266915 - navbar button with target in other frame opens the page in the navbar, not target.

Status: RESOLVED WORKSFORME

(Firefox :: General, defect)

Description

[rfb]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.7.3) Gecko/20041028 Firefox/1.0RC1
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.7.3) Gecko/20041028 Firefox/1.0RC1

This is similar to other frame related bugs like 235824, but they seem to have
too many "worksforme" as the last action on them.  I cannot accept that.  This
is a MAJOR problem that affects my ability to generate a website that is FF
compatible.

This definitely fails in FF, but works in IE 5.5, 6.1 and Netscape 7.1, 7.2 and
AOL. Two of my sites using frames on (RedHat, QuadCPU, MySQL, PHP, Apache).  One
site (http://www.badronaldstattooing.com) where the main page navbar using
.shtml pages sent from a button push in "bar" to open in "main" works fine. When
 pressing "galleries" button, the second set of frames for picture galleries has
16 buttons sending .php files to the "main" frame. All work except for two. 
They are all coded nearly identically.  The one odd thing is that the two that
fail have the word "face" in the GET string, and nothing else is different from
the other button statements.

My current addon to the kickstart-1 webpages is the extraction of a .5 gb Adobe
document into individual webpages of each catalog page.  I made buttons in the
nav-bar frame to open a set of page thumbnail buttons for the appropriate
document section in the target window.  When you click a button in the target
window, it brings up an image of that page of the document where a person can
then enter page #, catalog#, of an item in the shopping cart.
I've barely got a skeleton of the 1100 pages started, a main entry with three
buttons on left frame to open coverpage,eventspage and TOC in main window.  TOC
brings up an "index" page where area maps of some numbered items at top of index
page section are in the main window and a couple dozen section buttons on left.
  Sections buttons match areamapped text at top of TOC page for target code. 
All left frame section buttons from initial page of three buttons and second
level section buttons fail 100%.  



Reproducible: Always
Steps to Reproduce:
1.1. go to http://www.kickstart-1.com/dragpages/index.php
1.2. see left frame with DragSpecialties logo and three page thumbnails. Right
frame contains catalog cover image.
1.3. push any of the left frame buttons.
or
2.1. go to http://www.kickstart-1.com/dragpages/indexPages.php
2.2. See orange buttons in left frame, TableOfContents mapped image in right frame.
2.3. In right frame, press in area of "1. Windshields, Dresser and Touring  22-59"
You should see a new page of page thumbnails appear in right frame. That is the
correct output.
2.4. Press the orange "Windshields, Dresser & Touring" button in the left frame.
Do not try any other buttons because they are not programmed yet.  I'm just
setting up the skeleton right now.

or
3.1. go to http://www.badronaldstattooing.com/
3.2. see all left panel buttons open pages in right frame correctly.
3.3. press the Galleries button in left frame.
3.4. new set of left frame buttons and new right frame appears with right frame
items matching those buttons in left frame.  All left frame buttons work
correctly excepting the "Faces/Figures" and "Skulls" buttons.

Actual Results:  
section 1.  All of the three page thumbnails open their target pages inside left
frame instead of right frame.

section 2.  Orange Windows button opens target page in left frame instead of
right frame.

section 3.  All left from buttons bring up targets in right frame correctly
except the "Faces/Figures" and "Skulls" buttons, which open their target pages
inside the left frame.

Expected Results:  
section 1.  First button should duplicate initial entry with 3 left buttons and
catalog cover in right, second button should bring up Events page in right,
third button should bring up page of multiple index page buttons in right frame
and orange buttons in left. Don't know where orange buttons went.

section 2.  Left frame orange "Windshields" button should have opened up page
full of three initial left frame thumbnails and several thumbsize indexpages
buttons in right side frame. and reloaded all orange buttons in left frame.
Instead, the right frame contents come up in left nav frame and orange buttons
go away.

section 3.  The "Faces/Figures" and "Skulls" buttons should have brought up
their targets in right frame just like other buttons do. Instead they open in
the left frame.

The "Faces/Figures" and "Skulls" buttons both have the word "face" in the
SearchCategory string sent to the target PHP code for the target frame.
    <TD NOWRAP ALIGN=CENTER VALIGN=TOP><A onClick="setImage('FACES');"   
HREF="/imagedb/redirect.php?searchcategory=Faces/Figures"       
TARGET="main"><IMG SRC="/images/facesfigures.jpg"    NAME="FACES"    ALIGN=TOP
WIDTH=140 HEIGHT=30 BORDER=0></A></TD>
    <TD NOWRAP ALIGN=CENTER VALIGN=TOP><A onClick="setImage('SKULLS');"  
HREF="/imagedb/redirect.php?searchcategory=Skulls/EvilFaces"    
TARGET="main"><IMG SRC="/images/skulls.jpg"          NAME="SKULLS"   ALIGN=TOP
WIDTH=140 HEIGHT=30 BORDER=0></A></TD>

These are coded identically to other buttons and having slashes is not unique.
The onClick function simply sets/clears red or green button image in left frame.

Windshields, Dresser and Touring buttons are coded identically to others that do
work.

In the kickstart-1 pages, all frame declaration pages use this header:
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Frameset//EN"
"http://www.w3.org/TR/html4/frameset.dtd">

All target code sent to target frames uses this header:
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"
"http://www.w3.org/TR/1999/REC-html401-19991224/loose.dtd">

I read other bug descriptions where people have suggested things are not
enclosed in double quotes like they should be.  I've tried versions where there
are no things quoted or all things quoted and behavior is same.

I worried that calling the "navbar" as "navbar" and the main window frame as
"main" and the footer frame as "foot" may have been interfering like the word
"face" does in the tattoo site, but I renamed them to navWin, dspWin and tailWin
and the exact same errors result.

IE, Netscape and any other browser I have access to, works.  Maxthon performs
them correctly, then complains it can't find "self", but I do not use "_self" or
 "self" anywhere in my code.

Comment 1

[Giles Robertson]

Windows XP SP2/Firefox 1.0PR:

All cases work as expected, including when run under Windows 2000 compatibility
mode.

Either, therefore:
1) It's a vanilla Win 2000 issue
2) It's crept in between 1.0PR & 1.0RC1

Comment 2

[rfb]

(In reply to comment #1)
> Windows XP SP2/Firefox 1.0PR:
> 
> All cases work as expected, including when run under Windows 2000 compatibility
> mode.
> 
> Either, therefore:
> 1) It's a vanilla Win 2000 issue
> 2) It's crept in between 1.0PR & 1.0RC1

You cannot tell me that it works as expected.  While retesting myself, I noticed
I had switched two pages around and it was impossible to work as described until
I had fixed those pages a short while ago.  The same behavior of opening in the
wrong frame exists, but the contents of index and windshields were switched.

You've got to be kidding.  I've tried it on three different machines, and this
last time using the latest download without extensions and it does NOT work as
expected, unless you expect all your buttons to show target code in their own panel.

I delayed quite a while before deciding to put in a bug report, trying it in
many instances just to make sure I wasn't reporting a random fluke. I wouldn't
bother to take to time to go through tearing apart each and every line of code
in several websites investigating the problem and trying to get around it if it
wasn't a PROBLEM.

That is why I put in a separate but report from the others that keep ending in
this "works for me" ****.  It doesn't work for me and it will not work for
thousands of others.  If it is a win2K issue, then why don' t you try it on a
win2K box before you shoot your "its nothing" into the works?

Go to the bug entry page and search on "frame" and you will see many others have
had the exact same and similar issues.  I only put the report in to aid in
clearing something up, not to get another negative remark from someone in
England.  Every time I have put in a report on any tool I have received back a
negative, BS remark from some "know it all" who has a .uk at the end of their
email address.  What makes you think that just because it does not work in your
one platform, one time test that you should issue some statement indicating that
I'm just "high" and therefore their is no problem?

I've run call tracking/defect tracking systems for AMD, that's Advanced Micro
Devices to you non-silicon valley people, and I've had bugs and problems come in
that didn't happen on my personal station.  That is why I ran a test lab of some
50 systems with all the various O/Ss and installation configurations that we
supported and ran a blanket test across many machines.  Sometimes it was a
memory problem, sometimes other installed software interfered, sometimes it was
operator input and sometimes it was BAD CODE.  We never, ever simply dropped a
report within a couple of hour of hearing it without giving a lot of attention
and investigation into the problem.  When I say that I've run it on three
different machines with different extensions installed, and run in on a brand
new install without extensions in it, and the many bug reports in the system
about similar frame related problems, I would not respect anyone who responded
to my report.

If you people keep flaming and ignoring people who report problems, then you may
 end up keeping others from using the tool because of your ignorance in problem
testing and your arrogance that their couldn't be anything wrong.

