cherrypy-users Google Group

Re: caching and debugging

2013-05-21T14:04:48Z

Woops, dictionary config with valid dictionary :)
{'/': {'cherrypy.tools.caching.debug ' : True}}

Re: caching and debugging

2013-05-21T14:03:20Z

You can configure tool settings in a variety of ways:
Method decorators:
@cherrypy.tools.caching(debug= True)
Dictionary config:
{'/': 'cherrypy.tools.caching.debug' : True}
File Based:
[/]
tools.caching.debug = True
For more an more in depth overview of tools:
[link]

caching and debugging

2013-05-20T12:53:20Z

Hi
I spent a while trying to debug why my default cherrypy MemoryCache was not
caching and serving cached content.
I found out that the default max_obj size was too small for my case to
cache content, and also when i kept reloading a page to test the cache, the
request contained the header 'cache-control: max-age=0'.

Re: [cherrypy-users] MemoryError on cherrypy application hosted on pythonanywhere.com

2013-05-20T09:23:56Z

Hi Giorgos,

how much data are you storing in the sessions?, how much sessions files
does it have in the temporary directory and what is the timeout value
of the sessions?

Cheers.

On Sat, 18 May 2013 04:10:18 -0700 (PDT)

--
Rivera²

MemoryError on cherrypy application hosted on pythonanywhere.com

2013-05-18T11:10:18Z

I have a cherrypy application hosted on pythonanywhere.com. The application
receives very low traffic (very rarely do I see more than 2 concurrent
users), but still I see MemoryErrors very frequently, for example:

ENGINE Error in background task thread function <bound method
FileSession.clean_up of <cherrypy.lib.sessions.FileSes sion object at

Re: [cherrypy-users] do_logout gives 405 error

2013-05-10T18:58:30Z

That was it; I created a form using the post method, and that worked. Kind
of an ugly hack, unfortunately. Is there a way to override this behavior,
or if I expose a function how would I call do_logout()?

Re: [cherrypy-users] do_logout gives 405 error

2013-05-10T18:21:59Z

Yep, it looks like the do_logout request needs to be a POST. It's
not-idempotent, so GET isn't appropriate.

--
Joseph Tate

Re: [cherrypy-users] do_logout gives 405 error

2013-05-10T18:05:47Z

Thanks for the response.
Sorry, I don't understand what you're asking; I just enabled session_auth
in the config and (previously) it "just worked", just like do_login(). My
do_login() request uses POST, maybe I should use that for do_logout()?

Re: [cherrypy-users] do_logout gives 405 error

2013-05-10T17:49:24Z

What's the signature of your do_logout handler? Does it match the request?
Especially the HTTP Method? There are only a couple of reasons a 405
happen: If you are using method dispatcher, and ask for GET, and no method
is provided, the second reason is also related to HTTP Method not being
allowed.

do_logout gives 405 error

2013-05-10T15:57:50Z

I have an older cherrypy application that had been running on 3.1. I am
testing it against 3.2 (Debian upgrade, basically). It uses session_auth
to authenticate users, and login still works just fine.
Logout is broken, though. My code gives the user a "logout" hyperlink
which basically points to "do_logout?from_page=.". In 3.2 I get a 405

RE: [cherrypy-users] Debugging CherryPy applications with Werkzeug

2013-05-10T14:47:39Z

Yes, wsgi.pipeline is exactly that: a way to wrap the application for you.

Robert Brewer

fuman...@aminus.org

Sent: Thursday, May 09, 2013 10:08 AM
To: cherrypy-users@googlegroups.co m

Oi Joel,
nice post. I have the following question, Why not favor wrapping the application inside a middleware. The advantages in doing so is that it is a framework agnostic way of composing functionality. What are the advantages of doing it through Cherrypy's wsgi.pipeline? Or is it just a shorthand cherrypy provides to wrap the the application for you?

Re: [cherrypy-users] Debugging CherryPy applications with Werkzeug

2013-05-09T17:08:28Z

Oi Joel,
nice post. I have the following question, Why not favor wrapping the
application inside a middleware. The advantages in doing so is that it is a
framework agnostic way of composing functionality. What are the advantages
of doing it through Cherrypy's wsgi.pipeline? Or is it just a shorthand

Debugging CherryPy applications with Werkzeug

2013-05-09T05:46:58Z

Hi everyone

I wrote a post [1] about that topic on my blog, just to change the
usual request for help, hopefully some of you may find it useful.

