We need to get XMLHttpRequest into gklayout to be able to use some GC helper utilities that live there. And we're never going to ship gecko without XMLHttpRequest anyway so there's not much point to keeping it in mozilla/extensions. So we could either move it to mozilla/content/base, or to dom/src/base. Oppinions? Takers?

I vote for mozilla/content/base.

also note bug 315890

(In reply to comment #2) > also note bug 315890 Sure, but that didn't help for DOMCI.

We hit an assert in nsDOMClassInfo::PostCreate (http://bonsai.mozilla.org/cvsblame.cgi?file=mozilla/dom/src/base/nsDOMClassInfo.cpp&rev=1.351#3285) because nsIDOMParser should really be named nsIDOMDOMParser.

This one doesn't hit assertions in DOMCI.

The interface isn't frozen so we could in theory rename it nsIDOMDOMParser

Comment on attachment 211260 [details] CVS moves - v1 Looking for moa on the new locations. I thought about moving them to content/xml, but that'd require a new content/xml subdirectory (doesn't really fit in content or document).

Alternativly, we could merge the xml/content/* and xml/document/* directories into just xml/*. But that might be a bigger change then this bug.

reverting accidental owner change

Comment on attachment 211260 [details] CVS moves - v1 r+sr=jst

Comment on attachment 219137 [details] [diff] [review] v1.1 r+sr=jst

Finally done, sorry about the delays.

Comment on attachment 219137 [details] [diff] [review] v1.1 So I'd like to merge this to the 1.8 branch (see bug 336791), but I'm going to request approval for this one separately because it's a large patch, and I'd rather not have it be part of my diff in bug 336791.

Here's an actual merged patch. I'll also attach a diff showing what changed after doing the cvs adds and removes. I haven't done any runtime testing of this yet since I built with a bad configuration. I'll be able to test it in a few hours, but hoping to get approval before you're asleep, perhaps.

Here's the interdiff showing what actually changed other than the moves. Note that I had to merge in some constructor bits from bug 314931 as well.

Patch seems fine; Google Maps and GMail both work.

I landed this on MOZILLA_1_8_BRANCH on the basis of a discussion I had a few weeks ago with jst; I couldn't find anyone else around to approve, and I needed it in to finish merging bug 336791.

(Some after-the-fact review of the merging would still be good.)

This seems to have broken Camino: http://tinderbox.mozilla.org/showlog.cgi?log=Camino/1149943620.4759.gz

Ah, I missed the dependency on bug 335687; I'm not sure how easy it is to make that change without knowing the tools...

You also didn't pick up the allmakefiles.sh change and some of the packaging changes from bug 336244. Nothing major, but still nice to have.

> Ah, I missed the dependency on bug 335687; I'm not sure how easy it is to make > that change without knowing the tools... I had Mano check in Camino project changes that should unbust maya on the 18branch next cycle. (It's basically a 30sec fix, provided you're still running 10.3.9 and Xcode 1.5--which is basically just me these days....)