zathura - a PDF viewer

eschew · 2010-07-10 16:11:30

neldoreth wrote:

Hello,
Inxsible wrote:
zathura doesn't display this pdf correctly. Here are the screenshots of the same pdf using xpdf and zathura
xpdf shot : http://omploader.org/tNHcwOA
zathura shot : http://omploader.org/tNHcwOQ
Can you please give me a link where I can download this PDF? But I am pretty sure this is a poppler isue (try opening it with evince or epdfview).
Best regards

I'm not the original reporter, but you can download the PDF referenced here: http://www.mwbrooks.com/dvorak/dvorkeys.pdf. epdfview renders the PDF in the same way, so you're right about it likely being a poppler issue. Thanks for all your work on zathura and jumanji!

Inxsible · 2010-07-10 16:31:14

Thanks eschew and neldoreth. I guess then we will just have to live with it.

But if the issue is with poppler, then I will have to keep xpdf around just in case some other pdfs don't show up cleanly in zathura/evince/epdfview.

melon · 2010-07-13 05:22:08

I'm trying to remap the navigation (hjkl) keys for my dvorak keyboard, but I'm having issues getting this to work. For example, if I add "map h navigate_index down" to zathurarc or do ":m h navigate_index down" then press h, zathura crashes with

(zathura:6052): Gtk-CRITICAL **: gtk_container_foreach: assertion `GTK_IS_CONTAINER (container)' failed
zsh: segmentation fault  zathura test.pdf

If I map h to anything else (e.g. zoom in), it works. Am I doing the mapping wrong, or is this a bug?

In either case, thanks for the great program! I'm liking it much more than xpdf/apvlv so far.

neldoreth · 2010-07-13 06:37:43

Hello,

melon wrote:

If I map h to anything else (e.g. zoom in), it works. Am I doing the mapping wrong, or is this a bug?

Thanks for mentioning it - It was a bug and it is fixed in the git version right now - In addition you should add the INDEX mode for those mappings, because you only see the index when you are in it.

Best regards

Last edited by neldoreth (2010-07-13 06:38:04)

melon · 2010-07-13 08:51:52

Thanks! It works perfectly now.

firecat53 · 2010-07-17 00:19:35

Just wondering about a possible memory leak...I was browsing a 103-page pdf (mostly text, some graphic illustrations), doing a lot a scrolling and text searching, and zathura's memory use went up to 49% on my machine over about 10 minutes. Doing the same operations with evince resulted in only 7% memory use. PDF is from here.

Great reader!

Thanks,
Scott

neldoreth · 2010-07-18 18:10:37

Hello,

firecat53 wrote:

Just wondering about a possible memory leak...I was browsing a 103-page pdf (mostly text, some graphic illustrations), doing a lot a scrolling and text searching, and zathura's memory use went up to 49% on my machine over about 10 minutes. Doing the same operations with evince resulted in only 7% memory use. PDF is from here.

Thank you for this - Can you please check out the latest git version and try if this solved it?

Best regards

firecat53 · 2010-07-18 19:17:39

Hmm, interesting....I think it's improved substantially. Comparing to evince with the same document and doing different searches over and over, combined with zooming in and out, both programs use more memory over time, but in different ways. I haven't taken the time to understand the difference here, yet, but I thought you might find it useful. As I keep searching, the VIRT for evince is staying about the same, but %MEM is slowly rising, and with Zathura, VIRT keeps going up fairly quickly, and %MEM goes up too, but much more slowly.

Initally VIRT %MEM
zathura 127M 1.6
evince 314M 5.0

after searching
zathura 4019M 7.3
evince 354M 16.1

Hope this helps....or maybe it's a non-issue!

Scott

Last edited by firecat53 (2010-07-18 19:20:59)

neldoreth · 2010-07-18 20:56:16

Hello,

firecat53 wrote:

Hmm, interesting....I think it's improved substantially.

OK, good.

firecat53 wrote:

Hope this helps....or maybe it's a non-issue!

No, there are still some memory leaks that are caused through the command line completion and search function.

Best regards

neldoreth · 2010-07-27 07:43:02

Hello,

I just wanted to say that pwmt has a new website based on redmine (it is not finished yet) and brings a bug/feature tracker with it that I hope will be used wisely, as well as a wiki that will be used for further documentation.

