Re: administering directory exclusions for beagle
The problem now becomes how to exclude these directories from beagle indexing and to do it in a way that is automated by the system so that the exclusion is incorporated into the beagle config for each new user without some manual intervention by the user or admin.. The problem is that when a What version of beagle exists in SLED10 ? If its 0.3.x, then take a look at the global config files at /etc/beagle/config-files. Look specially for Daemon.xml and FilesQueryable.xml. FilesQueryable.xml contains the default excluded directories and they allow environment variables (e.g. you can specify $HOME/mapped_dir). You might also want to add Files (and other relevant backends) to Daemon.xml:DeniedBackends so that by default these backends are turned off. If it is 0.2.x or earlier, then the /etc/skel approach is the only way that comes to my mind. Although, I cant quite say off the top of my head what is going wrong. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: administering directory exclusions for beagle
Unfortunately it looks like its 0.2.x and that's SLED10-SP2. Any idea why the files from /etc/skel are over written or deleted? Nothing comes to my mind. Are you sure the location, name and the structure of the file is correct ? I dont remember the details of the 0.2.x series but I would suggest running the beagle-config as a user and then copying the generated file. Also, if the user homedirectories are not on a network filesystem, try running beagled with BEAGLE_SYNCHRONIZE_LOCALLY=1 set. If $HOME/.beagle is not a local fs, then beagled tries some things which could have removed the file. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Desktop Search Hackfest
Hey Max, I am thinking about joining in order to help hacking on the beagle side of things. So far it looks like Kevin is the only one joining for Kevin and Lukas have signed up till now. Unfortunately most of the other people who are involved with the project wont be able to make it. I didnt email the list since there were not a lot of people who were participating lately; so I wasnt sure if people are at all interested. beagle. Are there any thoughts on the concrete coding tasks proposed on the wiki - what should we focus on etc? I missed Kevin on the IRC today, but I think it would be wise to take advantage of the fact that other desktop search devs would be there (looks like full team for Strigi/Nepomuk and Tracker). I have a few stray thoughts in mind, not much. * Someone was proposing a sort of test-suite with lots of sample files and a database of what works and to what extent. There is already a bunch of files in svn/trunk/beagle/testing/files but we never managed to create a proper test suite. If you all can decide a common test suite then that will help all the projects. It will also help the developers in improving their filters/analyzers. * A gnome-do plugin would be good and a krunner plugin would be good. You dont really need to be at the hackfest to do these though :-) * I know Arun (beagle-xesam author) is a bit busy these days, so any kind of help in beagle-xesam would be nice. Since Mikkel and other xesam people would be there, this is a good chance to figure out what needs to be changed in beagle or where beagle wont conform to the spec. I havent put much thought into the meeting; also I will be gone for a month starting Wednesday so pretty much that is all I can help you with. Keep up posted on what happens, - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: beagle-search segfault on exit
I got this bugreport: https://bugzilla.redhat.com/show_bug.cgi?id=455978 and was able to reproduce it. Simply closing beagle-search results in a segfault. Any idea what could be causing this? It works just fine but crashes on shutdown, seems like it tryes to free a pointer when doing cleanup which causes this. There was a similar problem like this in one of the hardy alpha releases. It went away after they released an updated mono (or gtk-sharp, I dont remember). Is there any pattern in the crashing e.g. does it crash always or it crashes only if something is searched or only when something is searched and some result is clicked ? If you run it under gdb, does it give you some more information ? - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: beagle webinterface
ok, this don't give any error... but when I do an alert of mimetype value it only shows me $mimetype... Then use onClick=openhit('[EMAIL PROTECTED]', '{$mimetype}'); return false;) one question... is possible that user when clicks the link href changes for the modified [EMAIL PROTECTED] value? Not sure what you mean here. Can you explain with an example ? You might be able to change the href when the user clicks the link, but that would again involve writing an onClick handler. I think you should be able to change the href and then return true in which case the click will open the modified href. But this all javascript manipulation happening here. And I am not the best person to ask these :-). You should consult some advanced javascript tutorial or something. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: beagle webinterface
a href=return of function openhit() onClick=openhit(...) then, in function openhit function openhit(url){ direccio = modification of value url; return(direccio); } then in href, nor #, go the edited link, then we don't need to use mimetype and nothing more... From what I understand about HTTP servers, even then you would have the same problem. document.open(...), window.open(...) or clicking on an href uses the same HTTP calls and browser functions to open a new document. AFAIK, its the HTTP servers responsibility to inform the browser about the mimetype. You can test this, in openhit(), instead of opening a new window with the document, open a blank window and set its href to direccio. That will achieve the same effect as clicking on a link. See e.g. http://www.irt.org/articles/js128/index.htm#4.1 -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: beagle webinterface
Use the suggestion below ... a target=_blank href=# onClick=openhit(\[EMAIL PROTECTED]); return false; Make sure you have the return false. and add a method like function openhit (url) { //... open a new window using javascript // verify the javascript - I am writing from memory var newurl = url.replace (file:///, http://host/var/www/;); window.open(newurl); } -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: beagle webinterface
See if using single quotes help i.e. a target=_blank href=# onClick=openhit('[EMAIL PROTECTED]'); return false; when I put dashes on [EMAIL PROTECTED] always gives me the same error :( Error: no ben format Fitxer font: http://192.168.2.8:4000/hitresult.xsl Line: 60, column: 29 Codi font: a target=_blank href=# onClick=openhit(\[EMAIL PROTECTED]); return false; Error: uncaught exception: [Exception... Component returned failure code: 0x8061 [nsIXSLTProcessor.importStylesheet] nsresult: 0x8061 (unknown) location: JS frame :: http://192.168.2.8:4000/default.js :: TOP_LEVEL :: line 810 data: no] that is what I have on default.js function function openhit(url) //line 777 { alert (url); } Blackhold + Material SCG08 http://associacio-aoe.org/scg/charlas + Respect free software communities and use free and/or standard formats while sending attached files. Files such .doc, .xls, are formats that don't follow the standard rules and may cause problems on viewing or in the format of the text when these are opened with programs where thousands users communities develops free software. So that I ask for the files that are sent to me, to be odf, ods, odp, svg, png or in default pdf. Thanks you a lot... On Thu, Jul 17, 2008 at 1:06 PM, Debajyoti Bera [EMAIL PROTECTED] wrote: Use the suggestion below ... a target=_blank href=# onClick=openhit(\[EMAIL PROTECTED]); return false; Make sure you have the return false. and add a method like function openhit (url) { //... open a new window using javascript // verify the javascript - I am writing from memory var newurl = url.replace (file:///, http://host/var/www/;); window.open(newurl); } -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Beagle Craw not working correctly on beagle 0.3.3
It seems the problem is not 100% fixed for me : Beagle founds the directories which are in my crawl but do not found any mp3 files or jpeg files. What query are you giving ? And what results are you expecting (give one or two example paths that you expect to match) ? Can you try queries like .mp3 or .jpeg to match queries by extension ? These are queries that would produce sure hits. Also anytime you have such a problem, make a quick check with beagle-query. Sometimes beagle-search fails to show certain results (due to some bug). I have some exceptions from beagle in the logs (sorry for the spam) : You can ignore these debug-exceptions Debug: Caught an exception sending Beagle.HitsAddedResponse. You were probably using beagle-search or kerry. Here is one of my crawl configuration file : [EMAIL PROTECTED] beagle]# cat /etc/beagle/crawl-music ... CRAWL_DISABLE_FILTERING=yes I am assuming you disabled filtering on purpose. This will only index the filename and extension and not pass the file through the corresponding filter (the mp3 filter extracts the mp3 specific properties like artist, album etc.). - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Beagle won't index e-mails in KDE4 KMail
Beagle just won't index my mails stored in KDE4 KMail (Version 1.9.52), it did so on my previous openSUSE 10.3 with KDE3 KMail. The KMail slave is activated via the Kerry KDE GUI and thee-mail account type is disconnected IMAP as before on openSUSE 10.3. Any ideas why Beagle does not like my new KMail, at least so it seems? This could be due to change in the directory structure of the kde kmail cache, or change in their file format or both. Last time I checked, KDE4 KMail didnt yet move to Akonadi but they could have changed something else. If you can file a bug then we will try to identify what changed and fix the problem. Thanks, - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Crawl System and Beagle 0.3.3
Debug: Starting QueryDriver Debug: Found 0 backends in /usr/lib/beagle/Backends/ThunderbirdBackends.dll Debug: Found 0 backends in /usr/lib/beagle/BeagleDaemonLib.dll Debug: Reading mapping from filters Debug: Found 0 user-configured static indexes.. There it is. There should be a line saying ... Debug: Loading system static indexes. ... (and if it found any) Debug: Initializing static queryable: /var/cache/beagle/indexes/documentation Debug: Found 1 system-wide indexes. ... I suspect this is due to wrong path supplied when building the rpm. Check one thing, move one of the index to /usr/var/cache ... e.g. /usr/var/cache/beagle/indexes/documentation Usually ./configure --localstatedir=/var sets the path correctly, otherwise localstatedir defaults to $prefix/var - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Crawl System and Beagle 0.3.3
I am running Beagle 0.3.3 on Manriva 2008.1. I am wondering if the bug I already reported for beagle 0.3.6 (at the time I used another distro), http://bugzilla.gnome.org/show_bug.cgi?id=530235 also impact beagle 0.3.3 ? Should not be. #530235 was a regression in 0.3.6 as far as I remember. I am currently out of town so I might be mistaken, though unlikely. I cannot make the crawl system work with beagle 0.3.3, and I do not know how to investigate whether it is Beagle's fault, my fault or Mandriva's fault ? Hmm... run these quick tests to see if those indexes are built correctly and picked up by beagle $ ls -l /var/cache/beagle/indexes = should show documentation, application etc. system indexes $ beagle-manage-index /var/cache/beagle/indexes/documentation info = it should show count 0 = replace documentation with names of the other indexes, and it should be 0 for all the relevant ones $ beagled --fg --debug --backend documentation (from another terminal) $ beagle-query --max-hits 5 source:documentation = should return 5 results = replace documentation with other relevant indexes, and it should always return 5 results Let me know how this goes. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
indexing removable sources
Hi, I checked in one attempt at indexing removable media (CDs, USB keys, external hard drives, network mounts or even regular directories which can loaded, unloaded at runtime etc.). * Get svn trunk, build, install etc. * Start beagled (pass --backend none to run with only removable indexes) * Create a static index and pass --removable to make it removable $ beagle-build-index --removable and other usual options --target /path/to/index /path/to/mount/directory * you can pass --source source to use an easily recognizable name for the media, it is set to the directory name of /path/to/index by default * you can store the index inside the mounted directory or outside, as you wish * apart from the source, the Hal volume-label is stored as beagle:RemovableLabel and can be searched against. * Ask beagled to mount the removable index at runtime $ beagle-removable-index --indexdir /path/to/index --mount /path/to/mount/directory * Check beagle-index-info to verify * Perform queries * Ask beagled to unmount the removable index, again live $ beagle-removable-index --indexdir /path/to/index --unmount /path/to/mount/directory There might be some rough patches so I would really appreciate some testing. And some feedback on the command line options, the beagle-removable-index tool and any other information to set or store in the removable index. Thanks, - dBera PS: Why use two separate paths for the index directory and the mounted directory ? Because in the past discussions, we could not decide where to store the index for removable sources. This way users can decide themselves what works best for them. PPS: It is trivial to allow searching unmounted media. In that case, if a result is found in the unmounted medium, then it will be flagged and the UI can display a prompt asking the user to insert the specific medium. Sort of like searching catalogues. Would this be any useful ? -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Beagle license change proposal from MIT/X11 to GPLv3
