346980 - branch patch for cellmap

Status: RESOLVED FIXED

(Core :: Layout: Tables, defect)

Description

[Bernd]

I fixed a couple of bugs blocking bug 339128 on trunk. IMHO the patches should be ported to 1.8.1. I don't see any point in cherry picking the patches all of them lead to crashes. I will do the patch today or tomorrow so that it can be pushed towards branches.

Comment 1

[Bernd]

Attachment: patch against MOZILLA_1_8_BRANCH

this is the promised patch which I think should go on 1.8.1 and probably also on 1.8.0.7

Comment 2

[Bernd]

Boris could you have an eye on this while I am away (till 2006-08-21)

Comment 3

[Mike Schroepfer]

Bernd/Bz how long has this been on trunk?  Any risk for the branch?

Comment 4

[Boris Zbarsky [:bzbarsky]]

Variable amounts, from "as of this morning" for bug 343087 to "at least a month" for some of the other bugs.  There is probably some risk, but I'm not quite sure how much; the exact regression tracking and whatnot has been happening in the bugs this one depends on...

I do think we should let things bake on trunk for a few days before doing anything with branch, but the most recent change also looked reasonably safe to me, at least as this code goes.  

Comment 5

[Bernd]

One has to weight the risk. This will/should fix also the  other branch blocking bugs that block bug 339128 and do not have a patch till now. With the patches I can't reproduce them.

http://bonsai.mozilla.org/cvsview2.cgi?diff_mode=context&whitespace_mode=show&file=nsTableFrame.cpp&branch=&root=/cvsroot&subdir=mozilla/layout/tables&command=DIFF_FRAMESET&rev1=3.650&rev2=3.651

is not included in the patch but should be applied otherwise VC6 will complain.

A good way to estimate the risk would be to ask Jesse to run the tests from bug 339128 and see whether he sees regressions.

Comment 6

[Mike Schroepfer]

Comment on attachment 232193 [details] [diff] [review]
patch against MOZILLA_1_8_BRANCH

a=schrep for drivers.

Comment 7

[Boris Zbarsky [:bzbarsky]]

I had to edit the diff to remove the fixes for bug 339315 and bug 302911, which had already landed on the branch.  With those removed, this applied cleanly and all.  Checked in on 1.8 branch.

Comment 8

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

sayrer just landed a bustage fix for these changes -- for for loop scoping on Windows.

Comment 9

[Bernd]

Comment on attachment 232193 [details] [diff] [review]
patch against MOZILLA_1_8_BRANCH

I guess this should go on the 1.8.0 branch but I did not get enough baking on trunk, I fixed only today one of the regressions

Comment 10

[Bernd]

Comment on attachment 232193 [details] [diff] [review]
patch against MOZILLA_1_8_BRANCH

I had some discussion with dveditz, where I was more conservative with taking this onto 1.8.0.7. My main concern was that I did not do enough testing and together with still visible asserts. From a stability point of view this patch highly desirable especially when combined with the patch in bug 350081 and the regression fix ofr bug 347796. Just in case that the 1.8.0.7 release will slip asking for inclusion into 1.8.0.7. If I missed the boat this is definetly 1.8.0.8 material.

Comment 11

[Daniel Veditz [:dveditz] (OOO: back Aug 26)]

Comment on attachment 232193 [details] [diff] [review]
patch against MOZILLA_1_8_BRANCH

approved for 1.8.0 branch, a=dveditz for drivers

We're slipping to tomorrow, please land today.

Comment 12

[Bernd]

re [need testcase] 
Jay, you will not get a testcase, this bug accumulated the patches from the dependent bugs into one single patch as they partially superseded each other. The testcases are in the dependent bugs, this bug is verified once all dependent bugs are verified.