Closed Bug 433700 Opened 16 years ago Closed 16 years ago

Legend change its position while changing value of a select box in the same fieldset

Categories

(Core :: Layout: Form Controls, defect)

Product:
Core
Component:
Layout: Form Controls
Type:
defect
Priority:
Not set
Severity:
minor

Tracking

()

Status:
RESOLVED FIXED
Milestone:
mozilla1.9.1a1

People

(Reporter: noguer, Assigned: MatsPalmgren_bugz)

Details

(Keywords: regression, testcase, verified1.9.0.2)

Attachments

(5 files, 3 obsolete files)

Reporter's testcase
485 bytes, text/html
Details
Testcase #3
2.27 KB, text/html
Details
Patch rev. 2
6.81 KB, patch
dbaron
: review+
dbaron
: superreview+
samuel.sidler+old
: approval1.9.0.2+
Details | Diff | Splinter Review
reftest2.diff
6.22 KB, patch
Details | Diff | Splinter Review
testcase: legend with margin-top moves when browser resized
3.77 KB, text/html
Details 
User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; ca; rv:1.9pre) Gecko/2008040907 Minefield/3.0pre
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; ca; rv:1.9pre) Gecko/2008040907 Minefield/3.0pre

I have this simple HTML, every time I select a new value in the selectbox, the legend of the fieldset change its position.

But If I delete "top:110px", or  "bottom:0px" or "margin-left:-50px" it works fine.

<html><head>
	<style>
.cContingut {	position:fixed;	top:110px;	bottom:0px;}
.cContingut fieldset { padding-left:65px; }
.cContingut legend  { margin-left: -50px; }
</style>
</head> 
<body>
<div class="cContingut">
<fieldset>
<legend>2 &nbsp;&nbsp;	Esculli el compte vinculat</legend>
<select id="Compte" name="Compte" size="1">	
<option >a</option>
<option >a</option>
<option >a</option>
<option >a</option>
<option >a</option>
</select>
</fieldset>
</body></html>

Reproducible: Always

Steps to Reproduce:
1. Load the page in Firefox 3, in Firefox 2 it works fine.
2. Select a value in the selectbox

Actual Results:  
The lengend change its position

Expected Results:  
The legend should rest in the same position all the time
Attached file Reporter's testcase — 

    
  
Severity: normal → minor
Status: UNCONFIRMED → NEW
Component: General → Layout: Form Controls
Ever confirmed: true
Keywords: regression, 
          
            testcase
OS: Windows XP → All
Product: Firefox → Core
QA Contact: general → layout.form-controls
Hardware: PC → All
Version: unspecified → Trunk

    
  
Assignee: nobody → mats.palmgren

    
    

    
  
The problem is that we never get the real value for 'legendMargin'
if we don't reflow the legend frame (we will use a zero margin),
this makes us think we need to place the legend somewhere else:
http://bonsai.mozilla.org/cvsblame.cgi?file=/mozilla/layout/forms/nsFieldSetFrame.cpp&rev=3.168&root=/cvsroot&mark=446,448,450,553#444


    
    

    
  

      
      
      
      

        Attached patch
          
          
        Patch rev. 1 (obsolete)
        — 
          — Splinter Review
      

    


  
The fix is the last change.  The rest are just some style nits.

        Attachment #320956 -
        Flags: superreview?(dbaron)

        Attachment #320956 -
        Flags: review?(dbaron)

    
    

    
  

      
      
      
      

        Attached patch
          
          
        reftest.diff (obsolete)
        — 
          — Splinter Review
      

    


  

    
    

    
  
Comment on attachment 320956 [details] [diff] [review]
Patch rev. 1

r+sr=dbaron, except I think it would make more sense to put your new code as an additional else to the already-existing chain:

if (reflowLegend) {
  ...
} else if (!mLegendFrame) {
  ...
} else {
  // your new line to set legendMargin here
}

Any reason that wouldn't work?

        Attachment #320956 -
        Flags: superreview?(dbaron)

        Attachment #320956 -
        Flags: superreview+

        Attachment #320956 -
        Flags: review?(dbaron)

        Attachment #320956 -
        Flags: review+

    
    

    
  
Does that work for percentage margins?  I'd think this should use GetUsedMargin(), not GetStyleMargin()->GetMargin().

In any case, we need a reftest here that also tests percentage margins.

    
    

    
  
er, yeah, this should use GetUsedMargin.

    
    

    
  

      
      
      
      

        Attached file
          
        Testcase #2 (obsolete)
        — 
      

    


  
Yes, GetUsedMargin() was what I meant.  Actually, percentage margins isn't
working at all for legends.  I think the first call should also be using
GetUsedMargin() and moved after the reflow...  I'll fix that too...