The whole issue was/is solved with a SemWeb version that is GPLv2+ licensed ...snipped So does that mean as long as we use a Semweb that is dual licensed, there is _nothing_else_ that we need to do ? Right. I thought that was clear on IRC already. There is no need to relicense code of beagle, it can stay MIT/X11 as thats compatible with everything. Thanks to all of you, I think the issue is resolved now. We are continuing as it is (*) and everything is all right it seems. Thanks again and sorry for letting my confusion create this mess. - dBera (*) I have to update Semweb which I was going to do anyway. -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Beagle license change proposal from MIT/X11 to GPLv3
Recently we came to know that some distributions have difficulty with the current Beagle license. Specifically, Debian does not recognise Creative-Common Attribution 2.0 (CC-by-2.0) as a valid free software license [1]. However Beagle requires Semweb which is dual licensed under CC-by-2.0 and GPLv2-and-later. We used to include Semweb by accepting the CC-by-2.0 license. GPLv2 is incompatible with the Apache license (required by Lucene.Net). I can't quite understand why distributions have a problem taking the beagle code under MIT/X11, and combining it with the Semweb stuff under GPLv3 (as allowed if Semweb is, as you say, licensed as GPLv2-and-later). MIT/X11 is compatible (in this direction) with GPLv3, unless I'm very out of date with my software licensing knowledge, so they have a perfect right to do this, don't they? The resulting packages would be under GPLv3, but the source code in beagle's SVN would still be under MIT/X11. That does sound possible. After all, theoretically (e.g.) Debian is free is take the beagle tarball and release it under the GPLv3 license. However, perhaps your intent to change the license to GPLv3 to make it more obvious to distributions that they are entitled to create packages under the GPLv3. No, I don't have any such intention. Is there a debian bug / mailing list thread discussing this issue? A quick search on bugs.debian.org didn't show anything up, but perhaps it's hiding somehow! :) Nops ... Mirco 'meebey' Bauer of Debian brought it to my attention on the IRC. Mirco... any comments ? - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Beagle license change proposal from MIT/X11 to GPLv3
Hi Mirco, [please keep my CCed, I am not subscribed to dashboard-hackers (if thats a mailing list)] oops... sorry. This was my first post http://www.mail-archive.com/dashboard-hackers@gnome.org/msg04459.html The whole issue was/is solved with a SemWeb version that is GPLv2+ licensed (beagle uses/used a CC-by-2.0 only version though). So does that mean as long as we use a Semweb that is dual licensed, there is _nothing_else_ that we need to do ? Here a short copy/paste from the IRC 18:57:37 meebey http://www.apache.org/foundation/licence-FAQ.html#GPL 18:57:45 meebey SemWeb is GPL-2 or later 18:58:05 meebey so beagle as combined work: MIT + GPLv2+ + APL - (becomes as combined work) GPLv3 MIT being beagle itself GPLv2 being SemWeb APL being lucene.net - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: adding metadata to documents via web scraping
Updated my perl script (had a copy/paste error in the url somehow), now it seems to work. Nice. Unfortunately the Desktop Search doesn't show snippets for videos -- see attached screenshot: harrison ford is found for Indy.avi but it doesn't show where. Yeah, thats a bug :-( http://bugzilla.gnome.org/show_bug.cgi?id=371152 - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Firefox extension updated to FF3; testing requested
On Friday 11 April 2008 21:57:26 Jose daLuz wrote: Actually, while this doesn't throw any errors, I get no different results from modifying the old version to indicate 3.0.* was an acceptable max version for firefox. From what I understood from the commit messages, only bookmark indexing was changed for FF3. Nothing else was touched (except a sneaky code to detect at runtime whether FF2 or FF3 is running). A complete testing would still be good. But looks like webpage indexing is working for you so thats good. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: need help with testing: GMail live search (available in svn trunk)
I think you could even go one simpler and ask (a) if it's a google apps account and (b) if so, what is the domain name. The domain name in https://mail.google.com/a/joeshaw.org/#search/[EMAIL PROTECTED] would be joeshaw.org or https://mail.google.com/a/joeshaw.org/;. What I am asking is that is there is a notion of domain name for Google Apps that any google app user is supposed to know. And that I can safely add it to the end of https://mail.google.com/a/; to create the URL (safely - based on your experience). Yeah, it's pretty slow. It seems like we could cache the headers for certain IDs though? So at least largely overlapping services would be a bit quicker. Could you elaborate on this one ? Which IDs to cache ? Is this something specific to Google Apps ? There are log messages to the effect of Downloading headers for message ID 21. What I was suggesting was that if those IDs are stable, we could cache the headers locally so that we didn't have to download them again. I don't know anything about xemail-net, so I don't know if that is a possibility. I am not too keen on this one - these are IMAP IDs which could change between searches (if messages are deleted, moved or who knows on what other IMAP protocol). Furthermore, with caching comes all kind of messy details like how long to cache, how to synchronize cache, where to store the cache, could we store the cache on the disk across beagle searches, could we prefetch all headers ... etc. It a whole new game. Of course if anyone is interested, please go ahead. It would give a nice boost to performance and I would definitely like to see how it performs :-). - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: beagle r4658 - trunk/beagle/search/Beagle.Search.Tiles
I'm not sure if emblem-note is the best way to go, since ale the emblem-* icons are used by nautilus for file emblems. I'm not even sure if they provide the right size to fit our needs. I agree that the tomboy-note is pretty common in all GNOME desktops, but not neccessarily KDE ones w/ Gtk. The current icon used in beagle-search for Tomboy notes (tomboy-note) doesn't work, and there's no standard icon for notes specified in the Freedesktop Icon Naming Spec. So we'll use emblem-note instead which is fairly common among icon themes. I propose that we do what Tomboy did (yet again) for http://svn.gnome.org/viewvc/tomboy/trunk/data/icons/Makefile.am?revision=1842view=markup We should package our own icons for which their is no standard one defined. I personally like their tomboy*.{png,svg} icons :-). Also, if someone could file a bug with FD.org (or whoever maintains the spec) that will be awesome. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: search within a given path / paging mechanism
(CCing the mailing list) I have two questions concerning beagle-query: 1. I read your FAQs on beagle-project and noticed, that it is not well implemented to search within a given (sub-) path, I'd just like to know, if this will be implemented in future? Although I understand, that you will remove the boundaries between directories, I think this feature would be very useful. Currently searching among all files within a directory is possible but not within any subdirectories under that directory. Use like this inuri:name_of_dir or inuri:file:///full/path/of/dir. In the former case all directories with the given name are searched. (*) There is a long term plan to support searching under any directory, recursively including subdirectories but I dont see it getting implemented in the near future. The reason is it will involve a significant redesign of some parts. 2. beagle-query *.gif --stats-only tells me, that there are 1596 gifs in my indexed folders - With beagle-query *.gif --max-hits I can get as many hits as required - however: is there a possibility to specify a kind of offset, i.e. please give me 25 more hits, starting from hit 100 to implement some paging-mechanism (for browsing)? Apart from --stats-only, is there an option to get the total matches plus the hits matching the query? No, pagination is not supported and is unlikely to be implemented anytime soon. Pagination can be done by applications using the BeagleClient API; doing it on the beagled side will involve more or less the same steps so we left it out. The total matches is always returned, regardless of --stats-only. It is definitely there in BeagleClient API. If you are referring to beagle-query, the last set of hits should be followed/preceeded by a line saying how many total results were found. - dBera (*) I might have forgotten to add this to the wiki. -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
need help with testing: GMail live search
Hey folks, I wanted to see how a live GMail search backend will work so here is one. It just queries GMail IMAP server directly for searching. I need some help with testing. So if you are interested in helping out and you are not behind a proxy or anything, please read on. 1. Get this and unzip : http://cs-people.bu.edu/dbera/blogdata/GoogleBackends.zip 2. Copy *.dll.* to /usr/lib/beagle/Backends/ directory 3. Copy *.xml to /etc/beagle/config-files/ directory 4. $ beagle-config GoogleBackends GMailSearchEnabled true 5. $ beagle-config GoogleBackends GMailUsername [EMAIL PROTECTED] 6. $ beagle-config GoogleBackends GMailPassword password (Yes, its stored in plaintext in ~/.beagle/config/GoogleBackends.xml) 6.a Searches by default the folder INBOX, to search other folders (labels), use the GMailSearchFolder option of GoogleBackends 7. Verify that the binary file is registered correctly $ beagle-info --list-backends (should list GMailSearch) 8. beagled --fg --indexing-delay 0 --backend GMailSearch 9. $ beagle-query query_preferably_with_few_results 10. Should work with beagle-search too, but I cant access my Gnome VM right now. Thanks in advance, - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: need help with testing: GMail live search
I wanted to see how a live GMail search backend will work so here is one. It just queries GMail IMAP server directly for searching. I need some help with testing. So if you are interested in helping out and you are not behind a proxy or anything, please read on. ... 6. $ beagle-config GoogleBackends GMailPassword password (Yes, its stored in plaintext in ~/.beagle/config/GoogleBackends.xml) Just in case I scared anybody, the password is only stored in plaintext. However, all the network communication with the GMail IMAP server happens over SSL. So there is no worry of your account information getting stolen. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: need help with testing: GMail live search
Okay I have this running on Fedora 8 using Beagle 0.3.5 - it seems to work but the plaintext password thing, please tell me that will be replaced with gnome-keyring integration (yeah I know that sucks for KDE users but till we get a DE independant way to store these that is the best idea I have for safely storing the password and having it unlocked with the users session for the beagle process only). Yes, I have already received objections on that front :-(. I had that in mind too ... but there are a couple of limitations: - Requiring either gnome keyring or kde wallet support. Though I mildly prefer the but it is hard to use from C# as their C# bindings are incomplete. - I can avoid by asking the password directly but to get input from the user I need to show a dialog box - the daemon should not require X to run How to say this, I shamelessly ditched my Gnome mates and added KDE Wallet support :P. Well... dcop kded kwalletd readPassword wallet-id beagle [EMAIL PROTECTED] turned out to be a perfectly command-line way to get the password from KDE Wallet :-). - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: beagle r4658 - trunk/beagle/search/Beagle.Search.Tiles
I'm not sure if emblem-note is the best way to go, since ale the emblem-* icons are used by nautilus for file emblems. I'm not even sure if they provide the right size to fit our needs. I agree that the tomboy-note is pretty common in all GNOME desktops, but not neccessarily KDE ones w/ Gtk. The current icon used in beagle-search for Tomboy notes (tomboy-note) doesn't work, and there's no standard icon for notes specified in the Freedesktop Icon Naming Spec. So we'll use emblem-note instead which is fairly common among icon themes. (Assuming the size issue is not a concern, I dont know much about that) Currently we used tomboy-note. Some people were seeing blank icons with that. The Note tile is currently used for Tomboy hits, Labyrinth, EDS notes. Any suggestions for an icon which will more or less look like a note and will be present on any desktop (KDE or GNOME) ? I am sure we can assume the user is running Tomboy, Labyrinth or EDS if he/she gets a Note tile. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
How to deal with different encodings ?
