Re: NS 7.2 Duplicating Bookmarks File.

Moz Champion (Dan) wrote:
> Rufus wrote:
> 
>> Moz Champion (Dan) wrote:
>>
>>> Rufus wrote:
>>>
>>>> Moz Champion (Dan) wrote:
>>>>
>>>>> Rufus wrote:
>>>>>
>>>>>> Moz Champion (Dan) wrote:
>>>>>>
>>>>>>> Rufus wrote:
>>>>>>>
>>>>>>>> Moz Champion (Dan) wrote:
>>>>>>>>
>>>>>>>>> Rufus wrote:
>>>>>>>>>
>>>>>>>>>> For some reason, NS 7.2 (under OS X - 10.4.3 to be specific) 
>>>>>>>>>> is duplicating my Bookmarks file on Quit.  Any suggestions on 
>>>>>>>>>> how to make it stop?  I've tried fresh installs and checking 
>>>>>>>>>> Permissions on the file.
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Try this.
>>>>>>>>> Netscape OFF
>>>>>>>>> Remove the bookmarks.html file from the profile (and any 
>>>>>>>>> duplicates)
>>>>>>>>> store it on the desktop
>>>>>>>>> Restart Netscape
>>>>>>>>> your bookmark file should be empty (or revert to default list)
>>>>>>>>> add one bookmark (just one)
>>>>>>>>> Quit Netscape.
>>>>>>>>>
>>>>>>>>> If the bookmarks file was NOT duplicated on Quit, then
>>>>>>>>> with Netscape OFF
>>>>>>>>>      put the bookmarks.html file back into the profile, 
>>>>>>>>> allowing it to overwrite (replace) the 'newer' one already there
>>>>>>>>> Restare and quit again, let us know the results
>>>>>>>>>
>>>>>>>>> If it was duplicated, please report back
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> That didn't work - in fact, now I'm seeing something new.  It 
>>>>>>>> looks as if a duplicate bookmarks file is being created and 
>>>>>>>> deleted (the file name bookmarks-1.html flashes up and is then 
>>>>>>>> removed)...I saw this on a startup, and also on a few exits.  I 
>>>>>>>> also saw it during navigation to new sites.  On shut down, the 
>>>>>>>> duplicate file sometimes was deleted, and somtimes was not.  It 
>>>>>>>> didn't seem consistant in the few attempt I tried this morning.
>>>>>>>>
>>>>>>>> I'll also add that I looked at a NS profile on a friend's 
>>>>>>>> machine and found the same (numerous) duplication of the 
>>>>>>>> Bookmaks file.  He hadn't noticed because he mostly uses Safari.
>>>>>>>>
>>>>>>>
>>>>>>> What do you mean 'didnt work'?
>>>>>>>
>>>>>>> I asked you to do something and then report back on the results.
>>>>>>>
>>>>>>> Remove the bookmarks file from the profile with Netscape OFF
>>>>>>> (and remove any duplicates as well)
>>>>>>>
>>>>>>> Restart Netscape and quit
>>>>>>> did the duplicate bookmark file get generated?
>>>>>>>
>>>>>>> Restart Netsceap and add one bookmark, then quit
>>>>>>> did the duplicate bookmark file get generated?
>>>>>>>
>>>>>>> This is not a 'fix' this is a test to determine where the problem 
>>>>>>> lies
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> Yes - I did just what you described.  The problem/behavior still 
>>>>>> persists, with the additional observations I described above - 
>>>>>> that WAS the result of performing your suggestion.
>>>>>>
>>>>>> I've also had this input from another source - it's a Firefox bug 
>>>>>> report, but similar behavior:
>>>>>>
>>>>>> https://bugzilla.mozilla.org/show_bug.cgi?id=294584
>>>>>>
>>>>>> I'll add to this that I already have (and have had) my Mozilla and 
>>>>>> Profile folders included in my Spotlight Privacy list.
>>>>>>
>>>>>
>>>>> Care to let us in on the results of the test?
>>>>>
>>>>> Did or did not the duplicate bookmark file get recreated on quit?
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> Yes - it was recreated on quit.  But not consistantly.  (I ran 
>>>> several tests.)
>>>>
>>>>> With the default bookmark file?
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> Yes.
>>>>
>>>>> With the modified (add one bookmark) file?
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> Yes.
>>>>
>>>>>
>>>>> Answers for each step would be appreciatted.
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> AND - in addition, it was also being duplicated during transitions 
>>>> between web pages, which is something I hadn't seen before.
>>>>
>>>
>>> ah. 'but not consistantly'...
>>>
>>> try to ascertain the changes made that caused the bookmark 
>>> duplication to take place.  If it doesnt occur EVERY time then there 
>>> must be some variable that is contributing to the event, and tracking 
>>> down that variable is the key.
>>
>>
>>
>> ...and that's precisely the problem.  I had changed nothing when all 
>> this started.  The new behavior happened after I tryed your 
>> suggestion, and the duplication after navigation to different web 
>> pages was a new observation.  I'll add that I've since seen the 
>> duplication on startup as well...AND I've also seen it in two 
>> different Profiles under the same install where  the only difference 
>> is the size of the Bookmarks file (number of bookmarks).  I've also 
>> seen it on a fresh, pristine install on a new machine.  So it's really 
>> random.
>>
>>>
>>> A computer program is, in essence, a really 'dumb' thing. It will do 
>>> something over and over again, always the same way, if nothing 
>>> changes. So the fact that the duplication DOESNT occur at times 
>>> indicates that there is some variable in your setup that is 
>>> contributing to the problem.
>>
>>
>>
>> I evaluate operational aircraft system software for a living, and from 
>> my experience with conditional branching in near-realtime assembly 
>> language (and C++) programs - yes, programs can produce seemingly 
>> random and difficult to duplicate behaviors.  But I digress...this is 
>> a different program.
>>
>> Hence tracking down the bug report I cited.  I think it's also 
>> interesting that I've seen it on multiple machines - two of my own, 
>> and one that isn't.  It's definately a known NS-Mozilla bug, but no 
>> one seems to have squashed it yet.
>>
> 
> Something seems amiss here.
> 
> First you report it on one machine, then two, now three?

Yes - once I couldn't get a handle on it in my base set up, I looked at 
another machine; one belonging to a friend.  And as I just took delivery 
of a brand new 15" Powerbook I was also able to observe the same thing 
on a freash install.

> First you report it as occurring on quit, then on browsing, now on startup?

Yes - I am seeing new behavior presumably as a result of performing your 
suggestion.  Previously I had only looked for the behavior on Quit.  I 
took my observations further after performing your suggestion, and made 
additional notes.

> 
> The bug you cite seems to have been identified as a problem with 
> Spotlight, not Mozilla. Has been reported as
> Apple, radar ID 4188208.

Yes - I had seen this previously and thus had set Spotlight not to index 
my Mozilla folder as someone else had suggested.  The bug persists.

> While there are reports of this occuring on OS 10.3 as well as 10.4, the 
> 10.4 seems to have been satisfied by setting Spotlight not to index the 
> Firefox profile. 10.3 issues are not as clear, but seem related to 
> having the file locked or read only, a permissions issue

Yes - I have Repaired Permissions, and am still seeing it on a fresh 
install.  I have even gone so far as to verify and match Permission 
settings on the file manually between the fresh install and my old 
machine.  So if it IS a Permissions issue, then Mozillia is mis-setting 
them at install.  Either way, I can verify that the bug still is stilll 
extant with 10.4.3 - across three independant machines.

> 
> My suggestin was to remove the bookmarks file, quit and restart, where 
> did the browsing bit come from? 

It was something I decided to try once I noted that the file was being 
inconsistantly duplcated but did not persist (the "flashing" I noted) 
after several start/quit attempts.  I noted that on occasion, I would 
get multiple duplication within a single session, sometimes not.  So I 
persued the cause - what I found was that navigation between web pages 
sometimes caused the Bookmarks file to be duplicated, along with an 
additional duplication at shutdown.  BUT - the duplication appears to be 
"random" - it can depend on the length of the session, or the number of 
pages visited.  I tried the same page several times and it did not 
always occur.  I'm stumped.

And as I stated, it was a means of
> testing. By introducing unknown variables into the test sequence, you 
> have diminished any value of the test.

I was trying to both make all of the variables known, and to further 
quantify the extent of the behavior.  I added a new test and reported on 
it when I saw a new behavior.  Then I attempted to repeat all tests on 
another device.  That's what I do, and have done, professionally.  For 
the last 15-odd years.

> 
> If it was a NS or Mozilla bug, then why dont ALL users of OS X see it?
> It doesnt occur on 10.2, and turning off Spotlight indexing on 10.4 
> quashes it there, the jury is still out on 10.3 it seems.

Actually, I can personally report that it DOES occur in 10.2.  That was 
where I first saw it, and I still maintain my original 10.2 boot drive 
and am still seeing it there if I boot into my 10.2.8 install.  After 
looking at my friend's machine (10.4.3), I'm convinced that this is just 
under reported by users - how many people really look into the contents 
of their Mozilla folder and watchit while they are surfing around 
anyway?  My friend certainly didn't, and I was certainly surprized when 
I verified the behavior on his machine.

> 
> Turning off Spotlight indexing of the Firefox profile would seem to be 
> the answer in your case.

No - I can verify that it is not the answer, because that has been how I 
have been configured for ALL of these tests on my base machine - ever 
since I first saw the bug under 10.2.  That was the first "fix" I ever 
tried.

-- 
      - Rufus