ASP Error 0126 -vs- Firefox

[Stoic Joker]

Okay... I'm sure most folks will recognize ASP 0126 as the (server side) include file pathing error. The "catch" here is that not only is the path correct, the page loads flawlessly when using IE (5, 6, or 7).

Firefox on the other hand is tossing out the an ASP 0126 error (apparently) all by itself... Now how the hell a browser is managing to run afoul of a Server Side Script I've not a clue.  But I'm open for suggestions if anybody has an idea on what may be happening.

Thank You,
Stoic Joker

[f0dder]

Obviously firefox itself isn't going to bring up an ASP error, since that's done serverside. And I would be surprised if FF sends a regular HTTP request differently than IE (do check your server logs, though).

Is there any (clientside) scripting involved? Perhaps there's a subtle difference leading to a different page being requested? Or perhaps a cookie is being blocked?

[Stoic Joker]

Hence my quandry...

The visitor that reported the error was using Firefox, Safari, & OSX 10.5. Now Safari borking anything doesn't surprise me, but Firefox is written by folks that actually know what they're doing. And as already stated it's (an include statement in) a Server Side script (nothing client side)...which is supposed to make the scenario impossible. The site at that point uses only the browser session info, and the page just loops back to itself with SQL querys for different product types (e.g. it's really basic).

*Sigh*

I verified the page was working (in case the server glitched) and all was fine (in IE), I stepped outside the building (server is in-house) and tried the page from my home office which was also IE & fine. Our primary web developer (I'm more the server admin/backup coder), who always runs FF on a test machine, went to the site while we were on the phone and said it did indeed fail (ASP 0126) when loaded in FF ... even tho it was loading fine for me in IE.

So... Site broken -> Owner of company staring at me -> standing there with my D**k in my hand.

...Happy Times ...Not!

I know it "Can't" happen, but I've got a credible source verifying that it is happening.

[justice]

Does this help: http://www.webmaster...com/forum47/1620.htm make sure you use forward slashes where appropriate, IE might be more lenient.

[f0dder]

Again, have you checked server logs to see if FF does request differently from IE?

[Stoic Joker]

justice - I already ran across that page, enabling Parent Paths is only required if you're using the include file "../dir/file.ext" syntax. Parent Paths is the (flawed) feature that was used/exploited by Nimda, CodeRed and a few other server targeting broadcast bast viruses. So I never use it on a prodiction box.

f0dder - No I haven't checked the logs yet, but I'll do that when I get to the office this morning. Last night I installed FF on a XP SP3 VM and took it to the site only to find it was working fine ... for me ... which was no help at all as trying to troubleshoot a non-reproduceable error is a great way to break something (e.g. I was a bit PO'ed and decided it was best to just go watch TV).

I'll post what I find in the logs later this morning, now that I'm calmer...

P.S. Thank You - I'd been hoping to get your input when I originally posted.

[Stoic Joker]

New developments; Discovered how to reproduce error!

Okay... working 2 pronged attack, I'm going over the logs and for some reason I can not locate the user in questions session ... even tho it has to exist because he sent the report via the sites web form (the server is set to log everything).

And, The user specified that they were using FF3, I had installed FF3.0.1 on the test machine. Our other Dev guy just informed me that he was using FF2.02 during yesterday's testing (when it failed), but had installed FF3.0.1 and the site is now loading fine for him.

So... Apparently it is only FF3 and under that is failing to load the page.

[f0dder]

Nice that you've been able to reproduce - it's such a bizz when you can't, or it's only possible intermittently. So, I guess the trick is to make a request with IE, another with working-FF, and a third with faulting-FF, and then check the server logs. At least that's what I'd do, but I'm no web developer, mind you

[Stoic Joker]

Nice that you've been able to reproduce - it's such a bizz when you can't, or it's only possible intermittently. So, I guess the trick is to make a request with IE, another with working-FF, and a third with faulting-FF, and then check the server logs. At least that's what I'd do, but I'm no web developer, mind you

-f0dder (July 25, 2008, 05:43 PM)

Me either man, I'm a Network/Server Admin and I prefer to work in C++ on system utilities...I only do the website stuff with a gun to my head.

The part that was killing me was yesterday I was on the phone with our (official) company web guy who was using FF2 while I was using IE7 We were both coming from outside the building (in-house server) him from home & me using RDC, both doing forced refreshes, I loaded fine, and he failed consistently.

An exhaustive review (I'm talking 4 hours) of the server logs showed that 90% of his activities, and 100% of the outage reporters activities did not appear in the logs (even tho I have full logging enabled). I even used the time stamp on the Email he'd sent (to a constantly monitored mailbox) from the web form to match up the records and there was nothing there except for my internal visit (which was logged) 10 min after shit hit the fan.

So how the hell we managed to break half of something I've not a clue. So here's the weird part... Last night I managed to over complicate the diagnosis by noticing that there were Win updates available for the server ... Which of course required a reboot. now here's the now clear in retrospect time-line of the misc relevant events:

I exasperated toss FF3.0.1 in an XP VM and load the site correctly.

Trying to troubleshoot an issue that I know is impossible, I Email my results to the web guy hoping to confirm the results as I've no clue what version of FF he's using.

I go watch TV and wait for response from web guy.

Knowing I'll never be able to sleep I start dinking with the server notice and run the updates (server restarts...).

Web guy gets Email in morning, grabs latest copy of FF to confirm my results, him not knowing and me not thinking about the relevance of the restart (ShiT...) "we" both (separately) conclude that the issue lies with FF2.

Now the fun ensues... because I'm do to start a much needed vacation - so I gota get this MF fixed NOW!!!

I go at the Logs with a vengeance.

The anomalously missing log entries begin to paint a picture of some manner of intermittent systemic error ... This type of occurrence generally makes one thing of restarting something ... I go googling for an old copy of FF, install it on a Win2k VM, and confirm my now very strong suspicion that I'm a Friggin Idiot.

Problem Solved.

Because no matter how close one get to attaining a goal, there is always plenty of time for things to go horribly wrong...

[f0dder]

Hmmm, missing log entries - cached copies? caching proxy inbetween?

[Stoic Joker]

missing log entries - cached copies? caching proxy inbetween

-f0dder (July 25, 2008, 08:22 PM)

LOL - Oh the time I spent on that road ... Problem being The Email sent had to do a series of requests to the server in order to get through (and those were never found). Not to mention that I'm using CDOSYS to drop the (from) site mail directly into the Exchange pickup folder. ...And all the pages in question are set for no cacheing due to their containing dynamic content (Product catagories/pricing/etc.) pulled from a db.

With all that going something should have been logged ... but it wasn't.

Problem is I should have done:

iisreset /stop
iisreset /start

From the get-go (which would have resolved the "issue") ... but (being under-the-gun...) I came out of the gate over-thinking the problem, and (damn near) blew myself outa the water.

*Sigh* ...Guess that's why "they" call it a learning experience...

[Veign]

Can you post the source code causing the problem?