Best regards

Runiq · 2010-07-27 08:35:00

neldoreth wrote:

Hello,
…new website…
…tracker…
…wiki…
Best regards

Awesome.

neldoreth · 2010-07-29 15:17:34

Hello,

A new version of zathura has been released (0.0.8):

Changelog:
* Improved and fixed completion
* Optional parameters for the print command
* Escape filenames
* Resolved some memory leaks
* Global configuration file
* Specify configuration directory
* Configureable modes
* Smooth scrolling
* Extended statusbar string manipulation function
* Strip executable
* Updated manpage

Download:
Download 0.0.8

Best regards

Sirsurthur · 2010-07-30 08:22:21

Hello and thanks for this very good pdf viewer.

I have only one question : is it or will it be possible in the future to read .ps files with zathura as well ?

Thanks again,
Sirsurthur

bernarcher · 2010-07-30 08:32:15

The new version segfaults here if my zathurarc contains these lines:

map <PageUp> navigate previous
map <PageDown> navigate next

Is this functionality not provided any more?

neldoreth · 2010-07-30 10:21:58

Hello,

A new version of zathura has been released:

0.0.8.1
* Fixes bug in zathura 0.0.8

Best regards

bernarcher · 2010-07-30 12:25:59

Problem fixed.
Thanks for the quick reaction, neldoreth.

rent0n · 2010-07-31 18:14:19

After last zathura upgrade I can no longer print pdf files with:
:print <NAME_OF_PRINTER>

Error is as follows:

lp: Error - unable to access "" - No such file or directory

Other apps, including evince, print just fine.

Thanks!

Stalafin · 2010-08-03 18:41:01

There are a few things I have observed and would like to comment on:

1) There are some PDF documents, where there are two separate numberings - the preface is numbered with Roman letters, while the main part is numbered with Arabic characters. An example would be Kopka's 'A Guide to LaTeX'. Because of this, the number of pages displayed in zathura's buffer does not correspond to the page number as defined in the PDF. So, when I type 20G, I end up on page 20 in the PDF, which is displayed as 32/658 in zathura's buffer. I find this confusing

2) Some sort of prefetching of the previous and the next page would be amazing. I often find myself reading a page in a textbook or a paper, and I want to check an equation on a page above. Since many papers are rather old (so everything is an image, I figure), zathura takes a significant time to render that page. That is annoying, since I have to wait.

3) DjVu support would be amazing. :-)

EDIT: Also another thing I have noticed:

4) When opening a file, Tab will only give me a list of documents available in the current subfolder, if I have _not_yet_autocompleted_ what I wrote. For example
If I type /home/user/ and then press Tab, I will get a list of documents and/or folders. If I, however, choose to expand ~ or /home/us, such that after hitting Tab I will get /home/user/, another press of Tab will not give me a complete list of documents available (pressing another char followed by Tab, however, will give a list of documents starting with that char).

Last edited by Stalafin (2010-08-04 00:55:02)

OpenBSD · 2010-08-09 18:22:27

rent0n wrote:

After last zathura upgrade I can no longer print pdf files with:
:print <NAME_OF_PRINTER>
Error is as follows:
lp: Error - unable to access "" - No such file or directory
Other apps, including evince, print just fine.
Thanks!

+1
same problem

frabjous · 2010-08-11 03:44:48

SyncTeX kludge

I'm giving this interesting looking project a try as part of my "total immersion" strategy for forcing myself to finally get used to vim-keybindings.

I spend pretty much all my time on my computer in one of three places: browser (now using Vimperator), text editor (now gvim) and PDF viewer. This looks promising. But what I mainly use text editing/PDF for is LaTeX-ing, and one thing I miss from something like this as opposed to, e.g., Okular (or the patched evince-gtk in AUR) is direct support for SyncTeX, i.e., the ability to jump directly from a certain spot in the PDF to the matching part of the LaTeX source file and vice versa. It wouldn't be hard to add such a feature; really all that would be needed would be the ability to write the current page and cursor position to an external file, or call an external program with them as arguments. So consider that a feature request.

In the meantime, I'm worked up some very ugly hacks to provide partial support for SyncTeX just by exploiting the fact that the last active page is saved in the ~/.config/zathura/bookmarks file upon exit. I apologize that these are fairly badly done; it's more proof of concept than anything else. (I'm not a programmer; my experience with vim is only about a week only, and my experience with Bash not much older.)

The first script, when invoked from (g)vim, will close Zathura, change its bookmark to the page SyncTeX matches with the current cursor location for gvim, and restart Zathura to that page.