Cheers.

[1]
[link]
--
Rivera²

Re: Success stories request

2013-05-07T21:25:38Z

I use cherrypy is as webserver in bots open source edi translator
([link]).
Cherrypy is a really great product, very stable and trustworthy.
kind regards,
henk-jan ebbers
hjebb...@gmail.com

RE: [cherrypy-users] How do you serve a .msi file from cherrypy using staticfile tool

2013-05-07T15:46:29Z

Thanks for the reply.

I had tried multiple mime types...including the one suggested below.
But I still cannot get the issue resolved

I'm serving the URL request from Flash NavigateToURL(......,'_blank') ;

On My system and one of my colleagues using Chrome and IE(8) both render the msi as text in the browser

Re: [cherrypy-users] How to create "Fake File object" for Unit Testing?

2013-05-07T08:55:57Z

Hi Tyrone,

I had written a small recipe on the subject:
[link]

following this SO question:
[link]

Re: [cherrypy-users] How to create "Fake File object" for Unit Testing?

2013-05-06T19:07:00Z

Hello,

I'm not sure of what you want to do, but you can create a file-like
object (with read and readlines method) using StringIO objects.
This class is defined in the StringIO module for Python 2.7 and io
module for Python 3.X. Here's an example using Python 3:

'that works'

Hope that helps,

How to create "Fake File object" for Unit Testing?

2013-05-06T18:06:22Z

Hi there,
I'm trying to write a unit test to verify that I've uploaded a file
correctly via CherryPy. According to the CherryPy docs<[link]>,
I access the uploaded attachment like this:
attachment.file.read()
Does anyone have any familiarity writing unit tests for file objects?

Re: [cherrypy-users] Two WSGI servers with access to the same application instance?

2013-05-04T21:37:24Z

Thanks Derek, the first part is what I currently have, second part is more
what I was looking for.
My understanding is I still need to call cherrypy.engine.start() after
creating my app object?
Also curious, what's the difference between:
app = cherrypy.tree.mount(MyApp(), "", 'config.ini')
app = cherrypy.Application(MyApp(), "", 'config.ini')

RE: [cherrypy-users] How do you serve a .msi file from cherrypy using staticfile tool

2013-05-03T22:45:44Z

Hi Andrew,

Why are you using a very generic mime type of octet-stream instead of
something like "application/x-ole-storage" or "text/mspg-legacyinfo",
according to [1] that are the appropriates mime types.

'/support/Uplink.msi':
{'tools.staticfile.on': True,
'tools.staticfile.filename': '/apps/Uplink.msi',

RE: [cherrypy-users] How do you serve a .msi file from cherrypy using staticfile tool

2013-05-03T22:39:06Z

Can anybody help with this ? I would prefer to serve the msi as a static
link rather than as a zip file

[mailto:cherrypy-users@googleg roups.com] On Behalf Of Cooper, Andrew
Sent: 30 April 2013 16:14
To: cherrypy-users@googlegroups.co m
using staticfile tool

Given the following static file configurations

RE: [cherrypy-users] Custom logging handlers and DropPrivileges

2013-05-03T15:05:38Z

Hi Tim,

The recommended solution here is to wrap up your handler calls in a plugin and then subscribe it to run after the drop.

class Logging(plugins.SimplePlugin):

def start(self):
cherrypy.log.access_log.handle rs = []
rfh = logging.handlers.RotatingFileH andler(
my_log_path, "a", my_log_size, my_log_count)

Custom logging handlers and DropPrivileges

2013-05-02T20:38:30Z

Gang:
I've got a CherryPy server where, early in the startup process, I customize
the built-in logging handlers to support log rotation:
cherrypy.log.access_log.handle rs = []

cherrypy.log.access_log.addHan dler(logging.handlers.Rotating FileHandler(my_log_path,
"a", my_log_size, my_log_count))
... this has worked great for years. However, I recently also had to start

Re: [cherrypy-users] Two WSGI servers with access to the same application instance?

