with netscape 4.04 and 4.05, the javascript selection does not work. when you click on a program, the version column is updated, but the component column only displays the first correct entry and then a javascript error occurs. i am reporting this error second hand from my users. i saw and experienced the errors myself, but my users decided to immediately upgrade to netscape 4.51 before i could get the exact text (blasted users!) the error mentions something like "bla is undefined". could someone please verify this with v4.04 and v4.05?

Reassigning to dmose@mozilla.org, who now has front-line responsibility for all Bonsai and Bugzilla bugs.

Reassigning back to me. That stuff about me no longer being the front-line person responsible for Bugzilla and Bonsai turned out to be short-lived. Please pardon our confusion, and I'm very sorry about the spam.

this patch fixes (or works around) the problem: diff -r1.33 query.cgi 245a246,253 > // Netscape 4.04 and 4.05 also choke with an "undefined" > // error. if someone can figure out how to "define" the > // whatever, we'll remove this hack. in the mean time, we'll > // assume that v4.00-4.03 also die, so we'll disable the neat > // javascript stuff for Netscape 4.05 and earlier. > var agtver = parseFloat(navigator.appVersion); > if (agtver <= 4.05 ) return; >

*** Bug 6230 has been marked as a duplicate of this bug. ***

This seems to happen with NN3.04/NT4 too, except that the component list does not change at all and the error message is slightly different: "selectProduct is not defined". On IE4.0/NT4 there is no error message, but again, the component list does not change at all.

OK, I have applied the above patch.

I'm trusting that applying the above patch fixed this bug.

Hi, I'm using NS3.0 (my favourite ;), because it's fast and small. I have found possibly related bug-id's (listed in brackets) for the following two problems. I'm a little bit lazy, so i only post this text here: ---------- 1. problem with push-function in JavaScript-part created by query.cgi: (Bug #6230, #5249, #4133, #4106, #21381) JavaScript1.1 support the prototype-statement for objects, but if you loop through your objects 'cpts' or 'vers' (for (c in cpts) ...), you will also get the 'push'-string (remember the setting of the push-function with the 'Array.prototype'-call) as value for 'c'. But cpts['push'] is of kind 'function', not an (array-)object, so you will get an error "Object doesnt' support this property or method" (as Bug #4133), when accessing it: 'for (j in cpts[c]) ...' In JavaScript1.2 this works fine. No error. But you loose the availability for browser < V4 and get errors when you've set the language-attribute of the script-tag to JS1.2. :( ---------- 2. problem with selectProduct-function created by query.cgi: (Bug #6230, #5249, #22269, #21381) I don't understand, why anybody put JavaScript-functions within the body of a html-page and not to the header. sigh. Anyway, in the comments for Bug #22269, 'dave@intrec.com' mentioned the main-cause for not including javascript-code in the body-tag; because when the browser wants to execute the onload-functions, it can't find them. But there's an easy solution. We must simple define the used functions, BEFORE they are referenced (see below). -------------------- So, I suggest two changes: for 1: push-problem (sorry, only know how to 'cvs diff -u'): *** update 'query.cgi' *** ============================ allow JS1.1 @@ -200,3 +200,3 @@ my $jscript = << 'ENDSCRIPT'; -<script language="Javascript1.2" type="text/javascript"> +<script language="Javascript1.1" type="text/javascript"> ============================ don't check for Browser-type @@ -257,3 +257,3 @@ var agtver = parseFloat(navigator.appVersion); - if (agtver <= 4.05 ) return; + // if (agtver <= 4.05 ) return; ============================ check for 'function'-type in array-object ... @@ -279,2 +279,3 @@ for (c in cpts) { + if (typeof(cpts[c]) == 'function') continue; ============================ @@ -284,2 +285,3 @@ for (j in cpts[c]) { + if (typeof(cpts[c][j]) == 'function') continue; ============================ @@ -311,2 +313,3 @@ for (v in vers) { + if (typeof(vers[v]) == 'function') continue; ============================ @@ -316,2 +319,3 @@ for (j in vers[v]) { + if (typeof(vers[v][j]) == 'function') continue; ============================ Instead of using the check 'if (typeof(..) == 'function') continue;', you can use 'if (typeof(..) != 'object') continue;', if you like. I can't test it for IE or NS4.04/4.05, maybe you have the possibility to do this(?). -------------------- for 2: selectProduct-problem: *** update 'query.cgi' *** ============================ Print JavaScript in header (within head-tag) @@ -355,3 +359,4 @@ PutHeader("Bugzilla Query Page", "Query Page", "", - q{onLoad="selectProduct(document.forms[0]);"}); + q{onLoad="selectProduct(document.forms[0]);"}, + $jscript); ============================ @@ -359,4 +364,2 @@ push @::legal_target_milestone, "---"; # Oy, what a hack. - -print $jscript; ============================ *** update 'CGI.pl' *** ============================ Print JavaScript in header (within head-tag) @@ -490,3 +490,3 @@ sub PutHeader { - my ($title, $h1, $h2, $extra) = (@_); + my ($title, $h1, $h2, $extra, $jscript) = (@_); ============================ @@ -503,3 +503,3 @@ print "<HTML><HEAD>\n<TITLE>$title</TITLE>\n"; - print Param("headerhtml") . "\n</HEAD>\n"; + print Param("headerhtml") . "\n${jscript}\n</HEAD>\n"; ============================ Maybe the other bug-id's are solved too, maybe not. Ciao, Greetings by JUG :)

additionally change the following to avoid perl-warnings of unset-var: =========================== @@ -490,3 +490,4 @@ sub PutHeader { my ($title, $h1, $h2, $extra, $jscript) = (@_); + $jscript ||= ''; ===========================

OK, I have applied these patches. Thanks!

per Terry, auto-verifying any resolved bug that hasn't been touched since before 2.10 was released.

Moving to Bugzilla product