#!/bin/bash
# syncforward.sh full-path-of-.tex-file line column
pathfn="$1"
line=$2
column=$3
# reduce file name to basename
file="$(echo "$pathfn" | sed 's@.*/@@g')"
# change to directory containing file
cd "$(echo "$pathfn" | sed 's@\(.*\)/[^/]*@\1@g')"
# correct mismatch between numbering schemes
line=$((line+1))
column=$((column+1))
# escape slashes in path for sed-ding
filepath=`echo $PWD | sed 's@/@\\\/@g'`
# close Zathura for the output if already running
if wmctrl -x -l | grep -e "Zathura.*${file%.tex}.pdf" ; then 
    wmctrl -c "${file%.tex}.pdf"
    sleep 0.4
fi
# get page number to open using SyncTeX
page=`synctex view -i $line:$column:"$file" -o "${file%.tex}.pdf" | grep -m1 'Page:' | sed 's/Page://'`
page=$((page-1))
# backup bookmarks
cp ~/.config/zathura/bookmarks ~/.config/zathura/bookmarks.old
# Replace saved page number with SyncTeX result in bookmark
sed -i -e "$filepath\/${file%.tex}.pdf/,/^\[/ s/page=.*/page=$page/1"  ~/.config/zathura/bookmarks
# reopen Zathura and wait for it to map
zathura "${file%.tex}.pdf" &
sleep 0.5
# return focus to gvim
wmctrl -a "$file"
exit 0

The second script, when invoked from gvim, will close Zathura, and reopen it, simply so that the bookmark file will be written, then read the bookmark file to see where it was, and then find the line of the source matching approximately 1.5" down and 2" over on that page, and then make gvim jump to that line. This doesn't really work that well, but it at least gets you in the rough ballpark of where you should be most of the time.

#!/bin/bash
# syncreverse.sh full-path-of-PDF-minus-extension
pathfn="$1"
# reduce filename to base, add .pdf
pdffile="$(echo "$pathfn" | sed 's@.*/@@g').pdf"
# change to directory containing files
cd "$(echo "$pathfn" | sed 's@\(.*\)/[^/]*@\1@g')"
# close Zathura; wait for bookmarks file to write; reopen
if wmctrl -x -l | grep -e "Zathura.*$pdffile" ; then 
    wmctrl -c "$pdffile"
    sleep 0.4
    zathura "$pdffile" &
fi
# escape slashes in path for sed-ding
filepath=`echo $PWD | sed 's@/@\\\/@g'`
# read page number open in Zathura from bookmarks file
page=$(sed -n -e "$filepath\/$pdffile/,/^\[/ s/page=\(.*\)/\1/p"  ~/.config/zathura/bookmarks)
# use pdfinfo to get page dimensions
# not currently used but perhaps in future version
# pagewidth=$(pdfinfo -f $page -l $page "$pdffile" | grep -e 'Page.*size:' | sed 's/.*size: \(.*\)x.*/\1/')
# pageheight=$(pdfinfo -f $page -l $page "$pdffile" | grep -e 'Page.*size:' | sed 's/.*size:.*x\(.*\)pts.*/\1/')
# read input file and line number for page from SyncTeX
inputfile=$(synctex edit -o "$page:288:108:$pdffile" | grep 'Input:' | sed 's/Input://')
line=$(synctex edit -o "$page:288:108:$pdffile" | grep 'Line:' | sed 's/Line://')
# if nothing found, use basename for pdf
if [ -z $inputfile ] ; then
    inputfile="${pdffile%.pdf}.tex"
fi
# send command to gvim to open to that line
gvim --remote +$line "$inputfile" &
# wait for it
sleep 0.25
# refocus gvim
wmctrl -a "$inputfile"
exit

And here is the relevant part of my .vimrc for making use these:

function! SyncTexForward()
    let execstr = "silent !syncforward.sh %:p ".line(".")." ".col(".")
    exec execstr
endfunction
nmap <Leader>f :call SyncTexForward()<CR>
nmap <Leader>r :silent !syncreverse.sh %:p:r<CR>

So <leader>f invokes forward search, <leader>r backward search.

(I don't currently have the vim-LaTeX suite or other LaTeX plugin installed; I wanted to learn vim first before tackling one of those. Just be sure to compile the source using --synctex=1 as argument to pdflatex/xelatex.)

The scripts require wmctrl. I'm sure these scripts can be improved. In particular, they ought to be improved to allow for master and sub LaTeX documents, and to do some smarter error checking, etc.

Anyway, because I'm very new at this, I was hoping others might have suggestions along the same lines, or knew of ways of my scripts or processes could be improved. Would I be better off with apvlv or something else?

