881974 - pdf.js uses too much memory

Status: NEW

(Firefox :: PDF Viewer, defect, P3)

Description

[Nicholas Nethercote [:njn] (On Long Service Leave)]

pdf.js uses too much memory, at least in some cases.  This is bad, because PDFs are common, and the introduction of pdf.js made it hard to open PDFs in any other way.

This is a meta-bug for tracking specific problems.  I'll put the MemShrink tag on this bug, not on the individual blocking bugs.

Comment 2

[Florian Bender]

Is it possible to hibernate pdf.js (or parts of it) when not in use (i. e. in a background tab)? That will at least help reduce excessive memory usage in some use cases (e. g. when someone opens multiple PDFs at once). 

Would it be possible to implement such a hibernation mode?

Comment 4

[Liz Henry (:lizzard Please n-i to RyanVM, jcristau, or pascal)]

Here's a user on the support forums also reporting this issue: 
https://support.mozilla.org/en-US/questions/972366?esab=a&s=&r=0&as=s

Comment 5

[ryan.brothers]

I am also seeing extremely high memory usage of 1GB or more when viewing PDF files of scanned images.  See further details at https://github.com/mozilla/pdf.js/issues/3630.  Please let me know if I can provide more information to help resolve this issue.

Comment 6

[Nicholas Nethercote [:njn] (On Long Service Leave)]

Things are much better now, and I've recently landed a few more patches to the master pdf.js repo that haven't yet been merged into Firefox. I think this can be safely downgraded to a MemShrink:P2. If anyone has example documents that still cause high memory consumption please let me know.

Comment 7

[Nicholas Nethercote [:njn] (On Long Service Leave)]

I made some additional improvements, described here: https://blog.mozilla.org/nnethercote/2014/06/16/an-even-slimmer-pdf-js/

Comment 8

[mirh]

Should also depend on bug 1005005, bug 1165624 and bug 1262736

Comment 9

[mirh]

Bug 1182398, bug 1436734, bug 1536420 too