        Attachment #320956 -
        Attachment is obsolete: true

        Attachment #320957 -
        Attachment is obsolete: true

    
    

    
  

      
      
      
      

        Attached file
          
        Testcase #3
        — 
      

    


  

        Attachment #320968 -
        Attachment is obsolete: true

    
    

    
  

      
      
      
      

        Attached patch
          
          
        Patch rev. 2
        — 
          — Splinter Review
      

    


  
Use GetUsedMargin() in both places.

        Attachment #321001 -
        Flags: superreview?(dbaron)

        Attachment #321001 -
        Flags: review?(dbaron)

    
    

    
  

      
      
      
      

        Attached patch
          
          
        reftest2.diff
        — 
          — Splinter Review
      

    


  

    
    

    
  
Comment on attachment 321001 [details] [diff] [review]
Patch rev. 2

r+sr=dbaron assuming that you've tested that the % margins are %s of something sensible (i.e., the containing block width, for some reasonable definition of containing block)

        Attachment #321001 -
        Flags: superreview?(dbaron)

        Attachment #321001 -
        Flags: superreview+

        Attachment #321001 -
        Flags: review?(dbaron)

        Attachment #321001 -
        Flags: review+

    
    

    
  
Is there any solution for this ? The problem is the same in the last Firefox 3  -  Mozilla/5.0 (Windows; U; Windows NT 5.1; ca; rv:1.9pre) Gecko/2008060306 Minefield/3.0pre)  

    
    

    
  
Er... so how come we never checked this fix in?  :(
Flags: wanted1.9.1?
Flags: wanted1.9.0.x?

    
    

    
  
For RC2 you mean?  This wouldn't have been approved, so I didn't ask.
(I will land it on mozilla-central soon)

    
    

    
  
Ah, were we up to rc2 by then?  Yeah, this wouldn't have gone in at that late stage.

I still think this is worth fixing in 1.9.0.x.

    
    

    
  
Could you check this in?  And maybe add a testcase for comment 12, and double-check that it's sensible?

    
    

    
  
Both Linux unit test Tinderboxes are orange and I've been waiting HOURS for
them to go green...  The reftest covers % margins.

    
    

    
  
Well, I've added checkin-needed, so hopefully somebody else can land it for you at the next rare opportunity.

    
    

    
  
http://hg.mozilla.org/mozilla-central/index.cgi/rev/ba694e7645ea
http://hg.mozilla.org/mozilla-central/index.cgi/rev/69e0da0cdb2f

The reftest failed on qm-win2k3-03 for some reason so I disabled
it temporarily while I investigate:
http://hg.mozilla.org/mozilla-central/index.cgi/rev/b863a7474393

-> FIXED
Status: NEW → RESOLVED
Closed: 16 years ago
Flags: wanted1.9.1? → in-testsuite?
Keywords: checkin-needed
Resolution: --- → FIXED
Target Milestone: --- → mozilla1.9.1a1

    
    

    
  
Fixed and re-enabled the reftest:
http://hg.mozilla.org/mozilla-central/index.cgi/rev/b41ed6fdb780
Flags: in-testsuite? → in-testsuite+

    
  

        Attachment #321001 -
        Flags: approval1.9.0.2?

    
    

    
  
Comment on attachment 321001 [details] [diff] [review]
Patch rev. 2

Approved for 1.9.0.2. Please land in CVS. a=ss

Please be sure to land the tests in CVS as well.

        Attachment #321001 -
        Flags: approval1.9.0.2? → approval1.9.0.2+

    
    

    
  
Landed on CVS trunk for 1.9.0.2:
mozilla/layout/forms/nsFieldSetFrame.cpp 	3.169
mozilla/layout/reftests/bugs/433700-ref.html 	1.1
mozilla/layout/reftests/bugs/433700.html 	1.1
mozilla/layout/reftests/bugs/reftest.list 	1.457 

Flags: wanted1.9.0.x?
Keywords: fixed1.9.0.2

    
    

    
  


  
A repro of a related problem that may or may not have the same cause as the existing bug.  Legend moves when browser (Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.0.1) Gecko/2008070208 Firefox/3.0.1) is resized.

    
    

    
  
I hit a similar bug where the legend (set with position: relative and a
negative margin) of a fieldset moves. I see it when the browser is resized
rather than when a select element is used.  Can someone tell me if this is the
same bug and if the same fix will apply or if this needs to be entered as a new
bug?  Is the fix for this bug in a release version yet?

Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.0.1) Gecko/2008070208
Firefox/3.0.1

See attachment LegendMoveRepro.html "testcase: legend with margin-top moves
when browser resized."

    
    

    
  
Verified for 1.9.0.2 with Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-US; rv:1.9.0.2) Gecko/2008090212 Firefox/3.0.2.
Keywords: fixed1.9.0.2 → verified1.9.0.2

          You need to log in
          before you can comment on or make changes to this bug.
        






  

    
  







  

    

      
Attachment

      

      
      
      
      
    

    

      

        

          

            
General

            

              Creator: 



            

            
Created: 

            
Updated: 

            
Size: