William Noon wrote:
Last week I was setting up an Apache VirtualHost to point to a Zope
object tree using the ProxyPass directive and SiteAccess to get all the
urls right (great stuff!).
However, I noticed that the REMOTE_ADDR was always set to the
proxy host. This makes sense because
On Mon, 12 Feb 2001, William Noon wrote:
A simple fix to the proxy module (in proxy_http.c) will create a new
header called 'Original-Addr' to contain the client's ip address.
Here is a context diff against Apache 1.3.14 that should work even
in the new 2.0 servers:
FYI, there is an
Dieter Maurer wrote:
"manage_" methods are mapped to permissions (as are any other methods).
You can decide which roles have which permissions.
What does the 'View management screens' permission protect?
cheers,
Chris
___
Zope-Dev maillist
Steve Alexander wrote:
Also, take a look at mod_proxy_add_forward for Apache, and these other resources:
http://www.zope.org/Members/stephen/ApacheProxyPatch
http://modules.apache.org/search?id=124
ftp://ftp.netcetera.dk/pub/apache/mod_proxy_add_forward.c
I believe there is
Right,
I really like the idea of these things but I am concerned about something that
allows anonymous users to futz with traversal.
Can someone put my fears to rest that using these won't let anonymous users do
bad things to my sites?
cheers,
Chris
On Tue, 13 Feb 2001, Chris Withers wrote:
Right,
I really like the idea of these things but I am concerned about something that
allows anonymous users to futz with traversal.
Can someone put my fears to rest that using these won't let anonymous users do
bad things to my sites?
I use
On Tue, 13 Feb 2001 10:24:54 +, Chris Withers [EMAIL PROTECTED]
wrote:
I really like the idea of these things but I am concerned about something that
allows anonymous users to futz with traversal.
Can someone put my fears to rest that using these won't let anonymous users do
bad things to
On Tue, 13 Feb 2001 10:30:26 + (GMT), Matt Hamilton
[EMAIL PROTECTED] wrote:
I use them in conjunction with Apache's mod_proxy to rewrite
http://www.simpledomain.com to the long
http://zopehost.foo.com/blah/blah/VirtualHostMonstser/blah/blah. The Zope
host is behind a firewall, so
On 12 Feb 2001, 23:04 Chris Maresca wrote:
If I recall correctly, there is a tool for doing this call srvany,
part of the resource kit.
I use srvany for running some background services written in
Python on some of my servers. It works so far, but not as reliable
as I'd like - especially
Then again, there's the advantage of having something
included as a standard part of Zope.
Yes, thats true. I would like to see this being rolled into the standard
zope (and there is a Collector entry saying that), although I think its
unlikely given the 'competition' from VHM.
Having said
From: "Chris Withers" [EMAIL PROTECTED]
Well, it's easy enough to find out if a site is running Zope, then this
becomes
pretty easy attack to think of
I'm not going to claim that this is perfectly harmless, but I can't think of
any way in which this could be termed an "attack". You can
Hi,
I was to call a Python Script from within a Python product and I'm wondering
what to call it with so that context, container, script, namespace and subpath
all get bound to the right things?
cheers,
Chris
___
Zope-Dev maillist - [EMAIL
Morten W. Petersen wrote:
Hi guys,
indeed a strange problem, anyone experienced this?
Zope Error
Zope has encountered an error while publishing this resource.
AttributeError
Sorry, a Zope error occurred.
Traceback (innermost last):
From: "Chris Withers" [EMAIL PROTECTED]
We're actually phasing this hack out in favour of a Virtual Host Monster
which
seems like a much cleaner solution...
Sorry, Chris, VHM is irrelevent to this problem. If you want to know the
original remote IP, you have two choices:
1. Use one of
[Morten W. Petersen]
| indeed a strange problem, anyone experienced this?
What did you do to make this happen? Versions?
___
Zope-Dev maillist - [EMAIL PROTECTED]
http://lists.zope.org/mailman/listinfo/zope-dev
** No cross posts or HTML encoding!
[Morten W. Petersen]
| The ideal solution would be to use an object that lives in the ZODB,
| I wonder if there is a way to keep the 'object history' empty? That
| is, keeping the counter 'packed' while retaining 'object history'
| information on all other objects.