2013-05-02T15:43:12Z

Through cherrypy:
cherrypy.tree.mount(Root(), config=conf)
cherrypy.server.thread_pool = 30
cherrypy.server.ssl_private_ke y = 'something.key'
cherrypy.server.ssl_certificat e = 'something.crt'
cherrypy.server.socket_port = 443
cherrypy.server.socket_host = '0.0.0.0'
http_server = cherrypy._cpserver.Server()

Re: [cherrypy-users] Two WSGI servers with access to the same application instance?

2013-05-01T19:58:21Z

I do currently have my app running that way, was more just curious from an
experimentation point of view. On a similar note, was curious if I'd be
able to run a single instance of a cherrypy wsgi app behind something like
nginx with both http and https.

Re: [cherrypy-users] Two WSGI servers with access to the same application instance?

2013-04-30T18:54:21Z

Hi,

In my opinion you should use the CherryPy engine layer if you could. It'd
make your life easier. If not, have you look at this?

[link]

--
- Sylvain

Two WSGI servers with access to the same application instance?

2013-04-30T15:17:29Z

Hello all,
I'm looking to run two WSGI servers (one HTTP, one HTTPS) with access to
the same instance of my application. As far as I can tell, the only way to
do this is to split off two threads in my server file:
from myapp import app
from cherrypy import wsgiserver
from cherrypy.wsgiserver.ssl_builti n import BuiltinSSLAdapter

How do you serve a .msi file from cherrypy using staticfile tool

2013-04-30T15:14:18Z

Given the following static file configurations

Both the zip and exe files work but the msi file loads the file as text.

How can I force it to download ?

'/support/Uplink.msi':

{'tools.staticfile.on': True,

'tools.staticfile.filename': '/apps/Uplink.msi',

'tools.staticfile.content_type s': {'msi':

Re: [cherrypy-users] MIrror on Github?

2013-04-24T21:44:22Z

yesyesyes +1 :D

On Wed, Apr 24, 2013 at 5:43 PM, Lakin Wecker <

Re: [cherrypy-users] MIrror on Github?

2013-04-24T21:43:52Z

I am generally in favour of more exposure. MIchiel Overtoom's question
would be my only concern. It would be a shame if we generated pull requests
that were ignored because it was a pain to get them pulled all the way into
the mercurial repository.

Lakin

Re: [cherrypy-users] MIrror on Github?

2013-04-24T19:26:37Z

I don't mind either way but if people are interested, it might be a good
idea indeed :)

--
- Sylvain
[link]
[link]

Re: [cherrypy-users] MIrror on Github?

2013-04-24T17:04:45Z

Count me among those who'd love to see this. I think exposure wise it'd
really help.

Re: [cherrypy-users] CherryPy Sessions seems wrong

2013-04-22T17:21:06Z

Mark wrote:

Because they take up disk space. Now, we're talking about balancing an
infinitely small CPU load against an infinitely small amount of disk
space, so there is no reason to choose a particular time interval other
than whatever makes you comfortable.

If the session data contains sensitive information, then there may be a

Re: [cherrypy-users] CherryPy Sessions seems wrong

2013-04-21T19:34:54Z

Thanks,

I'm using SimpleDB as a session backend so I can keep track of state across
multiple machines. I just had to modify setting of the client cookie and
the clean up timeout to accomodate my use.

Mark

Re: [cherrypy-users] MIrror on Github?

2013-04-21T05:04:55Z

Is it two-way? What happens if someone forks CherryPy on github, modifies something and creates a pull request? How will that show up on Bitbucket? And what about the issue tracker? Are the issues shared too between Bitbucket and github?

--
"If you don't know, the thing to do is not to get scared, but to learn." - Ayn Rand

Re: [cherrypy-users] MIrror on Github?

2013-04-21T04:58:01Z

For what it's worth, I also think this would be awesome. Github has a
really nice interface for browsing code, and git itself is more ubiquitous
than mercurial.

Re: [cherrypy-users] MIrror on Github?

2013-04-20T23:28:42Z

I'm seriously considering to implement a mirror on github, It might give
more visibility to the project and it seems kinda easy with hg-git<[link]>
.
I'm not saying that we should move to github because is more popular, but
to have just a little bridge for the githubers, a lot of people look for

Re: [cherrypy-users] CherryPy Sessions seems wrong

2013-04-20T23:15:25Z

Hi Mark

You can change the clean up frequency with
'tools.sessions.clean_freq': 60 * 24 # a day
'tools.sessions.clean_freq': 60 * 24 * 7 # a week

Also this [1] previous thread might illustrate a little more how the
sessions are expected to be used.