Hey folks, We are having a bit of trouble in deciding (*) how to deal with files in an encoding different than the system encoding. By default, we use UTF8 everywhere and assume everything is in UTF8. Some file formats or data sources specify their encoding (emails, html files, office documents etc.) so those are not a problem. If non-UTF8 is used for filenames and such, a lot of non-beagle things also break; we are trying to use MONO_EXTERNAL_ENCODINGS to deal with this case. (**). For other files, depending on the file format, either UTF8 or the platform encoding is used. Its really a clumsy affair. Apparently Windows XP has a system setting how should I handle non-unicode programs where it is posible to assign a ISO8859-1 codepage. I have no idea how it determines if data is in non-UT8 encoding. So, even though someone could have a different system encoding, a completely different encoding could be used for file data and metadata. Its a perfect encoding mess :-/. I know its not possible to always determine the right encoding. We could have a BEAGLE_LANG variable, which if set, would specify the encoding to use while extracting data regardless of the System encoding. Probably most apps will fail while displaying that data, but being an indexer how far should beagle push its indexing ability. Any suggestions on what could be done to use the right encoding as closely as possible ? - dBera (*) http://bugzilla.gnome.org/show_bug.cgi?id=524077 (**) non UTF8 folders are not indexed - in progress - http://bugzilla.gnome.org/show_bug.cgi?id=440458 -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Extend snip when using verbose switch
Look for SnippetRequest variable = new SnippetRequest (); Then do, variable.ContextLength = number of context words; variable.SnippetLength = maximum length in number of characters; If you need exact code, tell me which tool (beagle-query or beagle-search) you need to change and I will describe the explicit changes. - dBera I see from that latest Beagle update that you have incorporated the API call. Excellent. . urm ... how do I edit to change the number of words and length of the snippet? Is it possible to extend the length, both front and rear, of the 'snip' when using the verbose switch with beagle query? If you mean by * using some parameter with beagle-query, then no * using some API call in beagle-query source - then not yet (blame my laziness) * by modifying source of beagled - then yes The infrastructure is there; I will try to remember to add an a API call to control it. -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Beagle on encrypted partitions yeilds horrible system performance
I am hitting kind of a nasty performance problem, my current test setup is a two disk mdraid RAID0 setup with lvm ontop of a dmcrypt, all partitions beagle touches are ext4. Now every time beagle 0.3.4 indexes a folder, the entire system becomes near non responsive, typing yeilds detection of multiple key presses and kcryptd and beagle-helper are combined using 100% CPU (in about a 80/20 split with beagle being the 20%). Hmm... and the encrypted partition is not a red-herring ? I mean, could it be ext4 ? Could it be extended attribute in ext4 ? Could it be just some undetected bug in beagle ? Ahh ... ok - kcryptd and beagle-helper are combined using 100% CPU (in about a 80/20 split with beagle being the 20%) - so there is something to do with kcryptd. Can you try this test ? Take a reasonably large text file. Run beagle-extract-content on it on a normal partition and your dmcrypted ext4 partition. If the time taken differs significantly, then there is a sure problem there and we can start from there. Unfortunately I am too far from using a dmcrypt partition on ext4 so it will be hard for me to test this directly :-( - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: ANNOUNCE: Beagle Libbeagle 0.3.4
/usr/bin/gmcs -debug -out:Beagle.Search.exe -target:exe ... ... ./Beagle.Search/Driver.cs(127,37): error CS1501: No overload for method `Register' takes `2' arguments Compilation failed: 1 error(s), 0 warnings I'm assuming you have ndesk-dbus-0.4. Here's a patch to fix this. I've committed this to svn too. The problem is that the Bus.Register() method changed between ndesk-dbus-0.4 and 0.6 (the 0.4 method being deprecated in 0.6). Thanks Arun. I will add this to the wiki Known Issues page. I like to beagle keep dependencies as low as possible but my machine sometimes (frequently) has more uptodate versions. *sigh* What an irony given that the second line of the release email said something about fixed build problems. /me prepares for another round of FTBFS storm -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: ANNOUNCE: Beagle Libbeagle 0.3.4
Hi Enrico, Your data will automatically re-indexed ... this behaviour should be reconsidered, since it exposes these two drawbacks: If you ran beagled 0.3.4, you can try to change the version from 18.x to 19.x in ~/.beagle/Indexes/*Index/version But read below to understand the risks with it. 1) (I think) the indices are dropped comletely on starting the new beagle version, so until everything is re-indexed, I cannot find many Right. And yes, sadly your old documents will not be found until everything is re-indexed. 2) (I think) I will loose all web pages that I ever visited, until I revist them, which I will not do when I can't remember them. So finding This is an even bigger problem. A solution I could imagine is not to drop the index, but to schedule all resources to be updated, und replace the lucene documents with the new ... First a lame disclaimer from old release emails At this point in development, we cannot commit to stable APIs or file formats. You will almost certainly need to reindex your data at some point in the future. :-D Wish I could use this in a serious way. Binary incompatibility is broken when the index format changes e.g. some new required internal field was added or some field structure changed. It would be hard to maintain backward ABI compatibility in those cases without add a pain, if at all possible. Now I absolutely do not remember if the version was increased this time for anything like that or just to force reindexing with newer lucene. If it was the latter, then it could have been avoided since lucene-2.1 is backward compatible and I hold myself responsible. I will be careful in the future and with lucene-2.3. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: ANNOUNCE: Beagle Libbeagle 0.3.4
To ensure that all indexed items match these new feautures we have to drop the index from time to time. However maybe implementing a migrator from our previous version to the new one should be the right thing to do. This would allow us to preserve web pages as you noted. This is actually not impossible for certain backends and I believe possible for all our web history backends. We store their cached text in the textcache, so we could use the previous properties in the index along with the textcache cached text to migrate the information for the new index. The reason we need the textcache data is we do not store the text content of the documents in the index itself. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Stemmed search configuration
Is it posible to configure the stemmed search feture to other languages than english (e.g. danish)? Not yet. The main problem seems to be how to decide the language of the data/metadata for each document. Only very few data sources (some html files, emails probably) specify the language of the data. Beagle has the means to use a different stemmer for each document but not with different metadata of a document. For most documents, only some data/metadata fields are in a different language and the others are generally in English. It will be hard to get it right everytime, so currently we just default to English. If you are using 0.3.x and you are willing to modify the source then change in beagled/LuceneCommon.cs: DEFAULT_STEMMER = English; to DEFAULT_STEMMER = Danish; Beware that this will use the Danish stemmer for every data/metadata indexed. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: image search with path data
The path to a image collection often teels a lot about the images in a folder. We could add this information in the EXIF comment of each image - but I do not like redundant text for each image when the information already is in the file path. E.g: /home/xxx/images/2007/barbeque/Brother and mother.JPG A searching for: barbeque mother Yes agreed and no, beagle does not search the query term in the path. This is due to a technical limitation (*). With 0.3.3 (IIRC), you can use a limited feature by using inuri:barbeque mother which will match the file above; however, the inuri only works with the current directory and not any directory above. While most people use the path information to categorize data, such information gets lost when files are moved (directory reorganization, email or upload a file etc.). Using embedded metadata seems to be a better solution in these current times. This is just my personal opinion. - dBera (*) For those who know, Joe has often mentioned this in the past. This is due to a design decision which optimized the file system backend to be robust against moves and renames in the filesystem but made path searching impossible. Fixing this will require something close to rewriting major parts of the file system backend. -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: jpeg exif comments with ascii char ord(127)
beagle-build-index/beagle-extract-content does not include ascii char ord(127) from jpeg exif comments. Is this a bug or are only ascii ord(127) allowed in exif comments? Its more of a feature but I dont know enough about this. I am told that Exif does not specify the encoding and ASCII is the most common one. However, for some reason which I do not remember anymore, we use UTF-8 for Jpeg JFIF comments. If you can find out any information about how the encoding is specified for Exif metadata, please file a bug and we will try to resolve this. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Extend snip when using verbose switch
Is it possible to extend the length, both front and rear, of the 'snip' when using the verbose switch with beagle query? If you mean by * using some parameter with beagle-query, then no * using some API call in beagle-query source - then not yet (blame my laziness) * by modifying source of beagled - then yes The infrastructure is there; I will try to remember to add an a API call to control it. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Bad hang in gtkfilechooser after running beagle for awhile
Okay...several hours later, plus one restart of X (unrelated to this bug)... - only one beagled and beagled-helper are running (ps ux | grep beagle) Here's the output of ps ux | grep beagle: sandy 1773 0.0 0.6 89768 28164 ?SNl 06:51 0:01 beagled /usr/lib/beagle/BeagleDaemon.exe --replace --bg sandy 1908 0.0 0.4 61648 18092 ?SNl Mar11 0:02 beagled-helper /usr/lib/beagle/IndexHelper.exe sandy 6135 0.0 0.4 60860 18396 ?SNl Mar11 0:02 beagled-helper /usr/lib/beagle/IndexHelper.exe sandy 7084 1.0 1.3 131428 55224 ?SNl Mar11 13:48 beagled /usr/lib/beagle/BeagleDaemon.exe --bg sandy10197 0.0 0.4 61772 18372 ?SNl 00:55 0:02 beagled-helper /usr/lib/beagle/IndexHelper.exe sandy23138 0.0 0.0 3040 728 pts/4R+ 10:08 0:00 grep beagle sandy30328 0.0 0.4 61896 18600 ?SNl Mar11 0:02 beagled-helper /usr/lib/beagle/IndexHelper.exe sandy32666 0.0 0.4 63292 18708 ?SNl 06:31 0:01 beagled-helper /usr/lib/beagle/IndexHelper.exe Multiple beagled and multiple indexhelper will certainly cause problem. If you killed X (session) by ctrl-alt-backspace kind of thing, then that might not have killed the existing beagled. The one started in the next session should have in turn killed the previous one (but looks like it didnt). And then there are lots of indexhelpers too. Most of this is fixed in 0.3.0 and some extreme corner cases are scheduled for 0.3.4. For now, if you can manage to somehow call beagle-shutdown at logout (e.g. session logout scripts), that will save you some trouble. With 0.2.x, the goal is to ensure that there is only one beagled and one index-helper. In the worst case, run a script at startup which will send SIGINT followed by a SIGKILL after 10sec to all beagled and beagled-helper. Are you still interested in seeing ~/.beagle/current*? I've copied them over, just in case. Now I'm going to restart beagle. No need for them. I believe the error in your case is solely due to beagle processes deadlocking while quitting. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
daemonless beagle
Hey, I added a tool beagle-static-query to query beagle index _without_ beagled. It is really meant to be used with static-indexes i.e now you can use beagle-build-index from a cron-job and beagle-static-query to query beagle (like updatedb/locate). It can still be used with any backend except you wont get live-query and latest changes (since beagled is not running, only the last built indexes are queried). beagle-static-query accepts arguments from beagled (regarding specying backends like --backend and --add-static-backend) and from beagle-query. The query syntax is same as beagle-query (for some reason snippet support is not working right now, I will fix it later). If no backend is specified by --backend all backends enabled in the config will be queried. Use --backend none to disable querying all enabled backends. Some examples: $ beagle-static-query --backend applications --verbose konqueror $ beagle-static-query --backend Files FileAttributesStore $ beagle-static-query --add-static-backend /path/to/personal/index --backend none kconfig Enjoy ! - dBera PS: This tool is lacking a manpage. If someone could write one, I will add it to svn trunk. Thanks. PPS: There is really no magic here. Internally, the tool starts a minimal beagled, queries it directly instead of using BeagleClient API and then stops it. The surprising thing is that it is fast to do so. -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: daemonless beagle
It is really meant to be used with static-indexes i.e now you can use beagle-build-index from a cron-job and beagle-static-query to query beagle (like updatedb/locate). It can still be used with any backend except you wont get live-query and latest changes (since beagled is not running, only the last built indexes are queried). And I forgot to mention how to run beagled in only-indexing-mode. Of course you can use beagle-build-index to build static index for files and directories. But for other backends, you can use beagled --indexing-test-mode to ask beagled to build/update the index and then exit. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE / Mandriva / Inspiron-1100 ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Could nemo get listed
Since 0.2 Nemo[1] has been able to use beagle as search backend using the beagle xesam adaptor. Is it possible we could get listed on this page under Projects using Beagle? Thanks :) http://beagle-project.org/Development /me ponders for 7 minutes 13 seconds ... consults the president of the inter-galactic beagle fan club ... and Of course, yes :) Please add it. - dBera PS: If anyone has any other projects using beagle, feel free to add it to too. -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Getting started with beagle
