Text 20, 99 rader
Skriven 2006-07-31 07:28:32 av Mike Luther (1:117/3001.0)
     Kommentar till text 19 av Bat Lang (1:382/61.0)
Ärende: eCS 1.2 + Mozilla
=========================
Hi Bat ..

 BL> Is there any such thing as 'teaching' Mozilla how to 'continue' an
 BL> aborted download? All I seem to get when retrying a dnld that has
 BL> stopped for any reason is a two choice window that asks if you wish to
 BL> continue (after it notes the same file name aready exists) 'Yes  No'.
 BL> If you say No, that's the end of it. If you say Yes, it automatically
 BL> overwrites the 'stubb' starting again from the beginning. !@#$%^
 BL> If there is some way to get Mozilla to "continue" where it left off, it
 BL> has escaped me.  Thanks for any reply/solution.
 BL> Good Modeming!  /\oo/\

Well .. answer number one ...  but read on after it, smiley ..

As for Moz itself, I've was told over and over *NO* and further that under no
circumstances will they change things so that the datestamp of the file is left
as it is on the original file.  Every time I asked for this I'm told to use a
separate download file utility which has that capability.

But answer number two is, AHA!  Are you running Mozilla 1.73 or the
evolutionary brainchild of it, Seamonkey?  Monkey Sea - monkey do! As is he
doo.

I have a broadband connection here over which almost all of my downloads move. 
The download procedure of the new Seamonkey is interesting.  When you go to
download a large file, it opens a download manager, just like the old one did. 
But at this time, as far as I can tell, With Seamonkey lately 1.0.2/3 at least
while you are deciding what directory that file will be put into, it is
actually 'downloading' some sort of a temporary slush portion for that file. 
That 'slush' portion is going into the cache directory if my thoughts are
correct.

I know that because if I take a long time to choose that target directory on a
larger file, then by the time I finalize the decision on where to place it,
here on broadband connections, the progress marker bar on the download manager
will be way toward completion!  But until today I never played with the broken
download bit.  I just did with Peter Weilberger's new unofficial version of
1.0.3.

Aha!  Same story about fuzzing the directory decision.  The marker is already
much of the way there on the 15MB transfer when the decision is made.  Further,
in the decided upon directory for the download, if you look at that directory
during the download you'll find the correct name file chunk for it as well as a
sub-extent '.part' now.  Such as the name 'myfile.zip.part' there.  Further, if
you bust the download with the Download Manager, or otherwise, even shutting
down Seamonkey 1.0.2/3,you will find that the partial file is still there.

Then if you connect back to that site and try again to download that same file,
SUPRISE!  You'll see the download progress bar go quickly chunk, chunk, chunk
to the as earlier download percentage.  And then you'll see the download resume
where it left off.

Plus, just like ZMODEM transfers, if you try to download it twice, even by
fudging it and renaming the old 'in-directory' previously gotten file, you will
see it rapidly whomp to the right and you'll get the 'new' download almost
instantly.  Duhhhhhh whatziz?

Answer.  One of the Moz/Seamonkey tactics that may be not so nice, is that, as
I've been told, the program is actually downloading a separate 'dupe' copy of
the file into the CACHE of all things.  It is only moved to your desired
directory when the download is finalized.  Then, of all things, there will
still be an original name duplicate of that file in the cache as long as there
is space in it to hold it.  Yes, it is removed from cache in due course when
newer things need space there or you manually flush it.  But until that time,
since the original download name slush file is in the cache, the game of 'get
me another' is circumvented by Seamonkey's current logic that 'I've already got
it and the CRC or whatever is the same!'  And .. in the cache, the file is not
'known' by its formal directory name.  It is only in there by an assigned
number which SM is able to hash to somehow for the rest of whatever it needs
for ID purposes as to what is there in cache.

So at least so seen here today in research for you on broadband.  If you simply
move to Seamonkey 1,0.2 or 1.0.3+ latest, you'll accomplish what you want.  But
be careful.  For the first time I've got an issue with Peter Weilberger's
enhanced unofficial version of the new 1.0.3+.  It will 'install' over his
wonderful 1.0.2 last version.  But it never gets past the initial screen splash
announcer which I have to kill with KILLFEATURE or something.  However Mike
Kaply's official 1.0.3 release overinstalls just fine and is working.  Watch
out though, in theory there is a major error in 1.0.3 that is to result in
1.0.4 real soon now.  And I don't know it Mike Kaply's official 1.0.3 has the
needed fix and there will be no 1.0.4 for OS/2 or not.  Peter's website says
the needed fix is already in his 1.0.3+ which is sadly busted here for me.


--> Sleep well; OS/2's still awake! ;)

Mike @ 1:117/3001






--> Sleep well; OS/2's still awake! ;)

Mike @ 1:117/3001

--- Maximus/2 3.01
 * Origin: Ziplog Public Port (1:117/3001)