185812 - refresh does not refresh all frames

Status: RESOLVED INVALID

(Core :: Layout: Images, Video, and HTML Frames, defect)

Description

[Ricardo Almeida]

User-Agent:       Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.3a) Gecko/20021207 Phoenix/0.5
Build Identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.3a) Gecko/20021207 Phoenix/0.5

When refreshing the page, only the first frame is refreshed...

The refered page (http://dialnp.di.fc.ul.pt/~ralmeida/) although having it's
content in portuguese, allows to see the bug...
The page uses 2 frames. One on the top for the menu and another for the content.

Reproducible: Always

Steps to Reproduce:
1. Go to http://dialnp.di.fc.ul.pt/~ralmeida/
    When the page starts it shows the home... 
2. Select some other tab in the menu (Links, for instance)
3. Press the refresh button on the browser

Actual Results:  
Refreshes only the menu frame, changing the selected tab to Home, but leaving
the contents frame with the links.

Expected Results:  
Refresh index.html, i.e., since there are two frames, both documents should be
reloaded with the pages stated in index.html. (In the example, home)

This report is made using Phoenix, but I also tested with Mozilla 1.2.1 and the
results were the same...

Comment 1

[José Jeria]

WFM 2002121608, Windows 2000, hitting refresh button reload both frames.

Comment 2

[Owen Marshall (Not reading bugmail)]

Actually, I do see this on an older build (2002112904). When CVS is done
building I will check there.

Comment 3

[R.K.Aa.]

testing with day old trunk CVS, Linux:

Seems to depend on cache settings. If i ctrl+reload, both frames are reset on a
reload. If i only reload and have prefs set to compare "when page is out of
date" - only the upper frame is reset.

Comment 4

[José Jeria]

Same behavior in IE 6 SP1, is this really a bug?

Comment 5

[Owen Marshall (Not reading bugmail)]

I dont know the expected behavior, and I don't know who would =) The strange
refreshing sounds like a netlib issue (if it is an issue at all), should we punt
it over?

Comment 6

[Reid Rankin]

Attachment: Bug #185812 testcase

Testcase, explains problem better, even though this is an error on the part of
the web page coder.

Comment 7

[Reid Rankin]

Attachment: Bug #185812 testcase (ancillary file)

Comment 8

[Reid Rankin]

Attachment: Bug #185812 testcase (ancillary file)

Comment 9

[Reid Rankin]

Attachment: Bug #185812 testcase (ancillary file)

Comment 10

[Reid Rankin]

Attachment: Bug #185812 testcase (ancillary file)

Comment 11

[Reid Rankin]

Comment on attachment 112588 [details]
Bug #185812 testcase

><HTML><HEAD><TITLE>Bugzilla bug 185812 testcase</TITLE>
>
><!--See nav.html for all details--></HEAD>
><FRAMESET cols="50%,*"><FRAME src="http://bugzilla.mozilla.org/attachment.cgi?id=112589&action=view"/><FRAME name="main" src="orig.html"/></FRAMESET></HTML>

Comment 12

[Reid Rankin]

Comment on attachment 112588 [details]
Bug #185812 testcase

<HTML><HEAD><TITLE>Bugzilla bug 185812 testcase</TITLE>

<!--See nav.html for all details--></HEAD>
<FRAMESET cols="50%,*"><FRAME
src="http://bugzilla.mozilla.org/attachment.cgi?id=112589&action=view"/><FRAME
name="main" src="orig.html"/></FRAMESET></HTML>

Comment 13

[Reid Rankin]

Attachment: Bug #185812 testcase (fixed ancilliary file)

Comment 14

[Reid Rankin]

Attachment: Bug #185812 testcase (fixed)

Comment 15

[Reid Rankin]

Attachment: Bug #185812 testcase (fixed)

Comment 16

[Reid Rankin]

I rather seriously messed up on the testcase. BTW, both fixed testcases are the
same.

Although the "bug" doesn't apply to CGI, (Browser logic: frameset page URL
contains GET string, therefore it's a dynamically generated page, so I must send
GET string again, so I must reload frameset page) so you will probably have to
download it to see the testcase in action.

Comment 17

[Owen Marshall (Not reading bugmail)]

CC'ing Darin. Darin, do you know what is expected? If you don't, do you know who
might?

Comment 18

[Boris Zbarsky [:bzbarsky]]

I talked to radha about this recently.  The expected behavior is that a simple
reload will reload the frames as they are (that is, if you clicked on a link in
one of the frames the _new_ thing should be reloaded in that frame). 
Shift-reload should reload the frameset and reset all the frames.