An architectural decision to be made, do we want to actually index the data off of every webservice, or just offer 'transparent' backends to query the existing query API's for each service. I'm more for a local A transparent proxy backend to query using webservice API (in beagle lingo, a QueryDriver) is fine for some kind of data but ideally a real backend that fetches the data and indexes it (backend) would be the best option. copy (makes it fast, and solid even when disconnected, but just my $0.02) I love writing/overhauling new backends, so I might stab at some of these (im actually thinking of maybe an out-of-process script that does its Beagle interaction like the Mozilla extensions ect, so we aren't responsible for its scheduling.) An out-of-process script will work but it is really not that complicated to do this in process. All you have to do is create an IndexableGenerator and feed indexables as asked in GetNextIndexable. Depending on how fast the data can be accessed from the webservice, either download some 30/40 indexables from the webservice in HasNextIndexable or use a separate thread to download them and put in a shared queue from which GetNextIndexable will get them. If you do it out of process, make sure you dont choke the internet by downloading all 10K emails in one go i.e. you can't ignore some kind of scheduling. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Beagle Properties
There is still a small amount of stuff that needs to be finished before we can fire off this event. Stay tuned. Changing these properties is pretty dangerous, because we will effectively be changing a string API. That means that apps that use Beagle will still compile, but they will silently break when their old string mappings don't line up to the new ones. So I think it is important for us to take the initiative ourselves to fix the applications and add-on backends and filters that use Beagle. I think it would be helpful to collect a list of these on the wiki page and have it be a core part of this work. That sounds overwhelming ! There are only a few that are within our reach ... yelp, nautilus (and possibly brassero) in gnome cvs and kerry in kde svn. There are more to which we dont have direct access. Isnt there a standard way of making string changes in a way to automatically make applications aware of it ? Can these changes qualify as breaking binary incompatibility; then we can increase library versions for both libbeagle1 and beagle-0.0. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Beagle Properties
is probably a good idea. And not to sound like a broken record, but this stuff should probably go on a branch (or create a 0.3.x branch and continue on trunk) so that we don't find ourselves in another rut. IIRC, there is already a branch dedicated to this ... don't remember the name right now, you can check it in websvn. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Beagle Properties
We could do this, yeah. We could bump the API versions of the C# assemblies and the so number of the libbeagle shared libraries. This And maybe also bump beagle version to 1.0 :) OSNews Headline Open source project finally leaves beta. We're doomed! Ok, ok ... I know I am dreaming ... - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: web client for beagle
I've been working on a web client for beagle. It's written in python and can be checked out from http://code.google.com/p/beagle-client-pylons/ Wow. Nice! I am always interested in knowing what people do with beagle specially the API libraries. Tell us more. Brief web search tells me pylons is a rails like framework. So what is the objective of the web client ? How does this run ? Is this more of an API (i.e. one can write web apps using it) or it is directly a client itself ? Pretty screenshots :) ? It's a client not an api. But probably not hard to expose some web services. I'll put some screenshots up tomorrow and send the link. Nothing too fancy, but it is ajax. Also, I'll put some instructions on how to invoke it. Ok. That will be good. I tried to browse the svn (there was some problem with the google code interface, i could only manage to see the different revisions) and saw only a few small python files ... I guess its the power of pylons working fulltime. You are correct. Pylons is doing much of the heavy lifting. And also the beagle lib. I just thought it would be nice to search my desktop and download from anywhere. Interesting. Waiting for more details. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: web client for beagle
I've been working on a web client for beagle. It's written in python and can be checked out from http://code.google.com/p/beagle-client-pylons/ Wow. Nice! I am always interested in knowing what people do with beagle specially the API libraries. Tell us more. Brief web search tells me pylons is a rails like framework. So what is the objective of the web client ? How does this run ? Is this more of an API (i.e. one can write web apps using it) or it is directly a client itself ? Pretty screenshots :) ? I tried to browse the svn (there was some problem with the google code interface, i could only manage to see the different revisions) and saw only a few small python files ... I guess its the power of pylons working fulltime. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Disabling Akregator and Liferea backends
Hey all, Akregator and Liferea backends do not work with the recent versions of the corresponding applications. I think there is no point in shipping them as they are right now; so how about disabling them (*) ? Recent Akregator versions use Metakit, a file-based database with C++ API. There has been proposals and some effort in writing a SWIG interface file for it to generate bindings for other languages, but I was not able to find anything concrete. Liferea moved to sqlite from xml since August last year. While its generally hard to get change notifications for sqlite based apps, there could be ways to get the backend working again. Liferea might be storing the time a feed was retrieved in sqlite - that can be used to get new feed information (removing feeds is tricky). Liferea also has a dbus interface, so something could be done using that. A rough scan in its websvn showed some avahi code to publish and sync data ... I havent look deeper but there could be something there. Lastly, I spotted some plugin code in Liferea; if it supports plugins then a thunderbird like backend could be written. There could be more possibilities. These are just my initial suggestions to anyone thinking of fixing the backend. - dBera (*) Actually moving them to /svn/beagle/trunk/attic where unused code is kept. -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: The beginnings of an advanced query UI
So, I think we all agree that while the system of text-based keywords for advanced queries can be extremely powerful, the issue of discoverability makes them more of a stepping stone to a graphical Agreed. Good thinking there. Anyways, feel free to dismiss, to heckle, to praise, Just trying to think of a way to be something more than just searching text ;) One thing (and possibly the only thing) I understood from the mockup is that you want every displayed information to be clickable and clicking them would refine the current search by adding that information. If that can be done without cluttering the UI (e.g. displaying the action icon only on hover), it might be useful without looking bad. Just a suggestion. I did not really understand the other proposed changes. I am a bit UI challenged, so dont worry too much. I want to mention in passing two things which bother me about the current GUI (during my brief testing sessions). One thing which I really really dont like is that it uses the available space in a very bad way. The number of tiles are fixed and so enlarging the window makes the tiles look very sparse. This is probably tricky to do since it is hard to figure out how many tiles to display based on desktop size, dpi, font-size etc. but there is got to be some clever hack to achieve this. The other thing is about adding some fanciness to the GUI. Gradients, shadows, sliding animations ... usual bling. Note that I am not hinting at meaningless effects here, but subtle one which draw the attraction of the user, make him/her aware of the different information and actions and make the actions look like real actions (not like a magic, here are the documents 1-8, you click here, they vanish and out of the blue appears document 9-12). - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
0.3.3 : For distributions and Ubuntu users
Hey folks, Wanted to point out a couple of things. * We had to bump ndesk-dbus requirement to 0.5.2 - the first version to be thread safe. Anything lower could have caused deadlocks. Its nearly a year that 0.5.2 was released so most distributions would be fine. Except Ubuntu Gutsy shipped with 0.4.0 and they are not providing any update. However, I was able to install ndesk-dbus from the Hardy repositories in Gutsy without any problem; just that single deb had to be upgraded, no dependencies. * For distributions, beagle now ships with several crawl-files to creating system indexes and updating them nightly. Once created, updating is fast and if done regularly should not take more than a couple of minutes. But user experience can be improved significantly by making a few observations. We cannot set these when we release and is best done at the distribution level. - Set the correct paths in the crawl files. We tend to add all possible known paths and all possible prefixes. - I am not sure about enabling the windows index by default. Indexing windows partition will definitely take a lot of time and lot of resources and might take more than a couple of hours. - Some users complained that the cron jobs started soon after they finished installing the distribution - that will suck big time. It might be a good idea to update the system with security updates and other cron jobs but definitely not updating the index (not only for beagle, but also updating index for man, updatedb) - Ship with default indexes (I think OpenSUSE already does this). Create a system wide index and ship it as a package. Then the users will not need the first time crawling and indexing. Subsequent changes will be updates and will be fast. - The first night after installation is going to be rough for users :) As of now, three heavy indexes will be built for the first time - applications, manpages and documentation (and windows if that is enabled). E.g. building manpages on my system took about 30-45 minutes. Documentation takes even longer ... these are all text heavy data. If something can be done to make it easier, then the users (myself included) will love it. I dont know much about cron, but maybe the different indexes could be scheduled with different intervals so that no two are on the same day for the first time. If there is anything we need to do upstream, please email your suggestions. Thanks. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: System.Xml.XmlException: Document element did not appear. Line 1, position 1.
20080111 10:38:30.0808 27307 Beagle WARN EX: Caught exception in DoTaskReal 20080111 10:38:30.0808 27307 Beagle WARN EX: Tag: Crawling Nautilus Metadata 20080111 10:38:30.0808 27307 Beagle WARN EX: Creator: 20080111 10:38:30.0808 27307 Beagle WARN EX: Description: 20080111 10:38:30.0808 27307 Beagle WARN EX:Priority: Delayed (0) 20080111 10:38:30.0808 27307 Beagle WARN EX: System.Xml.XmlException: Document element did not appear. Line 1, position 1. 20080111 You are probably facing http://bugzilla.gnome.org/show_bug.cgi?id=504192 . It was fixed shortly after 0.3.1 and is available in 0.3.2 This error aside, I think it's some for some intelligence in the beagle log handling. Too many times beagle has filled my ~ with a single error I have thought about this over and over again ... one option is to disable verbose debug logging by default, but people are against it. The argument given is, such verbose logging leads us to the errors - without them, we would not be able to know of certain errors. This argument is kind of valid, though it can be quite taxing on the users. Btw, most of the distros lower the loglevel so users installing from packages would not face this problem. The other options like last message repeated work against the above argument. Are there reasons to believe that reporting last message repeated is better than disabling debug output ? - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: beagle IPTC crawling question
I am using googles picasa to manage my image collection. I have been adding keywords to the images in the picasa interface, and inspection after the fact shows that the keywords are being stored as IPTC data in the image. For instance, one image with the keywords cindy and swatch has this for IPTC data (found using exiv2 pr -p i imagename): ... Now, Beagle does not pick up either of those two keywords at all. IIRC, IPTC indexing was added in 0.3.0. Are you using any of 0.3.x releases ? Even with 0.3.x, some of the IPTC tags are not extracted due to a limitation in the F-Spot code that beagle uses for IPTC extraction. If I recall correctly, its one of the lens/makernote related tag. If you are using 0.3.x and beagle is not indexing Application2, then it might be related to the F-Spot code. Let us know. BTW, you can use beagle-extract-content /path/to/filename to figure out what properties are extracted. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: suggest: make searches efficient
Aside the question of scalability, I am wondering how to display this in the beagle-search GUI ? Without cluttering the interface. Well; I suggest showing results like in that link :-) Keep a flat list of all results that can then be refined by clicking on the clusters. Beagle-search does not show a flat list of results, and I dont see how to implement a sidebar with all those extra links (buttons ?) without making it look cluttered. Clustering/faceting should not be hard to implement, its a widely known and used idea. I am more worried about the user interaction part. OTOH, I am planning on showing a cluster in the webinterface. One major difference there: users are used to seeing lists of clickable text (with scrollbars) in a browser. Thanks for your suggestion though, - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: suggest: make searches efficient
Ie cluster/facet extraction on the result set? It is far from trvial to do in an efficient and scalable way, but it can be done... The website I linked to have 10M items in the index. Its easy to do such fancy data-mining tricks for a webserver. On a desktop, such fancy things might cause annoying CPU spikes. But still this needs to be implemented to see how far can it work. Aside the question of scalability, I am wondering how to display this in the beagle-search GUI ? Without cluttering the interface. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
xchat backend
Apologies for the hope-packed subject, I am not announcing any xchat backend :) But I did start thinking about the backend. The log files are very similar to Konversation files, so the Konversation code can be easily adapted to make it an xchat backend. There are two problems, and major ones. xchat stores the timestamps in the log files in user-configurable way, using strftime format. Changing the strftime format to a .Net DateTime format is one obstacle. The second one is worse. The default format (and for any user that stores the timestamps in this way) only contains the time (no date). The Konversation backend and other IM backends rely on time-date to index the chat lines according to the timeline. One option is to refuse to index log files which dont mention the full date too ... but that will be too harse. The other option, ... don't know. Those are pretty much the problems with implementing an xchat backend right now. I know there is some demand for an xchat backend, so I thought to share with you my findings. Anyone is welcome to look into the issues. Either discuss here or at http://bugzilla.gnome.org/show_bug.cgi?id=487867 - dBera PS: In other news, svn trunk now has a backend for the IM client Empathy. It will be shipped soon with 0.3.2. -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: insufficient? beagle-0.3.1.tar.bz2/gz on ftp.gnome.org
