Reporting DOM Bugs? Be Sure to Read This!

Monday November 8th, 1999

Eric Krock has a request for anyone report DOM bugs:

"Please bookmark the below URLs!!!

If you hit a page that doesn't display well or has JavaScript errors in Nav5 (Mozilla) because the page uses MS IE4/5 DOM features, Nav4 Layer DOM features, or the LAYER/ILAYER tag, DON'T FILE A BUG in Bugzilla. Instead, use the below email creation templates to send an email to the page's owner asking them to upgrade the page to support W3C standards.

If the page supports IE4/5 but not HTML 4.0/W3C DOM:

If the page supports Nav4 but not HTML 4.0/W3C DOM:

If the page is breaking because of user agent detection problems:

(Use that one when a page doesn't work or has a JavaScript error because the client-side JavaScript isn't detecting the Navigator 5 Mozilla/5.0 userAgent string, or discover that a server-side CGI is returning the wrong page because it's not detecting the Mozilla/5.0 HTTP USER_AGENT string or is choking on its HTTP 1.1 CONTENT_TYPE string.)

Likewise, if you're examining a bug report in Bugzilla and realize it's caused by these issues, mark the bug INVALID. Then use the email creation templates to notify the bug reporter and the page owner of the need to upgrade. Bugs closed as INVALID in this way count half a point in the BugAThon!

Hint to avoid filing bogus bugs: if content on a page doesn't position correctly or there's a JavaScript error, do a View Source and look (or copy and paste and search in a text editor) for the strings document.all and layer (case insensitive). If you find those strings in the JavaScript or the HTML markup, think twice before filing a bug. Create a simplified test case without the proprietary features and see if the problems still occurs.

#35 Ugh

by Ben_Goodger

Tuesday November 9th, 1999 7:15 PM

You are replying to this message

Yes, right through 1998 I was designing dynamic elements for websites of mine initially for N4.x, and later for both N4.x and IE, so I understand as well as the next person the pain involved in doing something as fundamental as an expanding menu ;)

You raise good points, and I'm sure Mr Steinman will update his script with W3C DOM support (if he has not done so already). I think there should be some sort of stamp that a browser has that identifies it as fully <x standard> compliant so scripts can check ;)

What someone might like to do is develop emulation functions, e.g. create a function that searches through a document for an element with a specific ID (scanning through each level of layer in NN4.x, for instance, and each element in the all collection in IE4) and return the object. What you'd then do is attach this function to document with prototype:

document.prototype.getElementById = myGetElement;

You could then bundle this into a JS file. The function would be run on non-Mozilla/IE5 browsers (using the Devedge browser sniffer, say) which would emulate the appropriate function. This has been done the other way too, someone wrote a function a while back for Mozilla that imitated IE's innerHTML attribute.

Neither am I against the implementation of native code to do this, my point to the "doom and gloom" posters is that time does not permit the implementation of workarounds that may cause problems with the code most web developers want. I'm sure would welcome an implementation of this if someone provided the code.