- Zope - Zope lists

RE: [Zope] Re: Performance of webserver?
by Thaxter, Jason M. 01 Sep '99

01 Sep '99

That sounds great! Can it be built as --enable-module=pcgi --enable-shared=pcgi? If you work on benchmarking, try increasing the number of threads beyond the default of 4 (e.g. python -t8 -p./Zope.cgi ). If the threads exceeds by 2 or more the number of concurrent requests in your benchmarking app, you'll notice that the performance gap due to pcgi (vs the same test on zserver) narrows rather impressively, but it would be really nice to see it disappear more or less completely! > -----Original Message----- > From: Sam Gendler [mailto:sam.gendler@software.com] > Sent: Tuesday, August 31, 1999 3:36 PM > To: zope(a)zope.org > Subject: [Zope] Re: Performance of webserver? > > > I haven't done any benchmarking, yet, as I only downloaded zope > yesterday. I have, however, generated an Apache module based > on the pCGI > stuff, so that you can have persistent connections to the zope server, > and there is no overhead of exec'ing any new process. Everything runs > inside of Apache space, and every time an apache process exits, the > connection to zope is closed and cleaned up. This gives me > the piece of > mind of running on the Apache webserver, so I know the performance > baseline, and I trust the stability and security of the webserver. I > will make the module public just as soon as I add all of the handling > for real configuration settings, such as virtual hosts, etc. > Currently, > there are just global zope settings across the entire webserver, > regardless of host, directory, .htaccess, etc. > > Once it is done, you should get the baseline performance numbers of > Apache, with whatever added latency the zope executions require. > > --sam > > -- > ---------------------------------------------------- > "Some Scientists claim that hydrogen, because it is > so plentiful, is the basic building block of the > universe. I dispute that. I say there is more > stupidity than hydrogen, and that is the basic > building block of the universe. > --Frank Zappa > > > > > _______________________________________________ > Zope maillist - Zope(a)zope.org > http://www.zope.org/mailman/listinfo/zope > > (To receive general Zope announcements, see: > http://www.zope.org/mailman/listinfo/zope-announce > > For developer-specific issues, zope-dev(a)zope.org - > http://www.zope.org/mailman/listinfo/zope-dev ) >

1 0

Feedback: The namespace howto
by Michel Pelletier 01 Sep '99

01 Sep '99

> -----Original Message----- > From: Rik Hoekstra [mailto:hoekstra@fsw.leidenuniv.nl] > Sent: Tuesday, August 31, 1999 4:07 PM > To: info(a)digicool.com > Subject: Feedback: The namespace howto > > > Once again clearly explained, except for the most zennish of features. > Let me once again ventilate my frustrations: > > - Acquisition works mostly as expected - (though not always with DTML > Methods/Documents, see my other feedback) > - It has always excaped me how acquisition works with folders that are > on a same level? For example: > > ROOT > | index_html > | > ---------------------- > | | > OneFolder SecondFolder > | index_html > Subfolder > > > You can call this in various ways (for instance if > Secondfolder holds a > different sitedesign, and OneFolder holds the data; or with > multilanguage versions of a site): > > URL1 http://mysite/OneFolder/Subfolder/SecondFolder > > URL2 http://mysite/SecondFolder/OneFolder/Subfolder > > The riddle: in both cases which index_html is to be returned - I > struggled with this so much I can't remember which did what and why. I > can tell it was _not_ intuitive at all... The only way I found > In both cases, /SecondFolder/index_html is returned. In the first URL, second folder returns it's index_html directly. At that point, the *acquisition* name space stack is: Secondfolder Subfolder Onefolder ROOT During traversal, Onefolder finds subfolder, which then does not find secondfolder, so it searches back through the path until it finds 'secondfolder', which then returns it's index_html. In the second URL, SubFolder has no 'index_html' method, so it tries to acquire it. it will search the namespace stack from the top down: SubFolder OneFolder SecondFolder ROOT thus finding the first 'index_html' in secondfolder. The reason why you can acquire in this way is because the root folder is *allways* the bottom item on the acquisition name space. By that logic, that means that *all items in the root folder are acquirable by the whole object heirarchy*. > Hm, I suppose this was also a request for enlightenment, not just > feedback - oh well Were you enlightened? -Michel > > Rik Hoekstra >