svn version beagle-0.3.1 and tar.gz on ftp are differed. tar.gz without Util/AvahiBrowser.cs (not work configure --enable-avahi; make). Archive.cs not contained too. Yes, there was an error in the Makefile which didnt package the above two files. Sorry for the trouble. Please download those two files from svn and add them to the extracted tarball. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Support for more ontology foo in beagle xesam adaptor
Small question, why are there two different page-count fields? Couldn't they just be one like it is done in tracker. And what about pdf's, does beagle support extracting the page-count from that? Noticing the ongoing effort to add more ontology to the beagle-xesam adapter, I think I should mention one shortcoming (or feature, depending on how it looks like): Beagle was designed to be a desktop search infrastructure which can index any kind of data you want it to index. The backends and filters, which are responsible for actually generating the indexed data, are free to put the data in any format they like. We never had any restriction on the names of the fields that they use. We mildly _suggested_ that they use commonly used names (either looking at similar Filter or Backend or by following Dublin Core metadata names) so that the UIs can also display results from that source. But it was never a requirement. Developing the right namespace and name of the field for a particular property of a some kind of data is not an easy task. Also such a list was available an year or two ago, when most of the Filters and Backends were written. Thus it made sense to give the freedom to Filter and Backend authors. Note that nothing stops from anyone running Beagle to create a Backend or a Filter out of the Beagle tree and dropping it in the right place for Beagle to pick it up as a plugin. Probably the usual beagle UIs would not be able to display results from the new source, but thats the authors responsibility (BeagleClient API exists if he wants to create a special purpose UI for his specialized data). What the above caused was a hotchpotch of property names and abundunt use of the namespace fixme. I would not be surprised to find names like dc:author, fixme:author or fixme:tags and fixme:tag. Its stupid but it is a daunting task to decide a policy and make the changes everywhere :( Frankly I am without clues how to resolve this. That, given the fact that life has been going well even with this mess, at least till now ;-). - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: beagle problem
(yelp:11000): Yelp-WARNING **: beagled not running, using basic search support. but beagled is running. This happens with beagle 0.3.1 and yelp 2.20.0 under Ubuntu 7.10. This is probably the same issue as libbeagle0 vs libbeagle1. yelp needs to be rebuilt with libbeagle1 [1]. I compiled yelp with the suggested correction, but it still complains that beagled is not running. Perhaps there is something else to change in the yelp's source, but at this point I think it's a task for yelp's developers. Hmm... I tried yelp on gutsy and it was refusing to see the running beagled. Then I tried $ nm -D /usr/bin/yelp | grep beagle_ and it listed a lot of functions but not the crucial one beagle_client_new (which, according to the websvn code, is inside a #ifdef). I have no idea what it means since some of the other functions inside the #ifdef do show up in the symbol list. I suggest filing a bug with yelp. Most likely something is wrong there. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Spelling suggestions
* Lucene only stores stemmed forms of the words (beagle becomes beagl) We have to figure out a way to unstem the word: 1.) Hack the analyzer to get the unstemmed word 2.) Traverse through our TextCache and find a word which which contains the stem part. This is what I'll be looking into today/tomorrow. You might want to check the Highlighter.net package (in Lucene.Net/contrib from their website). They highlight matched words. They use StandardAnalyzer in their example but I wrapped a PorterStemmer around it and asked it to highlight words with same stem and it was able to do it. One way I had in mind was to create a tokenstream, check if the tokentext is the same as the suggested stem, if yes use the token.startoffset, token.endoffset to extract the actual text. Of course its easier said than done ;-) We need to only return the highest relevant suggestions, based on: 1.) Term frequency in index 2.) Levenshtein distance score Add to that there could be multiple indexes so results from multiple indexes need to be intelligently merged. Sorry, for the exhausting email and lets make Beagle rock! :-) Ya !!! - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Compiling beagle on mono 1.2.6
I had tried to compile beagle on newest mono (version 1.2.6). All related packages of mono were updated to (e.g. gtk-sharp, etc). But beagle doesn't compile, here is the error: Can you check if the suggestions given in http://bugzilla.gnome.org/show_bug.cgi?id=503639 work for you ? Thanks, - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: beagle xesam
So beagle-xesam-0.1 was also released yesterday. Congratulations and thanks. Would you mind giving a few details on what exactly it does now ? How can it be used ? Where does it sit between beagle and xesam (pun intended) ? Where does it go from here ? ... Basically the adaptor as it stands allows Xesam clients to connect to it, and pass it queries. The adaptor acts as a translator between the Xesam API and Beagle's native API. It takes Xesam queries over DBus, passes them on to beagled, and the hits back in Xesam-ese. Beagle-Xesam largely conforms to the latest (RC1) Xesam spec. ... To summarize, the basic adaptor is functional as it stands. The path ahead should become clearer as clients that use the spec arise. Ok so its functional and conforms to the spec. Is there any possibility of bringing the adapter into beagled itself ? - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: beagle r4293 - in trunk/beagle: BeagleClient beagled search search/Pages
Implement suggestions in an unobtrusive way using FuzzyTermEnum. Suggestions are only generated upon request. How expensive is the search for suggestions? If they're inexpensive, we might want to consider returning a SuggestionsResponse for every Query instead of or in addition to explicitly requesting them. I have a general question on generating suggestions; I emailed Lukas but I will also ask here in case someone familiar with language processing knows the answer. We only index the stemmed form of any word. We dont store the actual word anywhere. E.g. dancing very nicely gets indexed as danc veri nice. Using the stemmed words to find similar words will only give us similar stems - which is OK but how to get an actual sensible English word from the stemmed word ? E.g. say the incorrect query was vedy - it will stem to vedi - thay will return a similar term veri - but how to get back vedy (or some meaningful word, need not be of the exact form as very) ? Are there known solutions without using wordlists ? - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: beagle problem
noticed that yelp is not able to use beagle daemon, when a search is launched it warns: (yelp:11000): Yelp-WARNING **: beagled not running, using basic search support. but beagled is running. This happens with beagle 0.3.1 and yelp 2.20.0 under Ubuntu 7.10. This is probably the same issue as libbeagle0 vs libbeagle1. yelp needs to be rebuilt with libbeagle1 [1]. - dBera [1] Some distributions have patched yelp to build with libbeagle1 by commenting out beagle_query_add_source (query, documentation); This is incorrect; the API was removed and should be *replaced* by (as described in http://mail.gnome.org/archives/dashboard-hackers/2007-December/msg4.html): beagle_query_add_text (query, source:documentation); -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: beagle problem
However some other programs (nautilus and beaglefs) depend on libbeagle0 (i.e. the 0.2.18 version of the library) so I left this package installed and didn't install libbeagle1 (the name I gave to the package for the 0.3.0 version of the library). I have noticed some problems with the searches: beagle-search only finds files in my home directory, while nautilus finds files in my home directory *and* those listed in the static indexes. beagle-search seems to ignore the static indexes, although these are listed by beagle-index-info. Thats pretty weird! Are you sure about this ? What happens when you query using beagle-query ? Do a $ tailf -f ~/.beagle/Log/current-Beagle, then start a query in beagle-search - what gets printed in current-Beagle ? Another problem is related to beaglefs that doesn't work at all. When I run it with this command: libbeagle1 is needed for beagle-0.3.x. libbeagle0 might work in some cases but not guaranteed. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: beagle problem
However some other programs (nautilus and beaglefs) depend on libbeagle0 (i.e. the 0.2.18 version of the library) so I left this package installed and didn't install libbeagle1 (the name I gave to the package for the 0.3.0 version of the library). I have noticed some problems with the searches: beagle-search only finds files in my home directory, while nautilus finds files in my home directory *and* those listed in the static indexes. beagle-search seems to ignore the static indexes, although these are listed by beagle-index-info. Thats pretty weird! Are you sure about this ? What happens when you query using beagle-query ? Do a $ tailf -f ~/.beagle/Log/current-Beagle, then start a query in beagle-search - what gets printed in current-Beagle ? This is the output of searching for gfortran that is in the static indexes (6 hits) and in my homedir (4 hits). I searched with both beagle and nautilus. 20071214 01:09:56.3106 05816 Beagle DEBUG: Parsed query 'gfortran' as text_query 20071214 01:09:56.3111 05816 Beagle DEBUG EX: Caught an exception sending Beagle.SearchTermResponse. Shutting down socket. 20071214 01:09:56.3111 05816 Beagle DEBUG EX: System.IO.IOException: Write failure --- System.Net.Sockets.SocketException: The socket has been shut down 20071214 01:09:56.3111 05816 Beagle DEBUG EX: at System.Net.Sockets.Socket.Send (System.Byte[] buf, Int32 offset, Int32 size, SocketFlags flags) [0x0] 20071214 01:09:56.3111 05816 Beagle DEBUG EX: at System.Net.Sockets.NetworkStream.Write (System.Byte[] buffer, Int32 offset, Int32 size) [0x0] --- End of inner exception stack trace --- 20071214 01:09:56.3111 05816 Beagle DEBUG EX: 20071214 01:09:56.3111 05816 Beagle DEBUG EX: at System.Net.Sockets.NetworkStream.Write (System.Byte[] buffer, Int32 offset, Int32 size) [0x0] 20071214 01:09:56.3111 05816 Beagle DEBUG EX: at Beagle.Util.UnclosableStream.Write (System.Byte[] buffer, Int32 offset, Int32 count) [0x0] 20071214 01:09:56.3111 05816 Beagle DEBUG EX: at System.IO.BufferedStream.Flush () [0x0] 20071214 01:09:56.3111 05816 Beagle DEBUG EX: at System.IO.StreamWriter.Flush () [0x0] 20071214 01:09:56.3111 05816 Beagle DEBUG EX: at System.Xml.XmlTextWriter.Flush () [0x0] 20071214 01:09:56.3111 05816 Beagle DEBUG EX: at System.Xml.Serialization.XmlSerializer.Serialize (System.Xml.XmlWriter writer, System.Object o, System.Xml.Serialization.XmlSerializerNamespaces namespaces) [0x0] 20071214 01:09:56.3184 05816 Beagle DEBUG: Parsed query 'gfortran' as text_query 20071214 01:09:56.3996 05816 Beagle DEBUG EX: Caught an exception sending Beagle.HitsAddedResponse. Shutting down socket. 20071214 01:09:56.3996 05816 Beagle DEBUG EX: System.IO.IOException: Write failure --- System.Net.Sockets.SocketException: The socket has been shut down 20071214 01:09:56.3996 05816 Beagle DEBUG EX: at System.Net.Sockets.Socket.Send (System.Byte[] buf, Int32 offset, Int32 size, SocketFlags flags) [0x0] 20071214 01:09:56.3996 05816 Beagle DEBUG EX: at System.Net.Sockets.NetworkStream.Write (System.Byte[] buffer, Int32 offset, Int32 size) [0x0] --- End of inner exception stack trace --- 20071214 01:09:56.3996 05816 Beagle DEBUG EX: 20071214 01:09:56.3996 05816 Beagle DEBUG EX: at System.Net.Sockets.NetworkStream.Write (System.Byte[] buffer, Int32 offset, Int32 size) [0x0] 20071214 01:09:56.3996 05816 Beagle DEBUG EX: at Beagle.Util.UnclosableStream.Write (System.Byte[] buffer, Int32 offset, Int32 count) [0x0] 20071214 01:09:56.3996 05816 Beagle DEBUG EX: at System.IO.BufferedStream.Flush () [0x0] 20071214 01:09:56.3996 05816 Beagle DEBUG EX: at System.IO.StreamWriter.Flush () [0x0] 20071214 01:09:56.3996 05816 Beagle DEBUG EX: at System.Xml.XmlTextWriter.Flush () [0x0] 20071214 01:09:56.3996 05816 Beagle DEBUG EX: at System.Xml.Serialization.XmlSerializer.Serialize (System.Xml.XmlWriter writer, System.Object o, System.Xml.Serialization.XmlSerializerNamespaces namespaces) [0x0] 20071214 01:10:14.2454 05816 Beagle DEBUG: Parsed query 'gfortran' as text_query It still does not make sense :( I am surprised nautilus actually worked since it uses incompatible libbeagle0, but I am more surprised that nautilus worked and beagle-search did not! You built and installed everything correctly - right ? Do you get all the results when you do $ beagle-query gfortran -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: beagle problem
I've made some other test, and I've seen that beagle-search does find files in the applications static index, but fails in the documentation static index. beagle-query and nautilus work fine. Oh ... beagle-search explicitly excludes documentation index while searching. The documentation index is used by yelp (gnome help browser). Apparently documentation results in beagle-search was confusing users and so it was blacklisted from general search. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Working With Conduit
Which two birds ? I mean, I don't understand the goal here. From the link in your email, I understand conduit is a synchronization program for user data across machines. Roughly, at least. So where does beagle do in this picture ? Can you elaborate on that a little ? It's quite simple really. What does a beagle indexer do? It gets data out of an application, puts it into a format it understands and fires it at the index. What does a Conduit plugin do? One part is getting data out of an application, putting into a format the sync engine understands and fire it at the sync engine. What i'm offering is the ability to index data from any source that conduit supports as a sync source. KKubasik was also interested in the mapping data - oh, this photo came from flickr account foo bar, and this one from a digital camera. Ok. If I understand you correctly, you are proposing some kind of Conduit backend (or meta-backend if you will), which supplies data to beagle from all the different sources that Conduit handles. I think it is possible to do something like this. Either by using libbeagle or python bindings to send data to beagle from Conduit or pulling data from Conduit from beagled itself. Care has to be taken because Conduit and beagle might not be started at the same time, but I dont see logically this can be done. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Advisory: beagle-0.3.0 crashes at start
