311071 - treat <! as the start of a comment to end of line, unless e4x=1 (was: page keeps reloading)

Status: RESOLVED FIXED

(Core :: JavaScript Engine, defect, P1)

Description

[Peter van der Woude [:Peter6]]

Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.8b5) Gecko/20051004
Firefox/1.4.1 ID:2005100402

after patch Bug 310993 landed the url above keeps reloading

works in 20051003 2217pdt build
fails in 20051004 0252pdt build

Comment 1

[Peter van der Woude [:Peter6]]

same regressionrange
http://maps.google.com/ doesn't work anymore

Comment 2

[sashk]

https://riteaid.rebateplus.com/RebateListPage.asp?categories=20
check box next to, let's say razor and click add to list - doesn't work anymore
Mozilla/5.0 (Macintosh; U; PPC Mac OS X Mach-O; en-US; rv:1.8b5) Gecko/20051004
Firefox/1.4.1 ID:2005100405

worked fine on 20051003

Comment 3

[Blake Kaplan (:mrbkap) (inactive)]

comment 2 is a different bug, caused by bug 310351. I'll open a new bug to fix it.

Comment 4

[Brendan Eich [:brendan]]

Hrm, maps.google.com works for me still with an updated js/src and an older tree
(I am testing js changes in isolation from other recent Gecko, etc., changes). 
Peter, what doesn't work for you?  The maps.google.com problem should be filed
separately.

The riteaid URL and reproduce-by steps in comment 2 work for me, so I suspect
that's a separate issue too.

Patch in a minute for the problem affecting this bug's URL.

/be

Comment 5

[Brendan Eich [:brendan]]

Attachment: treat <! as the start of a comment to end of line, unless e4x=1

I feel like Bart Simpson, writing "I will respect backward compatibility
authority" 100 times on the blackboard.

/be

Comment 6

[Peter van der Woude [:Peter6]]

(In reply to comment #4)
> Hrm, maps.google.com works for me still with an updated js/src and an older tree
> (I am testing js changes in isolation from other recent Gecko, etc., changes). 
> Peter, what doesn't work for you?  The maps.google.com problem should be filed
> separately.

> /be

It just gives a gray empty screen where the map should be (confirmed by many)

Comment 7

[Brendan Eich [:brendan]]

(In reply to comment #6)
> It just gives a gray empty screen where the map should be (confirmed by many)

I believe you, and I said so -- it's just not due to the patch that regressed
this bug's URL, which was the patch for bug 310993.  I can prove that by backing
out that patch and fixing this bug, in my tree with an older Gecko but up to
date JS engine source.  In my tree, maps.google.com works fine with or without
the patch for bug 310993.  So the maps.google.com regression is due to a non-JS
change.

This bug is fixed by the attached patch.

/be

Comment 8

[Blake Kaplan (:mrbkap) (inactive)]

Comment on attachment 198475 [details] [diff] [review]
treat <! as the start of a comment to end of line, unless e4x=1

I still want my t-shirt. r=mrbkap

Comment 9

[Peter van der Woude [:Peter6]]

I filed Bug 311094 for maps.google.com

Comment 10

[Scott MacGregor]

Comment on attachment 198475 [details] [diff] [review]
treat <! as the start of a comment to end of line, unless e4x=1

mike can you give us an extra level of review here since we may spin the
release for this. Extra eyes is always good.

Comment 11

[Mike Shaver (:shaver emeritus)]

Comment on attachment 198475 [details] [diff] [review]
treat <! as the start of a comment to end of line, unless e4x=1

sr=shaver

Comment 12

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Comment on attachment 198475 [details] [diff] [review]
treat <! as the start of a comment to end of line, unless e4x=1

approving based on mscott's email that said this was blocking the respin

Comment 13

[Brendan Eich [:brendan]]

Fixed on branch and trunk.

/be

Comment 14

[Peter van der Woude [:Peter6]]

Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.8b5) Gecko/20051004
Firefox/1.4.1 ID:2005100415

Verified on win

Comment 15

[blitzreiter]

Please check also wto.int menu

Comment 16

[Brendan Eich [:brendan]]

(In reply to comment #15)
> Please check also wto.int menu

Eh?

/be

Comment 17

[Jesse Ruderman]

I assume comment 15 has something to do with http://www.wto.int/.

Comment 18

[Bob Clary [:bc] (inactive)]

Checking in regress-311071.js;
/cvsroot/mozilla/js/tests/js1_5/Regress/regress-311071.js,v  <--  regress-311071.js
initial revision: 1.1
done

Comment 19

[Boris Zbarsky [:bzbarsky]]

This causes JS_ASSERT failures and ensuing crashes -- see bug 311950.