Comments on: Extended File Attributes Rock! http://www.linux-mag.com/id/8741/ Open Source, Open Standards Sat, 05 Oct 2013 13:48:18 +0000 hourly 1 http://wordpress.org/?v=3.1 By: Damien Cuvillier http://www.linux-mag.com/id/8741/#comment-1233965 Damien Cuvillier Fri, 20 Sep 2013 07:07:19 +0000 http://www.linux-mag.com/?p=8741#comment-1233965 Hi. Indeed, it rocks. I imagine building metadata on file libraries, without db, by adding, for example <b>JSON</b> data in user.comment field. Is it possible to use <b><i>find</i></b> command to search for anything stored in these attributes ? thks Hi.
Indeed, it rocks. I imagine building metadata on file libraries, without db, by adding, for example JSON data in user.comment field.

Is it possible to use find command to search for anything stored in these attributes ?

thks

]]> By: CharlieBrown http://www.linux-mag.com/id/8741/#comment-784929 CharlieBrown Tue, 26 Feb 2013 14:31:24 +0000 http://www.linux-mag.com/?p=8741#comment-784929 Believe it nepomuk has many other little defects besides this, for example, nepomuk needs metadata parsers, you can have nepomuk but if you don't feed it any metadata parser or file parser, he won't do anything with the files you give him. Obviously, there is also the detail that nepomuk will slow or even crash, if you feed him tons of data and believe it I have KDE 4 installed (in fact I'm writing it from it), but nepomuk is disabled since he can't handle all of the files I have, like the photographs, the audio files, the code from my projects and the videos from the many series I like (it is an total of 16GB spread over >400K of files). Also, people want something that can organize files physically in the correct order, not only index them and nepomuk fails drastically on this. Believe it nepomuk has many other little defects besides this, for example, nepomuk needs metadata parsers, you can have nepomuk but if you don’t feed it any metadata parser or file parser, he won’t do anything with the files you give him.
Obviously, there is also the detail that nepomuk will slow or even crash, if you feed him tons of data and believe it I have KDE 4 installed (in fact I’m writing it from it), but nepomuk is disabled since he can’t handle all of the files I have, like the photographs, the audio files, the code from my projects and the videos from the many series I like (it is an total of 16GB spread over >400K of files).
Also, people want something that can organize files physically in the correct order, not only index them and nepomuk fails drastically on this.

]]> By: linuser http://www.linux-mag.com/id/8741/#comment-611417 linuser Fri, 30 Nov 2012 11:23:26 +0000 http://www.linux-mag.com/?p=8741#comment-611417 is there any way we can read and write system attributes in a C program. getxattr seems to do for user land I am looking for the system ? Thanks is there any way we can read and write system attributes in a C program. getxattr seems to do for user land I am looking for the system ?
Thanks

]]> By: John Viveiros http://www.linux-mag.com/id/8741/#comment-266553 John Viveiros Tue, 26 Jun 2012 12:35:00 +0000 http://www.linux-mag.com/?p=8741#comment-266553 It would be most useful if you could rather add labels as per GMail instead of using file names and directories. This allows you to add one or more labels to categorise your data. For example: Accounts\ Invoice\ Outstanding\ could be three separate labels. Which could be reclassified as Accounts\ Invoice\ Paid\. Furthermore, dynamic filters can then be created to automatically classify data on the fly. It would be most useful if you could rather add labels as per GMail instead of using file names and directories.

This allows you to add one or more labels to categorise your data. For example: Accounts\ Invoice\ Outstanding\ could be three separate labels. Which could be reclassified as Accounts\ Invoice\ Paid\.

Furthermore, dynamic filters can then be created to automatically classify data on the fly.

]]> By: totedati http://www.linux-mag.com/id/8741/#comment-81521 totedati Wed, 14 Dec 2011 00:05:49 +0000 http://www.linux-mag.com/?p=8741#comment-81521 yes user_xattr extended attributes is nice in theory but until systems like kde4 nepomuk avoid extended attributes and use a centralized datacentric path for metadata manipulation, console and manual attributes manipulations ( using getfattr and setfattr tools and so on ) will be an marginal, periferic and exotic computer geeks exercise ... we have a lot of mp3 players with mp3 tag editing support but i do not find one with extended attributes manipulation capabilities in parallel with oldfashioned mp3 id tags embedded directly in mp3 file! right-click in my kde4 dolphin file manager -> properties -> advanced properties for any file or directory ... where is xattr tags? nowhere! when i make custom tags for files in kde4 using nepomuk and dolphin this tags is saved and dumped in an mysql centralized database not in file extended attributes who can be read and reread and set with getfattr and setfattr ... so, xattr manipulation is still hidden deep on my fs even if all of my partitions has user_xattr mount option enabled years ago waiting and waiting for things like kde4 nepomuk infrastructure ... but right now nepomuk and all kde4 social desktop infrastructure is completely decoupled from it! nepomuk choice to use centralized database for extended tag storage and manipulations is a non-optimal solution ... if i lose my nepomuk database file ( file corruption is not a stranger man even in linux world ) i lose all my tags! ... if i want to move files to other os or distro with other extended tag management solution ( why not an win7 or osx !? ) i need a database dump and a database import and no guarantee of inter-compatibility! what will be more compatible is a double layer approach in nepomuk, a centralized database management ( for performance reasons ) and for every nepomuk tag which is file related in the same time an extended attribute for that file created and saved in an automatic routine ( things like custom tags, for web pages original url, file type, photos and mp3 tags and so on ) when i extract from files this tags for nepomuk database ... all this extended attributes can be saved with files, moved with files ( i think extended attributes is compatible even with ntfs fs from windows world! ) and later on rescans tags databases recreated in a reliable way directly from fs storage! all precious manual tags will be with us, saved in extended attributes, with no need to tag again and again with every os reinstall all files ... yes user_xattr extended attributes is nice in theory but until systems like kde4 nepomuk avoid extended attributes and use a centralized datacentric path for metadata manipulation, console and manual attributes manipulations ( using getfattr and setfattr tools and so on ) will be an marginal, periferic and exotic computer geeks exercise …