Last edited by frabjous (2010-08-11 03:46:59)

Jodell · 2010-08-12 01:40:22

I'm either a noob or this a bug

I've just installed zathura and I can't get it to read .pdf files. I use the shortcut 'o' followed by the path to the file but it says 'File does not exist.' The path is correct but zathura does not see it. Upon retrying zathura just freezes. I've tried both the stable and git versions and the problem occurs in both.

Last edited by Jodell (2010-08-12 04:50:53)

archguest · 2010-08-14 13:33:12

I've just tried 0.0.8.1-1 from community repository and was quite disappointed by a few things:

- the main window black with small green font in the status bar. would it not be better if by default it looked like any other gtk program?

- it does not display any sort of welcome message (like emacs does) and it could at least tell that it understands vi keybindings. it think apvlv does this quite well.

- when i tried to :open, tab completion did not work on files. i could tab complete /home/lex/books/ but not within the folder *unless* i input the first letter of a filename.

/me is still searching for a evince replacement

Ogion · 2010-08-14 13:53:01

"man zathura" for keybindings, and i always open files by giving the filenames as commandline arguments, but maybe that's because i'm much in the terminal anyway.
Also, it is not supposed to look like any other gtk program. It's supposed to look like the "vim-like" row of programs, like vimperator (and ofc vim itself).

Ogion

archguest · 2010-08-14 14:34:46

Ogion wrote:

"man zathura" for keybindings, and i always open files by giving the filenames as commandline arguments, but maybe that's because i'm much in the terminal anyway.
Also, it is not supposed to look like any other gtk program. It's supposed to look like the "vim-like" row of programs, like vimperator (and ofc vim itself).
Ogion

"man zathura" is there but I was talking about something else. It was a design suggestion and some vi implementation do display some helpful information on startup. It is not major issue of course.

I use vimperator and it respects default GTK settings to a high degree. For example, in :help section it does not use green on black colour scheme. BTW I have yet to see vi implementation which by default has green font for its interface. This was not a complaint about the colours (which are possible to change) but about a GTK application sticking out from other GTK applications for no good reason.

In regards to tab completion, if there is one for directories, and even for files once I type in a first letter of a filename, I would (reasonably) expect it to work without typing the first letter in a filename.

If all this is by design, I can perfectly understand it, after it's all about personal choice and move on and check out some other PDF viewer

Ogion · 2010-08-14 14:58:59

Right, sorry didn't want to say it like that.
You're right about green on black. I just happen to not notice, but then i don't read pdfs too often, and most of the time i hide the bars (ctrl-n and m).

Ogion

Arch Linux

#351 2010-07-10 16:11:30

Re: zathura - a PDF viewer

#352 2010-07-10 16:31:14

Re: zathura - a PDF viewer

#353 2010-07-13 05:22:08

Re: zathura - a PDF viewer

#354 2010-07-13 06:37:43

Re: zathura - a PDF viewer

#355 2010-07-13 08:51:52

Re: zathura - a PDF viewer

#356 2010-07-17 00:19:35

Re: zathura - a PDF viewer

#357 2010-07-18 18:10:37

Re: zathura - a PDF viewer

#358 2010-07-18 19:17:39

Re: zathura - a PDF viewer

#359 2010-07-18 20:56:16

Re: zathura - a PDF viewer

#360 2010-07-27 07:43:02

Re: zathura - a PDF viewer

#361 2010-07-27 08:35:00

Re: zathura - a PDF viewer

#362 2010-07-29 15:17:34

Re: zathura - a PDF viewer

#363 2010-07-30 08:22:21

Re: zathura - a PDF viewer

#364 2010-07-30 08:32:15

Re: zathura - a PDF viewer

#365 2010-07-30 10:21:58

Re: zathura - a PDF viewer

#366 2010-07-30 12:25:59

Re: zathura - a PDF viewer

#367 2010-07-31 18:14:19

Re: zathura - a PDF viewer

#368 2010-08-03 18:41:01

Re: zathura - a PDF viewer

#369 2010-08-09 18:22:27

Re: zathura - a PDF viewer

#370 2010-08-11 03:44:48

Re: zathura - a PDF viewer

#371 2010-08-12 01:40:22

Re: zathura - a PDF viewer

#372 2010-08-14 13:33:12

Re: zathura - a PDF viewer

#373 2010-08-14 13:53:01

Re: zathura - a PDF viewer

#374 2010-08-14 14:34:46

Re: zathura - a PDF viewer

#375 2010-08-14 14:58:59

Re: zathura - a PDF viewer

Board footer