Problem: For beagle-0.3.0, beagled crashes when started. The stacktraces look like (once for each backend): Error: Caught exception while instantiating Files backend System.Reflection.TargetInvocationException: Exception has been thrown by the target of an invocation. --- Mono.Data.Sqlite.SqliteException: Sqlite error no such table: textcache_data at Mono.Data.Sqlite.Sqlite3.Prepare (System.String strSql, Mono.Data.Sqlite.SqliteStatement previous, System.String strRemain) [0x0] at Mono.Data.Sqlite.SqliteCommand.BuildNextCommand () [0x0] --- End of inner exception stack trace --- Fix: Delete ~/.beagle directory. Also, r4251 fixes the problem. Explanation: The format of all data stored in ~/.beagle changed from 0.2.x to 0.3.0. It was ensured that all the different kinds of old data were purged or upgraded to the new format. Unfortunately, this check was missed for the TextCache.db textcache data. As a result, users who were using beagle-0.2.x with sqlite3 will see beagled crashing when trying to start the backends. This will not happen for users running sqlite2 earlier, since their database will be automatically deleted. Since almost all of beagle-0.2.x data is incompatible with beagle-0.3.0, all of them are anyway deleted when beagled is started; except config files, which are modified according to the new format. If there are no config files in ~/.beagle/config, deleting ~/.beagle will not cause any additional loss of data. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Install fails
How about I use System.Environment.MachineName instead of Mono.Unix.UnixEnvironment.MachineName ? Seems to work with long names too... (I tried with a 24-char name). Yeah, let's do that. S.E.MachineName just calls gethostname() in an internal call in the runtime. M.U.UE.MachineName calls the same thing, but does it using P/Invoke and apparently doesn't do it correctly. So even if it worked, the results would be exactly the same. Done. r4252. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Fwd: beagle trunk's e-d-s backend does not work with evo-sharp 0.14.0.1
Anyone knows anything about this ? Confirm or deny or how to make it work ? Is it yet another API change in Evo-sharp ? -- Forwarded Message -- Subject: beagle trunk's e-d-s backend does not work with evo-sharp 0.14.0.1 Date: Sunday 18 November 2007 Compiles just fine, but it doesn't work with it. I've got calendar events and contacts, but the webinterface shows EvolutionDataServer : 0 items, Currently not indexing This after running beagled with BEAGLE_EXERCISE_THE_DOG=1 which caused 1220 index items to be added from the evolution mail backend. beagled debug messages show the following: Warn: Unable to open addressbook file:///home/x/.evolution/addressbook/local/system: GLib.GException: EBookStatus returned 20 at Evolution.Book.Open (Boolean only_if_exists) [0x0] at Beagle.Daemon.EvolutionDataServerQueryable.BookContainer.OpenClient () [0x0] Debug: Getting calendar changes for file:///home/x/.evolution/calendar/local/system Debug: Calendar file:///home/x/.evolution/calendar/local/system: 0 added, 0 changed, 0 removed Warn: Unable to open calendar webcal://www.google.com/calendar/ical/[EMAIL PROTECTED]/x/basic.ics: GLib.GException: Authentication required at Evolution.Cal.Open (Boolean only_if_exists) [0x0] at Beagle.Daemon.EvolutionDataServerQueryable.CalContainer.OpenClient () [0x0] Debug: Inbox: Finished indexing 1 messages Warn: Unable to open calendar webcal://www.google.com/calendar/ical/x/x/basic.ics: GLib.GException: Authentication required at Evolution.Cal.Open (Boolean only_if_exists) [0x0] at Beagle.Daemon.EvolutionDataServerQueryable.CalContainer.OpenClient () [0x0] Debug: Getting calendar changes for contacts:/// Debug: Calendar contacts:///: 0 added, 0 changed, 0 removed Debug: Getting calendar changes for file:///home//.evolution/tasks/local/system Debug: Calendar file:///home/x/.evolution/tasks/local/system: 0 added, 0 changed, 0 removed Debug: Getting calendar changes for file:///home/x/.evolution/memos/local/system Debug: Calendar file:///home/x/.evolution/memos/local/system: 0 added, 0 changed, 0 removed Debug: Scanned EDS sources in .39s 0.14.0.1 is the current version used in ubuntu and fedora (and soon gentoo). -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Fwd: beagle trunk's e-d-s backend does not work with evo-sharp 0.14.0.1
Anyone knows anything about this ? Confirm or deny or how to make it work ? Is it yet another API change in Evo-sharp ? After a brief conversation with Varadhan in the IRC, we figured out the problem. Evo-sharp does not (yet) index remote addressbooks and contacts. And the addressbook mentioned below didnt have any local entries either. So, I guess that resolves the problem. If you have remote entries dont get surprised if beagle does not index them. - dBera PS: Apparently there are some tricks about opening the calendar in 'offline mode' but I dont know much about it. -- Forwarded Message -- Subject: beagle trunk's e-d-s backend does not work with evo-sharp 0.14.0.1 Date: Sunday 18 November 2007 Compiles just fine, but it doesn't work with it. I've got calendar events and contacts, but the webinterface shows EvolutionDataServer : 0 items, Currently not indexing This after running beagled with BEAGLE_EXERCISE_THE_DOG=1 which caused 1220 index items to be added from the evolution mail backend. beagled debug messages show the following: Warn: Unable to open addressbook file:///home/x/.evolution/addressbook/local/system: GLib.GException: EBookStatus returned 20 at Evolution.Book.Open (Boolean only_if_exists) [0x0] at Beagle.Daemon.EvolutionDataServerQueryable.BookContainer.OpenClient () [0x0] Debug: Getting calendar changes for file:///home/x/.evolution/calendar/local/system Debug: Calendar file:///home/x/.evolution/calendar/local/system: 0 added, 0 changed, 0 removed Warn: Unable to open calendar webcal://www.google.com/calendar/ical/[EMAIL PROTECTED]/x /basic.ics: GLib.GException: Authentication required at Evolution.Cal.Open (Boolean only_if_exists) [0x0] at Beagle.Daemon.EvolutionDataServerQueryable.CalContainer.OpenClient () [0x0] Debug: Inbox: Finished indexing 1 messages Warn: Unable to open calendar webcal://www.google.com/calendar/ical/x/x/basic.ics: GLib.GException: Authentication required at Evolution.Cal.Open (Boolean only_if_exists) [0x0] at Beagle.Daemon.EvolutionDataServerQueryable.CalContainer.OpenClient () [0x0] Debug: Getting calendar changes for contacts:/// Debug: Calendar contacts:///: 0 added, 0 changed, 0 removed Debug: Getting calendar changes for file:///home//.evolution/tasks/local/system Debug: Calendar file:///home/x/.evolution/tasks/local/system: 0 added, 0 changed, 0 removed Debug: Getting calendar changes for file:///home/x/.evolution/memos/local/system Debug: Calendar file:///home/x/.evolution/memos/local/system: 0 added, 0 changed, 0 removed Debug: Scanned EDS sources in .39s 0.14.0.1 is the current version used in ubuntu and fedora (and soon gentoo). -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Giant logfile again
The remaining lines are only messages like shown in the last line: Skipping over finished thread 1 of 1: EHT 05944 ... rocketed the log file size up to 20GByte in less than 2 hours. Its an extremely rare case which I noticed about a month ago. I checked in a preventive measure (I dont quite know why it happened) in r4055. I merged the change to the 0.2.x branch too but it was too late for 0.2.18. It should not happen anymore. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Opera backend for Beagle
warnings such as: 2007-11-16-18-01-19-Beagle:20071116 18:09:26.3378 21150 Beagle WARN: No timestamp on http://leprosorium.ru/comments/288395! These didnt get fixed after you synced with trunk ? Webpages indexed using the newer version should have the timestamp. -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Nautilus search, memory usage, hidden files, webinterface
guessing the problem was caused by some error in my user Daemon.xml and/or FilesQueryable.xml causing the system-wide configuration files not to be read. They have been carried through since a long time ago, so perhaps somewhere along the line an update to beagle-settings have been incompatible with the previous version and caused the corruption??? Good to know that it works now. You are quite correct in the above though, there were several copy-paste (*evil*) errors which would generate incorrect (even worse, incompatible) config files. Unfortunately, if there are already user config files overriding global ones, the global options dont get read - so once you have a wrong user config file generated, its hard to get it fixed without deleting the incorrect file. One of the reaons (actually, the only reason) I asked people to delete ~/.beagle if they were using trunk from before :) But all is good now, so nothing to worry. BTW, in the list of debug tips I mailed yesterday, the heap-shot tip is last resort and requires extra effort. The tips preceeding it are easy to perform and if anything is found there, easy to fix too. I am dearly hoping that you find some open file or some weird exception or a long list ot nautilus emblem tasks or something like that ... Analyzing heap-shot is very powerful but pretty cumbersome :-/ Thanks, - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Nautilus search, memory usage, hidden files, webinterface
3. Beagle indexes dot-files- and directories. Is this a new feature? Why? Was there a discussion on the list about this? Are you sure about this one ? dot-files and directories should not be indexed! Do you get those in search results ? Pretty much so. I thought it was very strange. For obvious reasons I had to create an exclude pattern for ~/.beagle. Still I suspect this causes some havoc for me. [EMAIL PROTECTED]:config$ beagle-query i7sfifd9.default Debug: Done reading conf from /home/bh/.beagle/config/Daemon.xml Debug: Done reading conf from /etc/beagle/config-files/Daemon.xml file:///home/bh/.bash_history file:///home/bh/.mozilla/firefox/i7sfifd9.default ... You are using the test tarball or latest svn trunk right ? And you did do a make install after you built ? One of the earlier config files had an error which was not excluding the dot files and directories - but that was fixed later. I have a feeling that somehow the old file is still in effect in your case. Could you do a fresh make install, remove/move your current config and then check ? $ beagle-config FilesQueryable ExcludePattern - ExcludePattern : (Exclude files matching these patters) Parameters: [Pattern] Values: - [.*] - [*~] - [#*#] ... The first one should be there... - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Open Mails with Thunderbird does not work
Beagle says this if I want to open a Mail: Unable to run : Failed to execute child process (No such file or directory) I don't use the thunderbird from debian, I have the original thunderbird installed in /opt/thunderbird. beagle-search looks for a file called thunderbird or mozilla-thunderbird in the directories in the environment variable PATH. Is the directory containing thunderbird exe present in your PATH ? -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Opera backend for Beagle
Okay, here's some output from beagled --fb --debug --backend Opera You did --fb or --fg ? --fg would run beagle on the foreground and print both the logs from beagled and indexhelper. Debug: Starting backend: 'Opera' Debug: Starting Opera history backend Debug: Opera history backend done in 0,205798s Debug: Finished thread EHT 19329 [19310 BeagleDaemon] Beagle.Daemon.OperaQueryable.OperaQueryable:StartWorker Warn: Couldn't store file attributes for /home/rikz/.opera/cache4/opr00UTD Debug: Caught ResponseMessageException: Connection refused Debug: InnerException is SocketException -- we probably need to launch a helper Debug: Launching helper process Debug: IndexHelper PID is 19331 Always: Starting Index Helper process (version 0.3.0) Always: Running on Mono 1.2.5.1 Always: Extended attributes are supported Always: Command Line: /usr/lib/beagle/IndexHelper.exe Debug: Found IndexHelper (19331) in 1,00s Warn: Unable to filter http://forums.gentoo.org/viewtopic-t-590705.html. /home/rikz/.opera/cache4/opr00UTD appears to have disappeared from underneath us Does this log give us any importaint info? Can I make beagled a little bit more verbose on what it is currently indexing and so on? I've visited lots of pages, but beagled was silent about them, even when they were indexed successfully. The indexhelper log (or if you start with --fg, then the output to the terminal) will show that webpages are being indexed. Maybe you can track that and see if there is anything there ? The msg xxx appears to have disappeared from underneath us is fine once in a while, it means opera removed the cache file for some reason even before beagle could get to it. But it should not happen always. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
webinterface updates
Hey folks, I just wanted to update you with the recently added features to the beagle webinterface. Its kind of ready for first release now. * Access it at http://localhost:4000/ * Directly search for foo bar by visiting http://localhost:4000/?search=foo+bar (so, you can create search-plugins, bookmarks or link from other places) * Groups the results into common categories like documents, images, im logs etc. * Shows all properties and uses nice property names (e.g. Author instead of fixme:author) * Shows snippets (on demand, to reduce load on machine) * Shows full text for emails, right in the browser itself. (On demand.) * For all the displayed properties, when hovering on them a clickable link is displayed which can be used to search for that property name and value. * Shows beagle-status and presents option to shutdown beagle. * Link to help (local link) * By adding some user_pref (explained in the 'Help' link), the search results are clickable. The results will be handled by the browser. * Web2.0 webapp :) - so usable in browsers that are web2.0 friendly (konqueror does not have xslt support :-/) * Practically no load on beagled, the entire UI is done in the browser and data obtained from beagled by mimicking BeagleClient API messages (!) Quick way to find the answer to the life, the universe and everything :) Hope you like the web interface useful. Suggestions welcome. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
test beagle-0.3 pre-release
Finally! 0.3 seems one step closer now. I made pre-release tarballs for beagle and libbeagle. You can get them from here: http://kubasik.net/beagle/beagle-0.3.0.tar.gz http://kubasik.net/beagle/libbeagle-1.0.tar.gz We would really appreciate if some of you could try this and let us know about any stopper bugs. If people can provide packages for the common distributions, then it might be easier for a lot of other people to test. Note that libbeagle is now distributed separately. Also the earlier libbeagle-0.0 will not work with beagle-0.3. Thanks in advance. Nearly every codepath underwent some change so just run it the usual way you do. Also this release has API and ABI changes from the 0.2.x series. If you are running any of the 0.2.x versions, everything should be automatically upgraded. If you are running trunk, you might want to erase ~/.beagle. Some of the new features that could get some testing, - beagle-settings and beagle-config. Besides there are now global configuration files at /etc/beagle/config-files/ - web interface (WebInterface option has to be enabled; either use beagle-settings or beagle-config Networking WebInterface true) at http://localhost:4000/ The webinterface is minimal but usable (and useful too). We are working on it to add standard features to it. See web-interface-help for how to open the result links - New tex filter, opera webhistory backend. ... the music filter should now index lot more tags and the image filter should now index several iptc tags too. Firefox extension to index firefox webhistory. Less troublesome Thunderbird extension. Nautilus metadata is also indexed. - beagle-search got some UI touches as well. I know some of you have comments about some design issues in the UI but for the time being, please focus on crashes/incorrect data. BTW, there is a bugzilla bug if you want to comment on the UI design of beagle-search. - Some additions to query syntax - network search (I am not sure how stable this is but from my experience with the webinterface this should work too) - the usual suspects - indexing some file takes 100% cpu for a long time, beagled/index-helper not killed after session logout, stale temporary files in ~/tmp etc. That reminds me, some OpenSuSE 10.3 users reported that certain emails hang gmime (basically means those emails wont be indexed and indexhelper will hang on that file) - http://bugzilla.gnome.org/show_bug.cgi?id=485005 Things to keep in mind: - the python and libbeagle apps out there _might_ not work (if they dont just work, then they need some renaming and tweaking). Specifically, kerry works (you have to tweak the build scripts to get it built with libbeagle-1.0). But the settings-gui with kerry and beagle-kontrol (that comes with kbeaglebar) should not be used. They use the old style config files. - I checked the latest shared-mime-info that beagle uses to detect mimetypes of files. The magic for the office types (doc/ppt/xls) are incomplete; so several valid office files will probably be misrecognised and ignored. - If you have a torrent download directory, its best to add it as an excluded subdirectory; some torrent clients continuously write/close the file as they receive data causing reindexing of the same file over and over. - Enabling avahi during ./configure will cause beagle to publish its search service via avahi and those machines can be searched remotely. Avahi is disabled by default because some of use experienced stability problems with avahi-sharp. Thats all for now. Please test the above tarballs and let us know if anything is seriously wrong. If things go smooth, we expect to get 0.3 out in about a week from now. Thanks in advance for your co-operation, - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: beagle-extract-content question: PDF docs
I've got a few PDF docs where beagle cannot find any contents in it: beagle-extract-content gebackene.zucchini.pdf Does pdftotext -q -nopgbrk -enc UTF-8 gebackene.zucchini.pdf output anything ? -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Opera backend for Beagle
Kevin, http://forums.gentoo.org/viewtopic-t-590705.html here we have a rare word LiveUSB. I just visited this page with Could you give this a try sometime ? I dont have Opera installed but I browsed the above webpage using firefox and beagle was able to successfully return the page when I queried for liveusb. Most probably a bug in the opera backend. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: How does beagle get stopped?
Hi Max, I took a look at the code and that is exactly what happens. But beagle --fg does not seem to return. I use it in a script that waits for beagled to return to stop the repository afterwards. This works fine with beagle-shutdown. However it does not seem to work when logging out /shutting down. The log at 9 prints the following: Starting beagle and watching it... Always: Starting Beagle Daemon (version 0.2.18) ... Debug: Lost our connection to the X server! Trying to shut down gracefully Always: Shutdown requested Debug: Stopping inotify threads Debug: Server '/home/max/.beagle++/socket' shut down Debug: All workers have finished. Exiting main loop. Debug: Xlib is forcing us to exit! Debug: Live ExceptionHandlingThread: EHT 14186 [14106 BeagleDaemon] Beagle.Util.Inotify:SnarfWorker EOF Any ideas why this does not return while beagle-shutdown does? Yesterday I found out a few places where index-helper/beagle could get caught during shutdown and not exit. That prompted me to re-read your email. If you have already figured it out, just ignore. The last line in the log above says - Debug: Live ExceptionHandlingThread: EHT 14186 [14106 BeagleDaemon] Beagle.Util.Inotify:SnarfWorker which means there is this SnarfWorker inotify thread that is still running, which is causing the process to not exit. SnarfWorker is a thread used in our Inotify handler to get inotify events asynchronously. During shutdown (after stopping inotify threads) that thread is killed. I have never seen that thread to hang before, but I generally dont run with Xss enabled (i.e. my beagled does not monitor X to figure out when to quit; I run it from a terminal). So its possible there is some issue there that I am not aware of. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Two ideas - up for adoption
Hey folks, If you are looking for 'simple' ideas to break into beagle development or just keeping yourself busy ;-), here are two ideas that might interest you. (1) GMail indexing: Figure out (reverse-enginneer ?) how gmail-desktop-search for linux indexes GMail emails. There is no public API and the usual gmail apis on the web are not search friendly. It should be an easy step of dumping the internet traffic when a search is performed followed by a harder step of finding out what the dump means. (*) All of the above assuming, Google doesn't download the emails using POP/IMAP and then index them. (2) Index Audio CDs: Build a beagle-build-index like beagle-index-audiocd tool (could even be some switch --audio-cd to beagle-build-index) that will fetch the CDDB or other relevant information from the web and index them. StaticQueryable needs to be modified a little to know that some results could be from audio cds, in which case do the right thing (e.g. return the result and the interface prompts the user to insert the audio cd with the specified name).(**) Don't hesitate to ask the list or me personally if you have any questions. In the mean time, wait a little bit more as I plan to put up a testing beagle-0.3-RC tarball in a week or so. - dBera (*) It might not be feasible to distribute the gmail indexer with beagle, but I am pretty sure it will become highly popular and useful through 'unofficial' means. Maybe it will even force Google to open the API. (**) This is very much like indexing-removable-media, which I described on the ML sometime back. You are welcome to work on it too. -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Opera backend for Beagle
Kevin, It was tested against the initial Opera 9.0 release, it really just needs some hardening, its not very stable and tends to go ape when it can't process a file properly. Anyways, all it needs is 20 minutes of love to bring it into the same ballpark as other backends, I'll look into updating it for the 0.3.1 release. That makes me nervous :) Should it be included in 0.3 ? I didnt realize that it was never released since it was added - so basically it never got enough testing. One thing we can do is to add it to the list of excluded backends in the global config. Users who want to use it can then change it locally. But if its not that ready, I would really like to disable it altogether. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: beagles eating my /var filesystem
Either there is some error in beagle-manage-index which could be revealed if you remove the redirection to /dev/null (along with --enable-deletion) or the index _is_ legitimate. Can you rerun with the redirection removed and attach the log (It should not be too large since nothing would have changed in the documentation directories). 520 MB index data looks a bit large though. But it is for 37090 files and it is not thoroughly improbable since the documentation files are mostly text data. Do you have a lot of files that are supposed to be indexed in the documentation index ? I ran the same on my machine (with 0.2.18): 285MB in the documentation directories specified in the crawl-documentation. Beagle made a index of size 27MB. I did some investigation (aka google search) and it looks like some error in our interaction with lucene. For some reasons, old files are not getting deleted or optimization (which automatically happens at the end of indexing) is failing silently! I would suggest deleting the documentation/ directory and re-creating the index. In the mean time, I am digging into lucene to see what could have caused this and how can this be prevented. Always: Starting beagle-build-index (pid 27679) at 28/10/2007 4:11:01 PM Debug: Set best effort IO priority to lowest level (7) Debug: Reniced process to 19 Debug: Loaded 284 records from /var/cache/beagle/indexes/applications/FileAttributesStore.db in 0.004s Debug: Starting IndexWorker Debug: Size: VmRSS=11.9 MB, size=1.00, 0.0% Debug: Flushing driver, 30 items in queue Debug: -file:///usr/share/applications/screensavers/distort.desktop Debug: -file:///usr/share/applications/screensavers/galaxy.desktop ... This run looks fine. It is probably worth noting that I always run the Ubuntu development version on the machine so package churn can be quite huge. Is garbage collection happening? i.e. when a documentation file disappears because the package is upgraded/removed is will the beagle index items be cleaned out? Thats with the --enable-deletion switch. It should be added to the crawl-scripts. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: beagles eating my /var filesystem
Always: Starting beagle-build-index (pid 27679) at 28/10/2007 4:11:01 PM Debug: Set best effort IO priority to lowest level (7) Debug: Reniced process to 19 Debug: Loaded 284 records from /var/cache/beagle/indexes/applications/FileAttributesStore.db in 0.004s Debug: Starting IndexWorker Debug: Size: VmRSS=11.9 MB, size=1.00, 0.0% You ran it against the application index, maybe the documentation index shows some error :) But I doubt it will, though. It is probably worth noting that I always run the Ubuntu development version on the machine so package churn can be quite huge. Is garbage collection happening? i.e. when a documentation file disappears because the package is upgraded/removed is will the beagle index items be cleaned out? Oh, one more thing. If you havent deleted the old index yet, you can do a (as any user) $ beagle-dump-index --uris --indexdir=/var/cache/beagle/indexes/documentation/ this will dump all the uris (11K for me) of the items in the index. You might be able to spot some anomaly there. I remember you already ran with --enable-deletion once, so there should not be any deleted directories. The --enable-deletion did not receive much testing, so there is some possibility of a bug there. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: beagles eating my /var filesystem
Quickly, whats the content of the file documentation/PrimaryIndex/segments ? (Its a binary file so attach it) -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Exclude patterns and subdirectories
(This is more relevant to distributions and sysadmins) Earlier beagle had a hardcoded list of file and directory name patterns (Makefile.in, .*, *.o etc) and directories (~/tmp) to ignore. There was a way to add more patterns and directories to the ignore list but nothing to enable indexing of a file with any of the default excluded patterns, e.g. Also, build-index was not using the full ignore list as that used by the file system backend. With the new configuration system, the list is now configurable. The global default is in the config file $SYSCONFDIR/beagle/config-files/FilesQueryable.xml Add or remove values at will. Note that for subdirectories to exclude, environment variables can be used (e.g. the default excluded subdirectories included in the FilesQueryable.xml are $HOME/tmp and $BEAGLE_HOME/tmp). BuildIndex now shares the same exclude pattern list as well. Users can further override the behaviour to their liking by using the beagle-settings GUI or beagle-config. If you can test the system and help us iron out any remaining bugs in the new config system, that will be appreciated. Thanks. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
webinterface enable: howto
Hi, There has been several more changes to the web-interface included in the trunk. Namely, the static html page is removed and all html is generated from an xml file and xml data (+xslt + css). The UI looks much better. It also provides a way to view information about the beagled process and shutdown beagled. The hits are now categorized and the categories can be shows/hidden by checkboxes. And most importantly it has a new (based on beagle-project logo) logo :-). Enough advertising. The webinterface is now available in trunk (with the files served by the server is in the location pointed to by BEAGLE_WEBSERVER_DIR) but disabled by default. There are two config options related to this: 1) config:Networking - option:ServiceEnabled = turns on/off network search. WebInterface is turned on at the first available port after 4000. This has to be set before beagled is started. Local beagled (thus the webinterface too) can be queried from anywhere in the network (http://ip_address/hostname:4000). $ beagle-config Networking ServiceEnabled 2) config:Networking - option:WebInterface = turns on/off webinterface, only if ServiceEnabled is false. This can be set/unset even when beagled is running and the webinterface will be accordingly started or stopped. The webinterface is only accessible from localhost (http://localhost:4000). $ beagle-config Networking WebInterface Due to a couple of mono bugs (fixed and to be available in mono-1.2.6), even when the web interface is apparently stopped (and trying to access http://localhost:4000 results in a failure), beagled keeps on listening on port 4000. This is harmless since nothing can be accessed via that port (but still an open port, so you should know this fact). And mono HttpListener listens on all interfaces - but again this is harmless, because illegal access is denied (e.g. only setting WebInterface to true causes beagled to listen on port 4000 on all interfaces, but only queries from localhost are allowed). The hits displayed in the webinterface are clickable links. On firefox 1.5, clicking the links to open in a new tab causes the file to open; on firefox 2.0 and above, opening of local files is completely disallowed due to security reasons. There are user configurable options to override this and allow the user to click on hits and open results - we will describe them in detail sometime soon. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: beagles eating my /var filesystem
# du -xk /var | sort -n ... 520054 /var/cache/beagle/indexes/documentation/PrimaryIndex Thats scary. Can you do the following diagnostic test ? (Take a note of the situation) $ beagle-manage-index /var/cache/beagle/indexes/documentation/ info $ du -xk /var | sort -n $ ls -lta /var/cache/beagle/indexes/documentation/ $ beagle-manage-index /var/cache/beagle/indexes/documentation/ optimize (take a note of the situation again) then run the command to build the documentation index but with --enable-deletion to beagle-build-index (take a note at this point) And tell me if anything changes. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Kerry: is it maintained?
Hi, I have been using Kerry for access to beagle through KDE for quite some time. I have also been building beagle from SVN. I upgraded to KDE 3.5.8 and needed to rebuild Kerry but it fails because it needs = libbeagle 0.2.5 and libbeagle 1.0.0 apparently doesn't make configure happy. I was looking for Kerry SVN but it is no longer in the KDE repo. I bumped the libbeagle version partly due to that. The kcm module of kerry (the settings GUI) wont compile with svn trunk. And there are some other API changes (some deprecated methods were removed) but those are simple to change. Is Kerry still maintained? If not, is there a good KDE interface to beagle? Any idea how to fix, I've never messed with autoconf scripts. KDE folks are all busy with KDE4. So in short, it is not maintained :( Hopefully once the next feature release of beagle is done, I will have some time to make kerry work with beagle. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: State of the Pooch
Hi Joe, Thanks a lot. I am delighted to be awarded the new responsibility. It has been a pleasure to work on beagle this far and I am sure it will be even better in the future. Work continues in trying to make a great 0.3.0 release, and in the meantime we're pushing out 0.2.x maintenance releases. I'd love it if people could be regularly running from SVN trunk so that we can stress test a lot of the features that I'll mention below and get a 0.3.0 release out there that the less adventurous users out there can enjoy. For others, there is no more new feature planned for 0.3.0. Lukas is giving some finishing touches to beagle-search and Nirbheek is working on making the webinterface a bit smoother experience (*). But there is no likelyhood of any changes to the core nor any completely new feature. Sort of like feature freeze but not quite. - dBera (*) e.g. I just came to know that even though my Firefox 1.5 allows me open local links, Firefox 2.0 explicitly deny that. We are thinking of other ways to open local files from the webinterface. -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Kerry: is it maintained?
KDE folks are all busy with KDE4. So in short, it is not maintained :( Hopefully once the next feature release of beagle is done, I will have some time to make kerry work with beagle. Lets hope, I too am using Kerry (Kubuntu) and would like to take advantage of the new features of Beagle 3.0 Kerry has been moved to extragear: http://websvn.kde.org/trunk/extragear/utils/kerry/ I had a checkout of kerry and I built it against latest beagle - it builds fine after I change (in configure.in) the libbeagle version from 0.2.x to 1.0 and libbeagle-0.0 to libbeagle-1.0. You might want to try that ... but WARNING! *do not* use the kerry gui for settings (which is implemented as a kcm module i.e. you can also find it in control center) - it might run correctly but is incompatible with the new config system. For configuration of beagle, use the command line tool beagle-config (or the gtk gui that comes with beagle, beagle-settings). - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: HTML mimetype
And no, beagle's HTML filter does not index application/x-mozilla-bookmarks file. Its trivial to add the mimetype to the HTML filter but I wonder if that is the right thing to do. Till this issue is resolved, don't be surprised if your html files are not indexed! The problem is partly due to shared-mime-info, so anybody with shared-mime-info-0.22 [1] will face the same problem. Anyone knows anything ? Found this 2 month old bug -- https://bugs.freedesktop.org/show_bug.cgi?id=11843. Aha - right. Further investigation reveals that the problem is due to a commit for the bug https://bugs.freedesktop.org/show_bug.cgi?id=266. IMO the fix for #266 is incorrect, but as always, it takes months and years to get attention of xdgmime/shared-mime-info developers. Reverting that particular change (mentioned in 266#5) fixes this problem. Tell your distribution to enable HTML filtering by temporarily reverting that change http://webcvs.freedesktop.org/mime/shared-mime-info/freedesktop.org.xml.in?r1=1.245r2=1.246 - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Exclude patterns in beagle-settings
Hi, Most you are aware of exclude patterns in beagle config (set via beagle-config or beagle-settings). Files matching these patterns are not indexed. I noticed that the exclude patterns are not regular expressions but simple patterns like .*, *.tmp, *.o, config.log. The equivalent regular expressions for these would be \..*, .*\.tmp, .*\.o, config\.log. The current values are easy to understand but are limited in power (e.g. if you read the source you will see that *akefile.* won't mean what you want and there are more). I was wondering if we should instead use regular expressions ? The only problem with regexes is that they are harder to parse and even simple patterns can result in a complicated regexes. What do you think ? One option is use regexes as the underlying values but ask the user to input simple patterns in the GUI and translate them to equivalent regexes when storing them. Then we also have to provide an advanced input box for the regex-friendly advanced users :). The whole things gets messy. This is on top of the fact that beagle-settings has not received any attention lately. - dBera PS: Regexes are in general slower, but in this particular case speed won't be a problem. -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Beagle and KTorrent
I've noticed that Beagle reindexes files that are currently in use by KTorrent. Not only those which are incomplete, but also those, which are already downloaded and now they are just seeded. As I understand, while files are seeded they do not change, so why should Beagle reindex them? How did you notice the re-indexing of already downloaded files ? Something in the log files ? In that case can you paste some sample lines, unlikely but there might be something there. Could also be a ktorrent bug ... ? -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Beagle and KTorrent
I noticed it with top. beagle-helper was using CPU. Then I decided to see what exactly does it index now. I used beagle-status for it. There's nothing interesting in there. If you think its important, I'll copy a part of beagle-status output here. No thats fine. I confirmed that with ktorrent, beagle with continuously index data, even those which are already downloaded. Its the same case with azereus and maybe other torrent clients Yes, now I presume that it's more KTorrent bug than Beagle. But I'd really like to hear any suggestions on why this bug exists and if it could be fixed easily by editing ktorrent's way of opening files. Yes it could be fixed if ktorrent (and other apps) did not needlessly open files RW but open them in read-only mode if only reading is needed. This requires some effort in these torrent apps, thus I doubt if they will fix it. Still... its worth a try. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Beagle and KTorrent
So we can presume that there are two available reasons of incorrect (in my opinion) re-indexing of unchanged files: - torrent client doesn't open file read-only and inotify correctly informs beagled about file change. The reason is torrent client bad realization. - torrent client opens file read-only but for some strange reason beagle's inotify watch tells beagled that the file was changed. I used beagle/Util/inotify-test to check what does ktorrent do its already downloaded files. (./inotify-test /path/to/dir/to/monitor) (The following is after the downloading is finished) *** inotify: Open 1 /mnt/extra/files/p2p/Aprar (file) *** inotify: Access 1 /mnt/extra/files/p2p/Aprar (file) *** inotify: CloseWrite 1 /mnt/extra/files/p2p/Aprar (file) *** inotify: Access 1 /mnt/extra/files/p2p/Aprar (file) *** inotify: CloseWrite 1 /mnt/extra/files/p2p/Aprar (file) *** inotify: Open 1 /mnt/extra/files/p2p/Aprar (file) *** inotify: Access 1 /mnt/extra/files/p2p/Aprar (file) *** inotify: CloseWrite 1 /mnt/extra/files/p2p/Aprar (file) I am sure if you check /proc/pid/maps or use lsof, then you will see files are opened read-write even after downloading is over. One reason they might be doing this is just because they are careless. Linux allows other apps to access the file just as easily if it opens the files RW or R. So there is no apparent harm for them; instead they dont need any extra checks if the file is already downloaded or not. I heard rumours that inotify/kernel could be modified to only report CloseWrite for files that are actually modified ... but I haven't read anything concrete about that. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: System.InvalidOperationException: Invalid connection string
This is probably related. Now my beagle log is filling up with: 20071018 00:42:49.9363 09041 Beagle DEBUG: Unable to determine account name for [EMAIL PROTECTED]:993 Pressumably one for each of the bogus folders under /home/brian/.evolution/mail/imap/[EMAIL PROTECTED]:993/folders/cur/subf olders/ Any ideas on how to clean this mess up? I've asked on the evolution list but nobody has responded. Its something to do with the account_names for those folders as stored in gconf. I dont know much about these things ... maybe you can try to check the list at gconf:/apps/evolution/mail/accounts and see if there is any suspicous entry. Could be some bug in the Evolution backend too ... - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: GSoC Weekly Report
A followup question, I didnot find any API documentation of Mono.Data.Sqlite :( #mono was also sleeping when I asked the question there. My understanding is that both M.D.SqliteClient and M.D.Sqlite follow the general ADO.Net API patterns and that the latter is more or less a drop-in replacement for the former. A few things may need to be tweaked, but in general just changing the using statements at the top of each source file should be all that's needed. I was more looking for some method for row-by-row retrieval, on demand. Real on-demand, where the implementation does not retrieve all the rows at once but returns one by one. You've always been able to get rows on demand via ADO.Net, it's just a matter of the implementation underneath. The old one (not modified by us) would load all of them into memory. I'm not sure how the new one performs memory-wise. If the Mono guys don't have any idea, the right I checked the source out of curiousity http://anonsvn.mono-project.com/viewcvs/trunk/mcs/class/Mono.Data.Sqlite/Mono.Data.Sqlite/ And the code for DataReader looks exactly the same (didnt do a diff, just visually) as the one in Mono.Data.SqliteClient. So even if we migrate (the migration would be easy), we still have to ship with a modified inhouse M.D.Sqlite and keep syncing in with upstream. *sigh* - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: GSoC Weekly Report
Ignore my previous email ... I was looking at the wrong place :( This is the right place for the new M.D.Sqlite http://anonsvn.mono-project.com/viewcvs/trunk/mcs/class/Mono.Data.Sqlite/Mono.Data.Sqlite_2.0/SQLiteDataReader.cs - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
HTML mimetype
Hey all, I recently noticed that *.html files are getting detected as application/x-mozilla-bookmarks instead of the correct text/html ! This is due to an xdgmime mime database (shared-mime-info) weirdness which recognizes *.html files as application/x-mozilla-bookmarks. Just for consolation, gnomevfs-info also makes the same mistake. I wonder what does nautilus do ? And no, beagle's HTML filter does not index application/x-mozilla-bookmarks file. Its trivial to add the mimetype to the HTML filter but I wonder if that is the right thing to do. Till this issue is resolved, don't be surprised if your html files are not indexed! The problem is partly due to shared-mime-info, so anybody with shared-mime-info-0.22 [1] will face the same problem. Anyone knows anything ? - dBera [1] http://webcvs.freedesktop.org/mime/shared-mime-info/freedesktop.org.xml.in?revision=1.246view=markup -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers