too long filenames

baris

I use two machines (work and mobile) sync their copy of Zotero storage with a third machine (home). I prefer to use unison for this.

I am having difficulty with transferring Zotero storage
from home: ext4 (rw,nosuid,nodev,uhelper=udisks)
to mobile: ext4 (rw)

However,
from work: ext4 (rw)
to home: ext4 (rw,nosuid,nodev,uhelper=udisks)
seem to work fine.

Any suggestion is welcome.

here is a log for errors:

Failed: Error in processing a transfer instruction:
File name too long [open(/home/futaz/Zotero/storage/.unison.EA8ANCBD.216e23d5f1eb3b900432d10b9d2e8107.unison.tmp/p0tlsvb29+fn540ih6qaaafxjrefuenq00ekogdama0b3b9nx8p+figohgnjc4omosqvgk4ivdziufojgig9fcuqscq00ndhmfjtfuhynqs64qlk3y2ry2pi1zdrygdd0tp3rqpwfp3upuwadailgjuqqwapqaaaaaelftksuqmcc@22)]
Failed: Error in processing a transfer instruction:
File name too long [open(/home/futaz/Zotero/storage/.unison.U9E8QTQI.216e23d5f1eb3b900432d10b9d2e8107.unison.tmp/site=cnn_money+cnn_money_position=1x1_over+cnn_money_rollup=business_news+cnn_money_section=jobs_and_economy+page.allowcompete=no+tile=1318952231992+page.allowcompete=yes)]

Failed: Error in processing a transfer instruction:
File name too long [open(/home/futaz/Zotero/storage/.unison.QVSM8PA5.216e23d5f1eb3b900432d10b9d2e8107.unison.tmp/go8laaaacv0rvh0y3jlyxrllwrhdguamjawos0woc0wovqxotownto0msswmdowmcyo2teaaaaldevydg1vzglmes1kyxrladiwmdktmdgtmdlumtk6mdu6ndermda6mdb5v6zlaaaagxrfwhrtb2z0d2fyzqbbzg9izsbjbwfnzvjlywr5ccllpaaaaabjru5erkjggg==@22)]
Failed: Error in processing a transfer instruction:
File name too long [open(/home/futaz/Zotero/storage/.unison.PS4TKHWG.216e23d5f1eb3b900432d10b9d2e8107.unison.tmp/site=cnn_money+cnn_money_position=1x1_over+cnn_money_rollup=business_news+cnn_money_section=jobs_and_economy+page.allowcompete=no+tile=1318952221512+page.allowcompete=yes)]
failed: storage/EA8ANCBD
failed: storage/U9E8QTQI
failed: storage/QVSM8PA5
failed: storage/PS4TKHWG
failed: storage/EXJ88C6Q

adamsmith

That really seems like a unison problem, no?
Zotero caps filenames that are too long for the file system, but if unison adds another lengthy string in there that might happen - since these aren't very many, I'd just manually shorten the filenames (from the Zotero interface so the link btw. file and Zotero doesn't get broken.).

baris

I think both Zotero and unison have problem since they both are adding to the same string. If there is a maxsize limit for a common resource, I think Zotero shouldn't use all the limit to itself.

Searching eveything I cannot locate go8laaaacv0rvh0y3jlyxrllwrhdguamjawos0woc0wovqxotownto0msswmdowmcyo2teaaaaldevydg1vzglmes1kyxrladiwmdktmdgtmdlumtk6mdu6ndermda6mdb5v6zlaaaagxrfwhrtb2z0d2fyzqbbzg9izsbjbwfnzvjlywr5ccllpaaaaabjru5erkjggg==@22) for instance.

Can you help locating them?

Thanks

adamsmith

it should be in the folder
QVSM8PA5 in the storage directory.
we'll have to disagree about the file length issue.
Unison is both creating the files that are too long (which I don't believe are necessary for rsync) and is hanging on the sync, so yeah, I'd say this is a Unison issue.
In any case, though - you're certainly free to use custom strategies for syncing files, but you can't expect Zotero devs to troubleshoot them for you.

baris

Sorry for missing information, part I couldn't figure out is to how to locate the file using Zotero interface.

I do not say unison doesn't have a problem. I say Zotero also have a problem. Do I get it wrong? Isn't it Zotero who name the file as: go8laaaacv0rvh0y3jlyxrllwrhdguamjawos0woc0wovqxotownto0msswmdowmcyo2teaaaaldevydg1vzglmes1kyxrladiwmdktmdgtmdlumtk6mdu6ndermda6mdb5v6zlaaaagxrfwhrtb2z0d2fyzqbbzg9izsbjbwfnzvjlywr5ccllpaaaaabjru5erkjggg

adamsmith

look at what file it is - maybe you won't even need it (in which case you can just delete it from the file system), otherwise that information should make it easier to fix things from Zotero.

As for the file name - Zotero saved it, it didn't create it. This is most likely from a website snapshot, so it was the designer of the webpage who created the filename.

ajlyon

You should be able to find it by entering zotero://select/QVSM8PA5 into the address bar, which will bring up the corresponding item in the Zotero pane.

dstillman

It's almost certainly just a useless ad file from some webpage you saved.

baris

Thanks for the workarounds, I'll try them. Although, I think
zotero:// URI won't work on the standalone version which I am using.

Aside, I do not see why Zotero must use the funny name a website author gave to some file while saving the snapshot.

adamsmith

These are usually files the parent .html links to - Zotero would have to correct all the links in the file if it started to rename files in a snapshot. That's not feasible.

Zotero Standalone should handle the zotero:// protocol on Mac OS X and Windows. It could support the zotero:// protocol on Linux as well with the appropriate packaging.

http://forums.zotero.org/discussion/8306/testing-zoteroselect/

dstillman

WebPageDump (the snapshot saving code we use) could probably be modified to just discard files with names that long, though—I suspect they're almost always worthless.