Cheers.

[1]
[link]

CherryPy Sessions seems wrong

2013-04-20T22:46:32Z

Sessions in cherrypy seem wrong to me.
1) Why is timeout set on the client and server? Client cookies should
default to session only so public computers don't retain the cookie after
the user closes the browser. Server just ignores expired cookies.
2) Why are we calling clean_up every 5 minutes? Can I change this? Do it

Re: CherryPy on PythonAnywhere

2013-04-18T09:05:23Z

Hi,
First of all install a Cherrypy.
then go through thae directory:CherryPy-3.2.2\cherr ypy\tutorial and try
that examples.
then try to implement on your own.
I am working on CherryPy and for web layouts HTML and my application works
good.
so, Need any help then plz post it..
Regards,

Re: [cherrypy-users] CherryPy wsgiserver - threadpool automatic grow/shrink

2013-04-10T18:01:44Z

My understanding is that the thread pool is limited. The thread pool use a model similar to pre-fork in Apache where the threads are created at start up and reused.

--
Eric Larson

Printing error stream but not access stream to stdout

2013-04-07T03:53:10Z

Hello,

I would like to make my cherrpy server log the 'error' messages to
stdout but not log the 'access' messages at all. If I set my cherrypy
config like this:

'log.access_file' : None,
'log.screen' : True

this appears to log both error *and* access messages to stdout, which

CherryPy wsgiserver - threadpool automatic grow/shrink

2013-04-09T13:56:46Z

I am using CherryPy wsgiserver (python 2.7.x) with one of my applications.
I configured 3 threads to begin with and default max threads, i.e., no
limit. When driven with concurrent requests, I donot see the thread pool
increased beyond the initial 3 threads. On further inspection into the
code, I see that ThreadPool.shrink() and grow() methods are not called

Re: How to start with varible adding on url

2013-04-06T22:58:21Z

Tim!!!!
look on this question
[link]
Thx for all

Re: Server's starting and add the querystring with redirect

2013-04-06T22:57:11Z

For all the solution is on
[link]

Re: Server's starting and add the querystring with redirect

2013-04-06T22:21:36Z

Hi ravij

Now I have new error:

import cherrypy
import urllib

class Root(object):
@cherrypy.expose
def index(self):
jsondict = [('foo', '1'), ('fo', '2')]
p = urllib.urlencode(jsondict)
raise cherrypy.HTTPRedirect("/index? " + p)

cherrypy.config.update({

Re: Server's starting and add the querystring with redirect

2013-04-06T22:12:15Z

Hi Mirko,
Please reconsider your "solution". What you are trying to do just seems
wrong. Lets walk through your code.
------------------------------ ------------------------------ ------------------------------ -----------------------
import cherrypy
import urllib
class Root(object):
@cherrypy.expose

Server's starting and add the querystring with redirect

2013-04-06T21:54:53Z

his is a server in Cherrypy that I programmed and I want add the query
string.
But when I use the redirect the site and it appears that does not work with
the localhost:8080/index?foo=1&foo =2
Why?
Where i wrong?
My project
import cherrypyimport urllib#import requests
class Root(object):
@cherrypy.expose

Re: [cherrypy-users] cherrypy.response.body is allways transformed to list

2013-04-03T18:34:26Z

Hello Eugene,

thank you for your help.

In an other app with the same genshi version, (Genshi-0.6) and a
different cherrypy version
it works as expected.

I tried also the older cherrypy version (3.2.2) and got the same
result.

Am Mittwoch, den 03.04.2013, 08:51 +1100 schrieb Eugene Van den Bulke:

Re: [cherrypy-users] cherrypy.response.body is allways transformed to list

2013-04-03T17:40:13Z

Hello Vincent,

I'm using Genshi in the same way (as tool, and plugin)
as it is described here:
[link]

I will follow the hint, that it can be caused by the templating engine.

Am Dienstag, den 02.04.2013, 22:15 +0200 schrieb Vincent Le Goff: