From: Zoran V. <zv...@ar...> - 2007-06-26 15:38:43
|
Hi! The IE7 is not able to talk to our server. Perhaps anybody has some idea why? I updated to last CVS status today (haven't been updating for about a month or so) and saw LOTS of changes... If nobody has any idea, I will have to pull up the sleeves and fix that asap. Cheers Zoran |
From: Stephen D. <sd...@gm...> - 2007-06-26 16:00:03
|
On 6/26/07, Zoran Vasiljevic <zv...@ar...> wrote: > Hi! > > The IE7 is not able to talk to our server. Error message? Log file? Does it fail on: ns_return 200 text/plain ok |
From: Zoran V. <zv...@ar...> - 2007-06-26 16:14:57
|
On 26.06.2007, at 17:59, Stephen Deasey wrote: > Error message? Log file? > > Does it fail on: > > ns_return 200 text/plain ok No errors, no access.log entries, nada. I created no-brain index.html like this: zvpb:/usr/local/ns zoran$ cat pages/index.html <html> <body> Hallo </body> </html> and this never gets served. The IE attempts the connection and stops, telling that it could not open the page. Other browsers get the page allright (including IE6). Weird. I haven't tried it on Linux. I'm just trying the Mac OS X. Any clues where I should start peeking? To me, it looks like a driver-thread issue but I'm not sure. Cheers Zoran |
From: Stephen D. <sd...@gm...> - 2007-06-26 16:34:54
|
On 6/26/07, Zoran Vasiljevic <zv...@ar...> wrote: > > On 26.06.2007, at 17:59, Stephen Deasey wrote: > > > Error message? Log file? > > > > Does it fail on: > > > > ns_return 200 text/plain ok > > No errors, no access.log entries, nada. > I created no-brain index.html like this: > > zvpb:/usr/local/ns zoran$ cat pages/index.html > <html> > <body> > Hallo > </body> > </html> > > and this never gets served. The IE attempts the > connection and stops, telling that it could not > open the page. > > Other browsers get the page allright (including > IE6). Weird. > > I haven't tried it on Linux. I'm just trying the > Mac OS X. Any clues where I should start peeking? > > To me, it looks like a driver-thread issue but I'm > not sure. I was thinking that maybe the Explorer developer toolbar let you spy on the headers like the Firebug extension for Firefox does, but it doesn't mention anything: http://www.microsoft.com/downloads/details.aspx?FamilyID=E59C3964-672D-4511-BB3E-2D5E1DB91038&displaylang=en You could try ngrep on your OSX box: sudo /usr/sbin/ngrep -W byline -d eth0 port 8000 ( http://ngrep.darwinports.com/ ) Maybe it's already installed? |
From: Zoran V. <zv...@ar...> - 2007-06-26 16:41:47
|
On 26.06.2007, at 18:34, Stephen Deasey wrote: > > I was thinking that maybe the Explorer developer toolbar let you spy > on the headers like the Firebug extension for Firefox does, but it > doesn't mention anything: Oh, the trouble is far more difficult than that... I removed the entire /usr/local/ns and reinstalled all fresh. This works now as expected. Including the IE7. But... our own app will not. And that is only on IE7. Other browsers work fine. The damn thing with IE7 it shows *nothing* as if the inital connection to the server failed. No logs whatsoever... There must be something that radically changed in the server. I did recompile all our private modules but no avail... So... this leaves me with a pleasant electronic archeology that I always loved that much... :-( |
From: Stephen D. <sd...@gm...> - 2007-06-26 17:02:24
|