I'm no ZODB expert, but I
From: Oliver Bleutgen [EMAIL PROTECTED]
Btw. with a small module it's even possible to log the usernames with
apache,
something medusa isn't capabable of afaik (as of zope version 2.3).
Cool! What module is this, and how do you use it?
Cheers,
Evan @ digicool 4-am
Evan Simpson wrote:
You shouldn't have to do anything special; You just call it.
How would I pass a DTML namespace that I might have lying around when I make the
call so that it gets bound to the appropriate Namespace variable in the script?
cheers,
Chris
Chris Withers wrote:
Shane Hathaway wrote:
But here's the biggest news: the Refresh product, which makes life just
a little bit easier for filesystem product developers, now has an
auto-refresh feature and the ability to set up dependent products. This
means changes to your Python
| The ideal solution would be to use an object that lives in the ZODB,
| I wonder if there is a way to keep the 'object history' empty? That
| is, keeping the counter 'packed' while retaining 'object history'
| information on all other objects.
That would work, however I just dont think
This doesn't use svrany.exe, but pythonservice.exe, which is part
of the win32 extensions written by Mark Hammond, now with
ActiveState. "Python programming on Win32" by Mark Hammond
and Andy Robinson contains some documentation on how to use
pythonservice.exe. The Windows version of Zope
Shane Hathaway wrote:
Any idea why that is?
What product are you refreshing? Maybe it's specific to the product.
An NDA'd one ;-)
Sorry, that's not a lot of help I know, but what sort of things should I be
looking for that'd cause refresh to not work properly?
cheers,
Chris
Chris Withers wrote:
Shane Hathaway wrote:
Any idea why that is?
What product are you refreshing? Maybe it's specific to the product.
An NDA'd one ;-)
Sorry, that's not a lot of help I know, but what sort of things should I be
looking for that'd cause refresh to not work
Shane Hathaway wrote:
apparently not. However, I have an idea that might solve it. Thanks
for the heads-up!
Pleasure :-)
(as a workaround we've developed the 'Pound on the Refresh button like a monkey'
technique *grinz*)
cheers,
Chris
___
I've just been reading through some academic papers by Tom Mens.
The paper at this link looks interesting, as it appears to me to
describe a problem that is addressed in Zope by acquisition wrappers.
http://prog.vub.ac.be/Research/ResearchPublicationsDetail2.asp?paperID=51
Abstract
Most
Would anyone volunteer to write a mounting HOWTO? This is a FAQ.
The ExternalMount product is kind of self instructing,
but a general Mounting HOWTO would be a good thing.
FileStorage is what the standard Data.fs is, and it supports undo.
Just one thing, ExternalMount product uses
A simpler solution is to use a mounted storage that doesn't do undo, and
stores changes in place.
You are the second person to answer with that. Would anyone care to give a
quick how-to on how to mount a non-undoing storage, such as FileStorage
since this doesnt seem immediately obvious to me.
Steve Alexander wrote:
Shane's ExternalMount product is here:
http://www.zope.org/Members/hathawsh/ExternalMount
However, I couldn't find it from searching on zope.org, and it isn't
listed on Shane's zope.org page.
Oops, I forgot to get it cataloged.
Shane
Chris Withers wrote:
Shane Hathaway wrote:
apparently not. However, I have an idea that might solve it. Thanks
for the heads-up!
Pleasure :-)
(as a workaround we've developed the 'Pound on the Refresh button like a monkey'
technique *grinz*)
I've added logic to sort the modules
"Andy McKay" [EMAIL PROTECTED] wrote:
| The ideal solution would be to use an object that lives in the ZODB,
| I wonder if there is a way to keep the 'object history' empty? That
| is, keeping the counter 'packed' while retaining 'object history'
| information on all other objects.
On 12 Feb 2001, Erik Enge wrote:
The rationale behind this is that the community at large would benefit
from this by having _real_life_ case studies so when their time has
come to implement an application in Zope, they don't fall into the
same traps and pitfalls we did. Instead of
Tres Seaver wrote:
...it is quite feasible to have one or more storages
inside your Zope ... often at a non-trivial time cost.
Where does the time cost come in? I've noticed a delay on first mounting
another storage, but I haven't looked for a delay on accessing objects
in the storage.
--
32 matches
Mail list logo
