316092 - Pre-parsed/Partially pre-rendered cache

Status: RESOLVED INVALID

(Core :: Layout, enhancement)

Description

[John R. Moser]

User-Agent:       Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8) Gecko/20051107 Firefox/1.5
Build Identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8) Gecko/20051107 Firefox/1.5

Bug #316089 explains the idea of a pre-parsed representation of Web pages in memory without utilizing full pre-rendering.  This is proported to be a potential way to speed up rendering without a full pre-rendered cache in memory.  This bug expands on the idea for on-disk caching.

An internal set of structures forming a tree representing the document in fair detail is good for internal use; but for storage on disk, this is overly complex, even though it's possible.  For disk storage, the internal tree could be easily rendered into a representative bytecode that retains all of the information about the pre-parsed page.  This bytecode would be designed in such a way that it could be read from cache and parsed more quickly than raw HTML or XML into the original internal tree.

A bytecode representation would follow a predictable flow, and thus would be much easier and faster to parse than raw markup.  It would also be partially pre-rendered, as CSS and HTML format specifiers would already be reduced to their equivalent natural format specifiers (bold, italic, font sizes, etc), and image and text placement would be hinted at already.

The bytecode may be larger or smaller than the corresponding markup; in either case, gzip compression on the bytecode should condense it down easily, as it would be fairly redundant at the byte level and thus would condense easily by the huffman pass.  The redundancy eliminator in gzip compression would also work well on the text strings embedded in the bytecode.

Reproducible: Always

Comment 1

[John R. Moser]

Remarked this to core, as it has to do with Gecko.  Target is the future, though this could be made into a more visible goal if somebody can produce some sort of technical evidence that this would actually be a good way to handle pre-rendering in cache.

Comment 2

[David Baron :dbaron: (⌚️UTC-5, no longer working on Mozilla)]

Given my knowledge of the algorithms used to lay out Web pages and the performance hotspots in those algorithms, I don't think this bug report makes any sense.  I suggest marking invalid, but apparently roc saw it and didn't think so.

Comment 3

[Robert O'Callahan (:roc) (email my personal email if necessary)]

I didn't really read it further after I realized it was in the wrong component.

I wouldn't say it's impossible that something resembling this idea could be useful. But I would advise against anyone spending time working on it.