Closed Bug 337292 Opened 18 years ago Closed 18 years ago

Since searchbox oval is a single hard-coded image, it breaks with the new searchbar flex behavior

Categories

(Firefox :: Search, defect)

2.0 Branch
PowerPC
macOS
defect
Not set
blocker

Tracking

()

VERIFIED FIXED
Firefox 2 alpha2

People

(Reporter: mozilla, Assigned: mozilla)

References

Details

(Keywords: fixed1.8.1)

Attachments

(7 files, 1 obsolete file)

We need to either disable the flex on the Mac or break the searchbox background into parts.
Joe, should we back out the flex patch from bug 258679 or IFDEF around it for Mac for A2? This ends up making it possible to have some really odd looking search boxes, and even makes it hard to switch engines in some cases.
I vote for IFDEFing it out on Mac OS X, so that we still get the benefit of the enlarged searchbox on other platforms, but let's do whatever we need to do, 'cause we shouldn't ship it like this.  Sorry for the breakage.
Attached patch ifdef it out for Mac (obsolete) — Splinter Review
I think this is all that's needed for reverting this on Mac (the flex=1 on the search bar won't matter, right?). Can someone test this patch on mac?
Attachment #221459 - Flags: ui-review?(beltzner)
Attachment #221459 - Flags: review?(joe)
Attachment #221459 - Attachment is obsolete: true
Attachment #221460 - Flags: ui-review?(beltzner)
Attachment #221460 - Flags: review?(joe)
Attachment #221459 - Flags: ui-review?(beltzner)
Attachment #221459 - Flags: review?(joe)
Comment on attachment 221460 [details] [diff] [review]
really ifdef it out for Mac

Tested OK on my Mac.
Attachment #221460 - Flags: review?(joe) → review+
Comment on attachment 221460 [details] [diff] [review]
really ifdef it out for Mac

Yup. This seems like the best case for A2, and we should fix it properly for B1.
Attachment #221460 - Flags: ui-review?(beltzner) → ui-review+
Assignee: nobody → gavin.sharp
Flags: blocking-firefox2?
Target Milestone: --- → Firefox 2 alpha2
BTW, marking severity as "blocker" (relative to a2) since that's what we decided at today's Bon Echo status meeting.
Severity: minor → blocker
Attachment #221523 - Flags: superreview?(bugs)
I've tested this on the 1.8 branch in OS X and it looks good.
Comment on attachment 221523 [details] [diff] [review]
This patch makes the searchbar look good at arbitrary widths on the Mac (removing the need for ifdefing the flex on the Mac)

I get the feeling there should be _a_ container around the textbox that can be bound to in the theme then have the binding include the textbox in its <children/>... but I'm not sold to that idea. Let's see what gavin says, then I'll mark sr.
Comment on attachment 221523 [details] [diff] [review]
This patch makes the searchbar look good at arbitrary widths on the Mac (removing the need for ifdefing the flex on the Mac)

>Index: browser/components/search/content/search.xml

>       <handler event="keypress" keycode="vk_up" modifiers="accel"
>                phase="capturing"
>-               action="this.parentNode.selectEngine(event, false);"/>
>+               action="this._getParentSearchbar().selectEngine(event, false);"
>+               />

leave the /> on the same line? This file seems to have lost the 80-char-line war a while ago :).

>Index: browser/themes/pinstripe/browser/jar.mn

>   skin/classic/browser/search-bar-background.png

Don't forget to cvs remove this file (in addition to the jar.mn removal from the new patch). Actually, maybe it's best to just change this file to be the "mid" you attached, and not rename it? How do the two differ?

I think this approach is fine. Having it be theme (and thus platform-) specific could lead to unexpected breakage when changing the search bar binding, and it's easier for themes to not have to deal with binding the search bar and instead just style the <box>s if desired.
Attachment #221523 - Flags: review?(gavin.sharp) → review+
Blocks: 337147
Assignee: gavin.sharp → joe
Comment on attachment 221523 [details] [diff] [review]
This patch makes the searchbar look good at arbitrary widths on the Mac (removing the need for ifdefing the flex on the Mac)

Ok, looks good enough to ship, getting this landed soon for a2
Attachment #221523 - Flags: superreview?(bugs)
Attachment #221523 - Flags: superreview+
Attachment #221523 - Flags: approval-branch-1.8.1+
Landed this on the 1.8 branch with the nit addressed. I haven't cvs removed the now-unused icon yet, can do that later if needed.
Keywords: fixed1.8.1
Gavin, can you post what actually got checked in?  I don't think it's working anymore: see attached composite screenshot.

I just did a sync and full build, and either my build is obscurely broken, or the final patch wasn't functionally equivalent to Joe's from Comment #7.  This fix doesn't appear to have made it into last night's Mac nightly, so I can't double-check there.
(In reply to comment #18)
> I just did a sync and full build, and either my build is obscurely broken, or
> the final patch wasn't functionally equivalent to Joe's from Comment #7.  This
> fix doesn't appear to have made it into last night's Mac nightly, so I can't
> double-check there.

Really?  It manifested properly for me in the nightlies I downloaded both at home and here...

The nightly looks good to me too now.  Must be something in my build.  Ignore my comment #18, please.
Just for the record, the only thing different between my branch checkin and attachment 221634 [details] [diff] [review] is my nit from comment 15 about the "/>", and attachment 221634 [details] [diff] [review] only differs from attachment 221523 [details] [diff] [review] in that it removes search-bar-background.png (unused) from jar.mn.
I also landed this on trunk, and removed the unused image that gavin had left in on his branch landing.
Status: NEW → RESOLVED
Closed: 18 years ago
Resolution: --- → FIXED
Status: RESOLVED → VERIFIED
Flags: blocking-firefox2? → blocking-firefox2+
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: