111036 - Cache DOM and layout in addition to page source

Status: RESOLVED DUPLICATE of bug 84286

(Core :: Layout, enhancement, P4)

Description

[Jonathan Briggs]

I think it would speed up cache hits on large pages quite a bit if the HTML
render and DOM processing didn't need to be done each time.  If all the
information needed to render the document was also saved in the cache then
document processing could start from that point.

I can see that this isn't being done now because when I load a page with a very
long table from cache, I can see the columns adjusting themselves as rows with
unusually wide columns are loaded.  The page takes 10 seconds to load without
cache and still takes 6 seconds to load with cache.  Internet Explorer on
similar hardware (actually, a VMware session) can load the same page from cache
in less than 2 seconds.

Comment 1

[gordon]

Changing component and owner.

Comment 2

[Oliver Klee]

This might be related to bug 33269.

Comment 3

[Boris Zbarsky [:bzbarsky]]

This seems like a footprint nightmare....

Note that we do not cache HTML in the memory cache (just in the disk cache) so
this could be a large part of the difference you see... (yes, there is a bug on
this, no I don't have the number offhand).

Comment 4

[rubydoo123]

adding the helpwanted keyword to see if someone in mozilla could jump in and 
help on this one, otherwise it may take some time for the assigned engineer to 
review this request.

Comment 5

[Matthew T (active 1999-2002)]

This would be a duplicate of bug 38486, except this bug refers to the cache in
general whereas 38486 is just about the memory cache.

Comment 6

[Marc Attinasi]

I strongly recommend that we NOT try to get this change made for Mozilla 1.0

Aside from the fact that we ar lacking resources to do it, it would potentially
destabilize Gecko, and would have potential memory footprint issues that we have
not even begun to explore yet.

By what criterion was this marked Mozilla1.0+ anyway?  It is currently marked
'future, enhancement, P4'.

Comment 7

[Caspy7]

This bug sounds very similar to Bug 84286, should one of these get duped?

Comment 8

[Asa Dotzler [:asa]]

not gonna happen for 1.0

Comment 9

[Christopher Hoess (gone)]

*** This bug has been marked as a duplicate of 84286 ***