no_serverpy

45551f3

No serverpy

pjz merged commit 45551f3 into master May 6, 2014

pjz deleted the no_serverpy branch May 6, 2014 19:48

chadwhitacre mentioned this pull request May 6, 2014

switch to gunicorn or uwsgi gratipay/gratipay.com#1868

Closed

This was referenced May 6, 2014

drop socket simplates #298

Closed

(Re-)Implement websocket/socket.io simplates #200

Open

document socket simplates #201

Closed

chadwhitacre mentioned this pull request Jun 18, 2015

python -m aspen doesn't work correctly on windows #440

Closed

chadwhitacre mentioned this pull request Nov 10, 2015

join 'em #526

Closed

chadwhitacre added a commit that referenced this pull request Nov 10, 2015

Flatten algorithms.website to algorithm

8a49093

We've only had one ever since we removed the server in #324.

Changaco pushed a commit that referenced this pull request Mar 10, 2016

Flatten algorithms.website to algorithm

917df06

We've only had one ever since we removed the server in #324.

Changaco pushed a commit that referenced this pull request Mar 11, 2016

Merge pull request #324 from gittip/no_serverpy

a20f3e6

No serverpy

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

No serverpy #324

No serverpy #324

pjz commented Apr 4, 2014

chadwhitacre commented Apr 22, 2014

chadwhitacre commented Apr 22, 2014

pjz commented Apr 22, 2014

chadwhitacre commented Apr 24, 2014

pjz commented Apr 25, 2014

chadwhitacre commented Apr 25, 2014

pjz commented Apr 25, 2014

chadwhitacre commented Apr 30, 2014

chadwhitacre commented Apr 30, 2014

chadwhitacre commented Apr 30, 2014

chadwhitacre commented Apr 30, 2014

chadwhitacre commented Apr 30, 2014

chadwhitacre commented Apr 30, 2014

chadwhitacre commented Apr 30, 2014

chadwhitacre commented Apr 30, 2014

chadwhitacre commented Apr 30, 2014

pjz commented Apr 30, 2014

chadwhitacre commented Apr 30, 2014

chadwhitacre commented Apr 30, 2014

chadwhitacre commented Apr 30, 2014

chadwhitacre commented May 6, 2014

No serverpy #324

No serverpy #324

Conversation

pjz commented Apr 4, 2014

chadwhitacre commented Apr 22, 2014

chadwhitacre commented Apr 22, 2014

pjz commented Apr 22, 2014

chadwhitacre commented Apr 24, 2014

pjz commented Apr 25, 2014

chadwhitacre commented Apr 25, 2014

pjz commented Apr 25, 2014

chadwhitacre commented Apr 30, 2014

chadwhitacre commented Apr 30, 2014

chadwhitacre commented Apr 30, 2014

chadwhitacre commented Apr 30, 2014

chadwhitacre commented Apr 30, 2014

chadwhitacre commented Apr 30, 2014

chadwhitacre commented Apr 30, 2014

chadwhitacre commented Apr 30, 2014

chadwhitacre commented Apr 30, 2014

pjz commented Apr 30, 2014

chadwhitacre commented Apr 30, 2014

chadwhitacre commented Apr 30, 2014

chadwhitacre commented Apr 30, 2014

chadwhitacre commented May 6, 2014