660282 - Map at gulesider.no not loading

Status: RESOLVED WORKSFORME

(Tech Evangelism Graveyard :: Norwegian, defect)

Description

[heraldo]

User-Agent:       Mozilla/5.0 (Windows NT 6.0; rv:5.0) Gecko/20100101 Firefox/5.0
Build Identifier:  Mozilla/5.0 (Windows NT 6.0; rv:5.0) Gecko/20100101 Firefox/5.0

1. Go to the mentioned URL
2. Map is not loaded, empty <div id="mymap" class="eniro"></div>

Reproducible: Always

Comment 1

[David Mandelin [:dmandelin]]

Confirmed. heraldo, would you be able to get us a regression range on this using nightly builds?

Comment 2

[heraldo]

Looks like it regressed somewhere between 2011-04-12-03-mozilla-central and 2011-04-13-03-mozilla-central

Comment 3

[Boris Zbarsky [:bzbarsky]]

heraldo, what are the changeset ids for those two builds (in about:buildconfig)?

Comment 4

[heraldo]

http://hg.mozilla.org/mozilla-central/rev/a174b86200d6
http://hg.mozilla.org/mozilla-central/rev/6bcaec19d09e

Comment 5

[heraldo]

In error console on my current install (5beta1):

Error: Eniro is not defined
Source File: http://kart.gulesider.no/rs/profiles/profile-no.js?v=40.14
Line: 2

Error: OpenLayers is not defined
Source File: http://kart.gulesider.no/rs/profiles/lang-nb_no.js?v=40.14
Line: 7

Error: Eniro is not defined
Source File: http://kart.gulesider.no/assets/D41D8CD98F00B204E9800998ECF8427E.js?profile=no&v=40.14
Line: 1

Error: Eniro is not defined
Source File: http://kart.gulesider.no/
Line: 3

Comment 6

[Boris Zbarsky [:bzbarsky]]

Range from comment 4: http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=a174b86200d6&tochange=6bcaec19d09e

Lots of interesting changes in here, but the one I seemed to recall happening around April 15 is definitely in that range: the version bump to 5.0a1.

Looking at the source of http://www.gulesider.no/rs/Eniro.js?v=40.14 we have this bit:

  fnug.loadBundle = function (a) {
    var b = navigator.userAgent.indexOf("AppleWebKit") > 0,
        c = navigator.userAgent.indexOf("MSIE") > 0,
        f = navigator.userAgent.indexOf("Firefox/4") > 0;
    f = b || c || f;
....
    fnug.loadScript(d, f);

and this bit:

  fnug.loadScript = function (a, b) {
    if (b) document.write('<script type="text/javascript" src="' +
                          fnug.resourcePath(a) + '"><\/script>');
    else {
        var c = document.createElement("script");
        c.type = "text/javascript";
        c.src = fnug.resourcePath(a);
        document.getElementsByTagName("head")[0].appendChild(c)
    }
  }; 

So these people "worked around" the fact that we switched to HTML5-specified handling of <script> tags by specifically detecting Firefox 4... instead of just using the codepath that works in all browsers for all browsers.  What they do instead is only use that codepath for WebKit and IE and Firefox 4, and use the codepath that HTML5 says shouldn't work for other browsers.

If I tack "Firefox/4" to the end of the UA string in a trunk build, the site works, as expected.

Comment 7

[Boris Zbarsky [:bzbarsky]]

Er, I meant the version bump happened around April 12, of course.

Comment 8

[Boris Zbarsky [:bzbarsky]]

heraldo, are you willing to contact the site and let them know about the problem?  I can try, but my ability to read Norwegian is not sufficient to find the "contact us" part of the page...

Comment 9

[heraldo]

Done, http://www.gulesider.no/hjelp/kontakt.html