So Giles, unless you can respond with other facts than "it works for me", don't
**** up my report so that someone who can test in multiple platforms and
versions can verify that it never, ever, ever happens, or it is something that
only happens with some installs and they are trying to figure out why.

Comment 3

[rfb]

After attempting to identify/fix the problem for too many days, I installed FF
on my kid's XP system and found the code was working there, as Robinson indicated.
I then began a cycle of completely removing Firefox from my system, reinstalling
DayXXX from a backup, and checked to see if the problem existed.

As far as I can tell, it has been there since I started making daily backups in
September (... 2004).  

I then completely removed every single piece of mozilla software from my system,
installed my latest "pretty good" FF backup which used an early .10 (10/1 or
10/4),went through the .10 to 1.0RC2 upgrade path again, updated all my
extensions that had been marked non-compilant and I still had the same problem!!

Again, removed all software. Put in plain 11/4/2004 1.0RC2 and it worked
correctly.  I have been re-installing extensions in two's and three's in an
attempt to determine which one might be the culprit.  Now the problem is, I have
re-installed most of the extensions that I want, leaving out some that I found
were redundant or I didn't use, or were ones that always spewed unknown object
errors into the Javascript Console.  The problem has not reoccurred.

Therefore, I have changed the severity to minor, since a complete rebuild:
throwing out both profile and program directories and removing all old,
redundant paths and garbage found in the windows registry, does "fix" the
problem.  However, I do not think the status should be changed from
unconfirmed/uncorroborated to any of the 4 in the resolved picklist because they
do not show what the proper state is: FISHY.  There is a problem.  I
accidentally created it on THREE different machines by doing the normal
uninstall program/install another nightly or remove/install|copyover extension
with new version style of keeping my system up-to-date with the nightly builds.
If it was only one bad system, then I would change the status to FIXED or
NOTABUG or whatever, but it is repeatable by the fact that I managed to do it
several times.  Just because no one else has a problem, doesn't mean that my set
of files didn't end up performing as I have stated.  Remember the old "If you
put a million monkeys in a room full of typewriters for a million years then one
of them will end up typing a Shakespear play" or something like that.  What is
the count of the number of people or actually, number of times FF has been
installed onto systems, such that NO ONE should ever respond with "that couldn't
happen" or some other such incredulous remark to any bug report ever entered.
**** does happen, for many reasons (possibly I failed to remove something, I
might have installed conflicting items, improperly initialized scripts (so many
javascript errors due to unknown search bars, unknown gBrowser, etc where the JS
code crashes before finishing what its supposed to do), I even had a couple of
powerfailures while I was working which may have hosed a disk file.

I have a few more extensions to install on the system I'm using now, and then I
will copy this profile to one of the other machines to see if it fixes it, and I
will do a similar total cleanout/reinstall fresh on the third system, but
install all the extensions in a different order/different groups to see if I can
determine any other faults.  I think I should send you a copy of my saved faulty
system (the profile and program directories) that someone may copy into their
system to see if it is my platform, registry, O/S or whatever that is the
problem, or perhaps they can find something hokey in an RDF file that is F'ing
something up.  My home system, just before the total flushout would not even
open a second window, even clicking the start icon multiple times; it would not
exit via the FILE:EXIT menu (in order to save session); if killed by the
top-right corner X, it would leave something with 100MB fingerprint running;
save image, save link rarely failed 25%; open in new tab only worked with some
types of links; while perusing Mozilla forums, no history was kept; and a host
of other problems from upgrading from a 10/1 to a 10/28 version of the core
software.  I probably still have a backup that was from 10/15, too.

So, if this bug gets updated to state: please add zipped backups from before
(bad-with the bug) and after (clean install-no bug) for comparison by someone
with a win2k box who is interested in checking this out......  I will comply.
On the other hand, if this bug gets updated to one of the "fixed" statuses and
says: not a bug, just an F'ing newbie;....... or something along that line, then
I will drop the matter and remove myself from the CC list.

Comment 4

[Shadowlady61]

I have no idea what is going on but firefox will not open for use on my 
computer even after reinstalling..any advice??

Comment 5

[Samuel Sidler (old account; do not CC)]

Based on comment 3 (and completely ignoring comment 4), this was an issue with an extension (or two). I'm closing this bug as WORKSFORME since it did, indeed, work for the reporter.

Reporter, if you're able to reproduce this using Firefox 2.0.0.3 and a fresh profile, please file a new bug with detailed steps.

http://kb.mozillazine.org/Profile_Manager