Firefox hangs/crashes when trying to store copy of file / linking file

__ls
Hey there,

I am using Firefox 3.0.6 and Zotero 1.5b2. Everytime when I try to store a copy of a file (as in Attachments Tab --> add) Firefox seems to hang, and won't respond to any input anymore. Most bizzare, the action itself, i.e. adding the file seems succesful, cause if I kill firefox via command line and restart it, the file is added as an attachment. This happens deterministically, and I am not really sure why, since adding files worked fine just yesterday.

if extensions.zotero.debug.log is set to true, this is the last output befora nothing happens anymore:

zotero(3): Running pdfinfo "$HOME/.mozilla/firefox/t3nhz7hx.default/zotero/storage/HA2DQQRT/bla.pdf" "$HOME/.mozilla/firefox/t3nhz7hx.default/zotero/storage/HA2DQQRT/.zotero-ft-info"

The bahvior is pretty much the same if I try to link to files instead of storing copies (but I didn't check the console output for this one).

Any ideas what this could be due to?
  • __ls
    I'm on gentoo linux btw, if that adds any information.
  • noksagt
    Full-text indexing of files can take a while, but there shouldn't be any reason to kill firefox.

    You can also check that the pdfinfo and the pdftotext that are in your zotero directory are able to run to completion. I'd check with a very short PDF, both from inside zotero & by invoking those utilities from your command line.
  • __ls
    Hey there,

    yeah I've tried it offline, this time with a pdf containing just "foo" ;)
    pdfinfo and pdftotext works just fine, also within the zotero directory, I tested it on some of the pdfs under storage/.

    But you might have found a valid point there, I guess it might be due to indexing that ff crashes (or hangs longer than 15min after which i just lose patience).

    attaching foo.txt works
    attaching foo.pdf does not

    is there some way to disable full text indexing (i don't use it anyway) without deinstalling those pdf utilites? I already set max no of page to index to 0, but that apparantly didn't help. Am I overlooking sth really simple here?

    Thanks so far
  • noksagt
    Did you test the versions of the utilities that are installed in your Zotero directory (and not those that are installed system-wide)?

    If you aren't indexing PDFs, I don't know why you would want to keep the Zotero versions installed....
  • __ls
    hey,

    yeah you were right, was using the systemwide utilities (0.10.x instead of 3.02 as zotero did) .. but still -- if i wanted to use those features it should work, shouldn't it? :)

    and maybe one could also have a look at why those utilites are invoked at all if one sets the to-be-indexed max to 0 pages?

    thanks though ...
  • noksagt
    was using the systemwide utilities (0.10.x instead of 3.02 as zotero did) .. but still -- if i wanted to use those features it should work, shouldn't it? :)
    Your system-wide versions won't work with zotero, as the Zotero devs have made slight modifications to xpdftools to make them emit files (there are suggested ways to avoid having to do this, but I think it is low priority; the zotero-compiled versions do work for most people. You should also be able to get the zotero-patched version to work for you. Depending on what you've done with your gentoo box, you may need to either roll your own binaries of these versions or add on some support libraries to allow the zotero-compiled binaries to run.
    and maybe one could also have a look at why those utilites are invoked at all if one sets the to-be-indexed max to 0 pages?
    I haven't actually looked at the code, but I can easily see this: the xpd utilities work fairly quickly & the max word/page limits may be intended to keep the zotero database size down. So, perhaps the size checks are only done after the info and fulltext have been obtained.
  • dstillman
    Workaround: http://forums.zotero.org/discussion/4935/crash-when-attaching-document/#Item_18

    The system pdfinfo won't generate the text file Zotero reads to determine the number of pages anyway, so you'd get the same behavior you were already getting (if you were using PDF indexing at all, which it sounds like you weren't).

    If you were having this problem with the system pdfinfo, though, that suggests that the problem isn't in our patched version but rather in either the original code or (probably more likely) in Firefox.
    and maybe one could also have a look at why those utilites are invoked at all if one sets the to-be-indexed max to 0 pages?
    Yeah, it probably shouldn't do this. I'll change it to bail before pdfinfo if pdfMaxPages is 0.

This is an old discussion that has not been active in a long time. Instead of commenting here, you should start a new discussion. If you think the content of this discussion is still relevant, you can link to it from your new discussion.