On 6/26/07, Zoran Vasiljevic <zv...@ar...> wrote: > > On 26.06.2007, at 18:34, Stephen Deasey wrote: > > > > > I was thinking that maybe the Explorer developer toolbar let you spy > > on the headers like the Firebug extension for Firefox does, but it > > doesn't mention anything: > > Oh, the trouble is far more difficult than that... > I removed the entire /usr/local/ns and reinstalled > all fresh. This works now as expected. Including the > IE7. > > But... our own app will not. And that is only on > IE7. Other browsers work fine. The damn thing with > IE7 it shows *nothing* as if the inital connection > to the server failed. No logs whatsoever... > There must be something that radically changed in > the server. I did recompile all our private modules > but no avail... > > So... this leaves me with a pleasant electronic > archeology that I always loved that much... > > :-( FYI: freebsd seems borked. Immediately after the call to Tcl_FSLoadFile() in modload.c as nslog.so is being loaded (which is the first module to be loaded) all log entries are printed twice and module loading fails. Which maybe points to some corruption somewhere. I've run it under valgrind and mudflap. Works fine on RHEL 5, SLES 10, Debian 4, Fedora 7, 32/64 bit, and itanium :-) Fails for me on: td150.testdrive.hp.com, td152.testdrive.hp.com |
From: Zoran V. <zv...@ar...> - 2007-06-26 17:42:56
|
On 26.06.2007, at 19:02, Stephen Deasey wrote: > FYI: freebsd seems borked. Immediately after the call to > Tcl_FSLoadFile() in modload.c as nslog.so is being loaded (which is > the first module to be loaded) all log entries are printed twice and > module loading fails. Which maybe points to some corruption somewhere. > > I've run it under valgrind and mudflap. Works fine on RHEL 5, SLES 10, > Debian 4, Fedora 7, 32/64 bit, and itanium :-) > > Fails for me on: td150.testdrive.hp.com, td152.testdrive.hp.com Great :-( So we need more freebsd testing ( have no installation here ) BTW: the IE7 trouble... I did not compile ALL the modules in our app (I keep forgetting one all the time). It compiled now and it works fine. Sorry for the noise on the list (my fault). Generally we are still "in flux" and the public API is changing so one always need to remember to recompile *everything*. Lately I was very occupied so I did little (none?) work on the server. I wanted to get this init stuff entirely done with interp traces and trash that Tcl introspection stuff for good. Perhaps this will be a good time to "get to know" all the changes done so far. Some months ago I suggested to make a 5.0 release... Or, at least some other like 4.99.3 or such.... will make things simpler for us. What do you all think? Cheers Zoran |
From: Stephen D. <sd...@gm...> - 2007-06-27 16:56:46
|
On 6/26/07, Zoran Vasiljevic <zv...@ar...> wrote: > > Lately I was very occupied so I did little (none?) work > on the server. I wanted to get this init stuff entirely done > with interp traces and trash that Tcl introspection stuff for good. > Perhaps this will be a good time to "get to know" all the > changes done so far. It would be great if we can get some kind of coherent story about interps, packages etc. and how you manage them into ns_ictl.man: http://www.crystalballinc.com/vlad/software/naviserver/files/mann/ns_ictl.html (when Vlad reruns his script) > Some months ago I suggested to make a 5.0 release... Or, at least > some other like 4.99.3 or such.... will make things simpler for > us. What do you all think? It's a nice idea... :-) |
From: Zoran V. <zv...@ar...> - 2007-06-27 19:12:44
|
On 27.06.2007, at 18:56, Stephen Deasey wrote: > On 6/26/07, Zoran Vasiljevic <zv...@ar...> wrote: >> >> Lately I was very occupied so I did little (none?) work >> on the server. I wanted to get this init stuff entirely done >> with interp traces and trash that Tcl introspection stuff for good. >> Perhaps this will be a good time to "get to know" all the >> changes done so far. > > > It would be great if we can get some kind of coherent story about > interps, packages etc. and how you manage them into ns_ictl.man: > > http://www.crystalballinc.com/vlad/software/naviserver/files/mann/ > ns_ictl.html I agree entirely. Let me do that in the next few days. At the same time I will like to ditch that ugly introspection mechanism. This is a long time due. Hopefully I will not step on somebody's toes with that ;-) Cheers Zoran |