Open Bug 90741 Opened 24 years ago Updated 3 years ago

XBL <xul:box> tags that don't have inherits="orient" should be changed to <xul:hbox> or <xul:vbox>

Categories

(Core :: XUL, defect)

x86
All
defect

Tracking

()

People

(Reporter: hamfastgamgee, Unassigned)

Details

Attachments

(9 files)

This bug tracks finishing off bug 73671. It will encompass the XBL changes, one change that was missed on the trunk, and some mailnews and security changes that were left for later due to (now shown to be invalid) bug 88332. These have all been tested locally, but I'm Cc:ing everyone I think would want to test and review this stuff. The major part of this bug are the XBL changes; the rest of them are real easy to test and should be able to be cleared quickly. (Also Cc:ing rginda, to let him know that he can change Venkman at his convenience. :) )
I'm not gonna post a diff for mailnews until the folder outliner lands. :)
Status: NEW → ASSIGNED
r=ddrinan.
sr=hewitt
The security changes (r=ddrinan, sr=hewitt) have been checked in. Still waiting on r= for the other small change (the sr=hewitt was for both this and the security patch). No reviews as yet for the XBL stuff, which is not surprising. :)
I swept through every overlay in the entire tree and found any instance of a <vbox> that overlayed an <hbox> in the main file and corrected it (hence the latest patch). Only a couple of these are actually visible in the product (the Edit List function in the Address Book, for instance), thanks to additional <hbox> or <vbox> elements inside of the opening <vbox>, but these changes are good to make nonetheless. This does not touch pref-applications-new.xul (which already has a patch submitted) and the mailnews changes which are contingent on the folder outliner (I don't want to screw up their landing by relatively small changes in two XUL files).
Seth should look at the mailnews changes, and someone from the editor team (probably cmanske) needs to see the editor ones. sr=blake for the lone xpfe change
I would be happy if we could wait with this for a while (at least the mailnews part)... It's low gain and pretty high risk.
Fixes for regressions are high risk?
Last week you horked us... I just want to make sure you don't do that again. Especially not now when we're anticipating the folder_outliner branch.
Mark explained to me that the XBL changes are for later. I looked through the patch from 07/14/01 13:17, and it looks good to me. It fixes regressions (as opposed to spawning new ones. :) r=hwaara on that one, (attachment 42301 [details] [diff] [review])
r=cmanske for all composer and mailnews changes. All of this looks good.
sr=kin@netscape.com on the editor changes.
I checked the editor patch in to the trunk for Mark, per our IRC conversation.
http://bugzilla.mozilla.org/showattachment.cgi?attach_id=43322 seems ok to me. last time mailnews got horked by innocent looking changes. please double check that all affected dialogs are not going to be messed up. after that, sr=sspitzer
I double and then triple checked it before hwaara would give me an r=. It's quite well tested. :)
Yay! All the regression fixes are in (as of timeless' checkin today, July 24)! Updating summary appropriately.
Summary: Whack the rest of the <box> tags in the trunk → XBL <xul:box> tags that don't have inherits="orient" should be changed to <xul:hbox> or <xul:vbox>
Attached patch XBL changesSplinter Review
The most recent attachment is a set of smaller XBL changes that only do three things. 1) <xul:box orient="vertical"> -> <xul:vbox> 2) <xul:box orient="horizontal"> -> <xul:hbox> 3) A bunch of minor whitespace fixes (including whacking some tabs) This excludes the fourth, and much more risky, change of <xul:box> (without orient="horizontal") -> <xul:hbox>. hewitt didn't think there was a risk in that, but this new patch is much easier to review (and I've been running with it locally for some time with no regressions [actually, I've been running the full patch for some time with no regressions] that I could see; there aren't test cases for all XBL widgets, though, especially on platforms that I can't test on) and is safer. Unless <xul:box> is going to have an alignment change at some point, I'm happy with just doing the smaller change (after the last time and causing regressions, I'm being more cautious). <xul:box> == <xul:hbox>, but it's just not worth it.
sr=hewitt
fixed
whoops, not really fixed, I only landed the xbl portion of mark's changes.
Component: XP Toolkit/Widgets: XUL → XUL
QA Contact: jrgmorrison → xptoolkit.widgets
Resetting assignee to default after no activity for 7 years
Assignee: andersma → nobody
Status: ASSIGNED → NEW
Severity: normal → S3
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: