[zathura] zathura memory usage

[Matthieu CASTET]

Hi,

doing a seach on big pdf (for example arm_arm
https://www.scss.tcd.ie/~waldroj/3d1/arm_arm.pdf), need lots of
memory (1.5 GB with arm_arm).

With other pdf reader like xpdf, there is no such problem.

This make search unusable with big pdf.

Matthieu

[Moritz Lipp]

Hi Matthieu,

Post by Matthieu CASTET
doing a seach on big pdf (for example arm_arm
https://www.scss.tcd.ie/~waldroj/3d1/arm_arm.pdf), need lots of
memory (1.5 GB with arm_arm).

I am not able to reproduce this.

Which version of zathura are you using? And which PDF
backend?

Best regards,
Moritz

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: signature
URL: <http://lists.pwmt.org/archive/zathura/attachments/20140729/5d12a07d/attachment.sig>

[Matthieu CASTET]

Hi Moritz,

Post by Moritz Lipp
Hi Matthieu,

Post by Matthieu CASTET
doing a seach on big pdf (for example arm_arm
https://www.scss.tcd.ie/~waldroj/3d1/arm_arm.pdf), need lots of
memory (1.5 GB with arm_arm).

I am not able to reproduce this.

In fact this pdf use 500MB.
This one (http://liris.cnrs.fr/~mmrissa/lib/exe/fetch.php?media=armv7-a-r-manual.pdf) use 1.5GB.

For that I open the pdf. Type "/" to enter in search mode. Search a word (let's say foo). And then zathura eats cpu and memory.

Post by Moritz Lipp
Which version of zathura are you using? And which PDF
backend?

I am using the debian pakaged version for sid :
https://packages.debian.org/sid/zathura

This is the 0.2.9 version with poppler backend.

Matthieu

[Matthieu CASTET]

Hi,

Nobody is able do reproduce it ?

I tried valgrind, it show no memleack, but lot's of memory allocation.

Matthieu

Le Wed, 30 Jul 2014 14:35:12 +0200,

Post by Matthieu CASTET
Hi Moritz,

Post by Moritz Lipp
Hi Matthieu,

Post by Matthieu CASTET
doing a seach on big pdf (for example arm_arm
https://www.scss.tcd.ie/~waldroj/3d1/arm_arm.pdf), need lots of
memory (1.5 GB with arm_arm).

I am not able to reproduce this.

In fact this pdf use 500MB.
This one (http://liris.cnrs.fr/~mmrissa/lib/exe/fetch.php?media=armv7-a-r-manual.pdf) use 1.5GB.
For that I open the pdf. Type "/" to enter in search mode. Search a word (let's say foo). And then zathura eats cpu and memory.

Post by Moritz Lipp
Which version of zathura are you using? And which PDF
backend?

https://packages.debian.org/sid/zathura
This is the 0.2.9 version with poppler backend.
Matthieu

[Sebastian Ramacher]

Post by Matthieu CASTET
Hi,
Nobody is able do reproduce it ?
I tried valgrind, it show no memleack, but lot's of memory allocation.

Well, that's somewhat expected. We iterate over every page and search
for the string. So of course, we need to load all the pages into memory.

We could be a bit more intelligent and just load the current page,
search for the string and wait for the user to hit n and then continue
searching. But even that wouldn't help if the result is at the other end
of the document.

Does disabling incremental search help a bit? At least it won't search
for every string while you're still entering the search string.

Cheers