Snapshot storage files with @22 extensions

Slackenerny
Just wondering, each time I take a snapshot of a webpage, the storage folder ends up full of files with @22 extensions, sometimes over a hundred. I've been deleting them, and the snapshots look mostly the same - is this ok? Or am i deleting useful files?
  • adamsmith
    that's OK - they're likely something the webpage does - if it looks OK after deleting they're almost certainly not useful.
  • Slackenerny
    Thanks! That's a relief - deleting them makes manually syncing a library to a portable drive much quicker - it can reduce a big library by thousands of files!
  • baldwint
    I see this behavior too - only for Nature journals. I think there is something wrong with the translator.

    The files have names like "01.gif@22" and are all the same size. They aren't even .gif files - if you look inside, it's an html file of the nature.com homepage.

    So, instead of saving .gif files for the snapshot, the translator is saving ~200 copies of the same web page at ~150KB a pop. Obviously this is not right, and it eats disk space.

    Is this bug known?
  • dstillman
    Provide an example URL. (Nothing to do with the translator, though—it'd be an issue with the third-party snapshot saving code we use.)
  • baldwint
    I just reproduced the problem from my university computer.

    URL (e.g.): http://www.nature.com/nature/journal/v415/n6867/full/415039a.html
    Zotero: 2.1.10
    Firefox: 3.6.13
    OS: Xubuntu Linux 10.10
  • baldwint
    I reproduced it from my home computer.

    URL: same
    Zotero: same
    Firefox: 5.0.1
    OS: Mac OS X 10.6.8

    Curious to hear if anyone else can see this.

    1) navigate to the nature URL above
    2) save page to Zotero
    3) find item in library, click + to expand
    4) right click on Nature Snapshot and choose "show file"

    See if there's a bunch of @22 files in that directory.
  • dstillman
    The snapshot code would work the same on all computers. No real mystery here—just some bug in how the snapshot code handles the site code. We'll look into it. Thanks.
  • adamsmith
    (just as an aside - you're using Firefox versions that no longer receive critical security update - both of your computers have browsers that haven't been updated for 2years or so. Not just will you experience bugs - though as Dan says this one has nothing to do with the browser - you're also exposing yourself to exploits and known vulnerabilities.)
  • dstillman
    Fixed on the 3.0 branch.

    You can likely delete any @22 files you find without any dire consequences (and we might want to clean them up automatically at some point). I believe they'd all be invalid.

    If you're using syncing, after deleting the files you can resave the main HTML file (or, in OS X or Linux, use the touch command) such that the modification time changes, and Zotero will resync the pruned version.
  • baldwint
    Awesome, Dan.

    I believe you have just set a bug-fix speed record. Well done.
  • mtwulf-privat
    I also have thousands of files like this (3620 to be exact). Mostly invalid graphic files like png@22, gif@22, jpg@22 or jpeg@22. Consuming 80 MB alltogether.

    Because these are located in many different folders, manual touch is not very suitable.

    Some sample content (vid-collapsed.png@22)
    404 Not Found
    ------------------------------------
    nginx/0.6.36
    or (gvy8rtidnbk.gif@22)
    /*bcs*/


    So automatic cleaning would be much better.

    Besides, I just noticed, that there are also valid png and gif files among old snapshots. So I'm no longer sure, what I will miss after cleanup.

    Example of a useful file (created by a zotero-snapshot in october 2010):

    "key_visual8.jpg@22" is a true .jpg-file
    from URL http://www.paarkommunikation.pflugweb.de/2.html

    If a create a snapshot with the present version of zotero (3.0.3), no image is displayed, when I show the snapshot by double-clicking it inside the library (URL:zotero://attachment/723/).

    The images are stored though, and are being displayed correctly, if I open the stored page as a file. (URL:file:///C:/Zotero-Data/storage/WF2KGXZR/2.html)

    In this case this is not a problem, but certainly, there is a bug in the snapshot code. I don't know about the bugs on the webpage itself, though.

    Apart from this little bug, zotero works great.

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

Sign In or Register to comment.