Refactor getPage (in the worker), and attempt to use the Linearization dictionary to lookup the first Page
#9931
Conversation
|
/botio-windows test |
…ent` instead Addresses an existing TODO, and avoids having to pass in a `pageFactory` when creating `Catalog` instances.
…irst Page Since PDF.js already supports range requests and streaming, not to mention chunked rendering, attempting to use the `Linearization` dictionary in `PDFDocument.getPage` probably isn't going to improve performance in any noticeable way. Nonetheless, when `Linearization` data is available, it will allow looking up the first Page *directly* without having to descend into the `Pages` tree to find the correct object.
Snuffleupagus
force-pushed
the
refactor-getPage
branch
from
915d26f
to
ec3728b
Compare
|
/botio-linux preview |
From: Bot.io (Linux m4)ReceivedCommand cmd_preview from @timvandermeij received. Current queue size: 0 Live output at: http://54.67.70.0:8877/77c2813403c7e04/output.txt |
From: Bot.io (Linux m4)SuccessFull output at http://54.67.70.0:8877/77c2813403c7e04/output.txt Total script time: 2.85 mins Published |
|
/botio test |
From: Bot.io (Linux m4)ReceivedCommand cmd_test from @timvandermeij received. Current queue size: 0 Live output at: http://54.67.70.0:8877/f773d67175a8472/output.txt |
From: Bot.io (Windows)ReceivedCommand cmd_test from @timvandermeij received. Current queue size: 0 Live output at: http://54.215.176.217:8877/7ef07cc460ed5a7/output.txt |
From: Bot.io (Linux m4)SuccessFull output at http://54.67.70.0:8877/f773d67175a8472/output.txt Total script time: 19.49 mins
|
From: Bot.io (Windows)SuccessFull output at http://54.215.176.217:8877/7ef07cc460ed5a7/output.txt Total script time: 26.97 mins
|
|
Good to have this; thanks! |
As expected, using the
Linearizationdata doesn't appear to have had any (noticeable) performance impact. This can probably be attributed toCatalog.getPageDictonly fetching the required nodes, and not the entirePagestree, in combination with caching of already resolved nodes (inpageKidsCountCache).Edit: To clarify the above, for Linearized files this patch does result in (at most) a handful of fewer invocations of
this codesince the first Page can be accessed directly.So while the amount of data being loaded when fetching the first Page is reduced, the difference is really tiny in practice. Hence it seems that unless the server/connection is really slow, the difference would most likely not be seen/felt. (Also, keeping in mind that the default viewer will pre-render the next/previous page.)
Fixes #9716.