1 0

Zope WebDAV e Apache
by Carlos Henrique Bauer 01 Sep '99

01 Sep '99

Hi, Is there any way to access Zope WebDAV under Apache? Best regards, -- Carlos Henrique Bauer

1 0

RE: [Zope] Zope Installation pains..
by Rob Page 01 Sep '99

01 Sep '99

> Anyways, onto the meat of the matter.. > > I am trying to install Zope onto a fresh Mandrake (Redhat > 6.0) 6.1 box, > running Apache, the ultimate goal being to run SquishDot. At first I What version of zope are yo trying? Pls provide the apache rewrite rule you are using. --Rob

2 1

ZCatalog questions
by Jason Jones 01 Sep '99

01 Sep '99

Is it possible with ZCatalog to index subsets of page data? For instance, I would like to index at a paragraph level, including anchors so that I could return users to the relevant part of a matching document. If the phrase "this is some phrase" is found three quarters of the way down the page, and I have anchors every paragraph or so, I'd like to send the user to the closest anchor. Can ZCatalog index on anchors, or on in-file meta data that it can build anchors off of? Thanks, Jason jason_j(a)countermedia.org

1 0

How Hotmail blew it
by Jeff Bauer 01 Sep '99

01 Sep '99

Dear Zopistas, While not directly relevant to Zope development, some people may find the following article of value. http://news.excite.com/news/zd/990901/06/how-hotmail-blew It describes how Hotmail was accessed by placing a simple line in a browser's location field. "No computer voodoo required, no hackerish expertise needed." microsoft-should've-used-zope-ly yr's, Jeff Bauer Rubicon Research

1 0

Zope Installation pains..
by Nicolas Pottier 01 Sep '99

01 Sep '99

Hello there.. First off let me say that I have RTFM, and it doesn't seem to be helping.. I've also searched the list and (surprisingly) not found any info on similar problems, so this is my last resource.. Trust me I understand how bad lists can get due to annoying newbie questions, but alas, here I am. Anyways, onto the meat of the matter.. I am trying to install Zope onto a fresh Mandrake (Redhat 6.0) 6.1 box, running Apache, the ultimate goal being to run SquishDot. At first I wanted to use pcgi to serve it up, but after struggling with that for a while unsuccessfuly I decided I would settle with using zerver. Now, I've tried both the rpm distribution of zope as well as the standard tgz distribution with the same effect. After the install, I go to http://localhost:9673/ to find my happy Zope home page waiting for me, so far so good. I now try going to the management screen link, and am challenged for a username and password. For some reason, it does not take the default username and password, so I change the access file to contain nic:test. I now shutdown zerver and try again. It now challenges me, and comes up with the following error: Sorry, an error occurred.<p>  Can someone help me out? I'm willing to blow everything away and start over given instructions that are known to work, but I've never had such trouble installing anything before given instructions, so i suspect something is wrong somewhere. Thanks! -Nic PS Please reply directly, as I am not yet on the list

1 0

Portal Toolkit
by davidbro＠namshub.org 01 Sep '99

01 Sep '99

Perhaps this is a FAQ, but since I've been off of the list for a while, I probably didn't see the answer fly by. But what is the plan for the Portal Toolkit? Is it part of Zoep 2.0? Or is it an add-on product? Will there be a fee for it? dave (coming back into the Zope fold)

1 0

zope hangs after "a while"
by Elliot Lee 01 Sep '99

01 Sep '99

http://news.gnome.org/gnome-news/ - the GNOME News site is up and running, thanks to help from list members. The only remaining obstacle between the site and a good slashdotting is the fact that after a period of operation, the server hangs (the same symptoms as when I overloaded 2.0.0b4 with requests). This period can be as short as 20-30 minutes, and I think averages around an hour. The site has received 20k page hits over the past 23.5 hours (I've got a script that restarts it every hour). I know this is very vague - if I can provide any specifics, please say so. -- Elliot The time for action is past! Now is the time for senseless bickering!

3 2

#zope irc?
by aggarwal 01 Sep '99

01 Sep '99

hello, i remember reading on the zope site that there was a #zope channel set up on efnet... i have been periodically checking there, but no one seems to be logged on there.. i am in no way stepping on any ones toes here, but for simplicity sake, i did put a bot in the channel, to keep it alive (topic set, etc.) i will gladly remove the bot to the proper owner, i just thought it would a cool place to meet fellow zope'rs... -Samir

3 3