we have a lot of mp3 players with mp3 tag editing support but i do not find one with extended attributes manipulation capabilities in parallel with oldfashioned mp3 id tags embedded directly in mp3 file!

right-click in my kde4 dolphin file manager -> properties -> advanced properties for any file or directory … where is xattr tags? nowhere! when i make custom tags for files in kde4 using nepomuk and dolphin this tags is saved and dumped in an mysql centralized database not in file extended attributes who can be read and reread and set with getfattr and setfattr … so, xattr manipulation is still hidden deep on my fs even if all of my partitions has user_xattr mount option enabled years ago waiting and waiting for things like kde4 nepomuk infrastructure … but right now nepomuk and all kde4 social desktop infrastructure is completely decoupled from it!

nepomuk choice to use centralized database for extended tag storage and manipulations is a non-optimal solution … if i lose my nepomuk database file ( file corruption is not a stranger man even in linux world ) i lose all my tags! … if i want to move files to other os or distro with other extended tag management solution ( why not an win7 or osx !? ) i need a database dump and a database import and no guarantee of inter-compatibility! what will be more compatible is a double layer approach in nepomuk, a centralized database management ( for performance reasons ) and for every nepomuk tag which is file related in the same time an extended attribute for that file created and saved in an automatic routine ( things like custom tags, for web pages original url, file type, photos and mp3 tags and so on ) when i extract from files this tags for nepomuk database … all this extended attributes can be saved with files, moved with files ( i think extended attributes is compatible even with ntfs fs from windows world! ) and later on rescans tags databases recreated in a reliable way directly from fs storage!

all precious manual tags will be with us, saved in extended attributes, with no need to tag again and again with every os reinstall all files …

]]> By: perl2ruby http://www.linux-mag.com/id/8741/#comment-9980 perl2ruby Sun, 14 Aug 2011 12:48:14 +0000 http://www.linux-mag.com/?p=8741#comment-9980 It would be useful to have options for find command that support check for extended attributes. It would be useful to have options for find command that support check for extended attributes.

]]> By: evanlec http://www.linux-mag.com/id/8741/#comment-9729 evanlec Tue, 19 Jul 2011 10:59:36 +0000 http://www.linux-mag.com/?p=8741#comment-9729 This is a very nice bit of info about Linux filesystems. Directories and filenames are just simply not enough these days to properly organize files in a way that makes sense to humans! I am curious to see what tools/scripts/etc have been written to try and take advantage of the user namespace attributes in order to better organize a users home (~) directory. I think I shall start taking advantage of this immediately for myself! This is a very nice bit of info about Linux filesystems. Directories and filenames are just simply not enough these days to properly organize files in a way that makes sense to humans!
I am curious to see what tools/scripts/etc have been written to try and take advantage of the user namespace attributes in order to better organize a users home (~) directory. I think I shall start taking advantage of this immediately for myself!

]]> By: lakshmipathi http://www.linux-mag.com/id/8741/#comment-9726 lakshmipathi Tue, 19 Jul 2011 08:12:59 +0000 http://www.linux-mag.com/?p=8741#comment-9726 >Is it possible to read these attributes from C code? Are you looking for getxattr() ? Check man getxattr for more >Is it possible to read these attributes from C code?
Are you looking for getxattr() ? Check man getxattr for more

]]> By: mukiwa http://www.linux-mag.com/id/8741/#comment-9668 mukiwa Mon, 11 Jul 2011 12:29:59 +0000 http://www.linux-mag.com/?p=8741#comment-9668 Is it possible to read these attributes from C code? Is it possible to read these attributes from C code?

]]> By: fsl http://www.linux-mag.com/id/8741/#comment-9628 fsl Thu, 30 Jun 2011 15:27:14 +0000 http://www.linux-mag.com/?p=8741#comment-9628 Very nice, but given the majority of files come from desktop users running Windows, how can I tive a practical to those extended attributes? Does samba support them and present then in a way a windows user can query / set from Windows Explorer? Is there any integration to Microsoft Office or Open/Libre Office document properties? Any document management system supports them? And what about header / tags embebed in multimedia files? Very nice, but given the majority of files come from desktop users running Windows, how can I tive a practical to those extended attributes? Does samba support them and present then in a way a windows user can query / set from Windows Explorer? Is there any integration to Microsoft Office or Open/Libre Office document properties? Any document management system supports them? And what about header / tags embebed in multimedia files?

]]>