Require.js parsing errors

[jfsiii]

Require.js parsing errors

[cramforce]

I think we should revert the experimental changes in the gh-pages branch. This is a side effect of the bespin thing I think

[cramforce]

Please pull from gh-pages branch and see if error persists

[jfsiii]

Deleted repo and forked anew. No issues observed in Chrome 7.0.503.0 dev. However, issues are still seen in FF 3.6.6 and 4.0b4.

The issues still seem related to require.js. http://jfsiii.streamie.org/ext/require.js and http://cramforce.streamie.org/ext/require.js are both exhibiting the odd and disconcerting behavior of returning different responses on each load. One time the response is 750 lines, the next is 850 lines, another is 1100, and so on.

Even more strange is that the responses are "mangled", for lack of a better term. See http://pastie.org/1125912 for a representative example.

Accessing http://cramforce.streamie.org/ext/require.js via curl (and Chrome) does not show any of these inconsistencies. http://screencast.com/t/NmM2MTQwYW shows the a mangled portion of the response when http://cramforce.streamie.org/ext/require.js is entered in the location bar.

My copy of FF4b4 is brand new, with no extensions installed. FF 3.6.6 behaves the same with all extensions disabled.

I will re-apply my fixed for the unterminated multi-line comment and see if, perhaps, that is what's causing FF behave so oddly and erratically.

[antimatter15]

I'm getting the issue on chrome, and chromium. Not on wget however.

[cramforce]

Can you pull in jfsiii's fixes to see whether that works?

[cramforce]

Or by pulling the minified version. I could commit that one, but I prefer unminified code.

[cramforce]

Should be fine now

[antimatter15]

AAARRGGHHH!!!
It's happening again!!!!

FInstall&&A())NIaTU
88i3
i3ranifLene8m \dl
8tall&&A())NIaTU
88i3
i3ranifLene8m \dl
8tall&&A())NIaTU
88i3
id==3
id==3
id==3
id==3
id==3
id==3
id==3
idi|i{
}|i{
}|i{
}|i{
}|i{
}|i{
}|i{
}|i{
iAef\lUref\l\lUpDlUm

|i{

[cramforce]

Yeah, I've seen it too. Then it goes away again. It is probably a bug in http-proxy. Do you have time to investigate? Maybe it only occurs under heavy load. The server is getting quite a hammering :)