[00:59] TheMuso, is there a way to search the archive for packages that provide files in a given directory? I want to see what packages but files in /usr/lib/deskbar-applet. packages.ubuntu.com doesn't seem to be able to do this [01:20] robert_ancell: apt-file [01:20] robert_ancell: i.e apt-file is the package that will help you find files in packages that you don't have installed. [03:57] kenvandine, hey, how much do you know about pkg-config? Can you look at bug 455614 and see if you can tell why someone would modify .pc dependencies? [03:57] Launchpad bug 455614 in libsm "don't mess with the pc file" [Undecided,Incomplete] https://launchpad.net/bugs/455614 === robert_ancell_ is now known as robert_ancell [06:49] Good morning === ara_ is now known as ara [07:09] hello al-maisan [07:09] hola didrocks :) [07:43] Good morning [07:58] hi pitti [08:32] morning everyone [08:48] MacSlow: ping? :) [08:48] slomo, ja? [08:50] MacSlow: hi, you know cairo well, and i saw some interesting font rendering from you iirc... what's the best and fastest way to draw only the outline of text, best when using pango? :) [08:50] hey slomo [08:50] hi seb128 [08:50] slomo, thanks for the quick gst fix! [08:51] seb128: np :) are your non-ac3 dvds working too now btw? [08:52] slomo, I didn't try again but I will do that later today and let you know [08:52] but looking on the bug seems it was on good track and somebody else was able to test [08:52] seb128: yes, but it's not optimal yet... it "should" work :) [08:53] slomo, do pango_cairo_layout_path() ... then cairo_stoke() it [08:54] slomo, bug #450946 is the sort of bugs you are interested to look at too? [08:54] Launchpad bug 450946 in totem "requires a restart to use the newly installed codecs" [Low,Confirmed] https://launchpad.net/bugs/450946 [08:55] MacSlow: that's what i'm doing now, unfortunately it's very slow compared to pango_cairo_show_layout()... because pango_cairo_show_layout() does all kinds of caching and stuff that gets lost when rendering the font to a cairo path [08:56] MacSlow: cairo developers said to use user fonts, unfortunately pango doesn't support them yet... so i thought you might know some pango trick to do it faster for now :) [08:57] slomo, you just ask how to do it... not how to make it fast :) [08:58] slomo, also I don't know what you rendering requirements are (fps, fill-rate limits etc.) [08:58] MacSlow: ok, sorry :) so you don't have a better idea either? the final goal is to draw white text with black outline [08:58] hey slomo - seb128 asked yesterday about the stability of fludendo-mp3 compared to the other one [08:59] MacSlow: well, just drawing the outline shouldn't take 60% of all processing time (the 30% include creating a much larger test screen, drawing the normal text, putting a shaded box around it, ...) :) [09:00] slomo, approximate it (but will not look as good)... draw with pango_cairo_show_layout() in black... decrease the size.. draw again with pango_cairo_show_layout() in white [09:00] mvo: it's worse but has a better license (GPL vs. BSD or MIT) [09:00] slomo, how much text do you have to draw and at what frame-rate? [09:01] MacSlow: ok, i guess i have to do that then... and think about how much i should increase the text size. would you change the text size with cairo (cairo_scale) or with pango? [09:01] MacSlow: subtitles on video, so any framerate and a bit text [09:05] slomo, hm... I'd say you should certainly surface-buffer the text you draw... meaning draw it once to a surface-buffer... reuse that as long as the text doesn't change... and only re-render text tothe surface-buffer if new text needs to be displayed [09:06] slomo, I hope you don't call pango_cairo_layout_path() every frame :) [09:07] MacSlow: no, only everytime it changes of course :) my testcase was a bad choice here because the text changes every frame probably ;) [09:07] slomo: thanks, we fixed it recently to have the g-s-t meta-data and now it shows up in the codec installer. but it also sets it self as default if multiple ones are installed. is that safe/sane? or should we instead lower its priority or break it again to not show up at all? [09:08] slomo, you should certainly use surface-buffering... I don't expect subtitle-text to change every 1/24 th of a second :) [09:08] mvo: i don't know, depends on your needs i guess :) [09:09] need: no-crashes ;) [09:09] but thanks for the info I will think about it a bit and discuss with seb128 [09:09] slomo, surface-buffering you certainly will need to do for optimizatino reasons anyway [09:10] mvo: then better use the mad one, it's tested more ;) the only advantage of the fluendo one really is the license (or if you bought it from fluendo the patent license) [09:10] slomo, well, it's basically "what should user get when clicking on a mp3" [09:10] slomo, ok, that's what I though too, it's late for a such change now [09:10] there is a g-c-i bug there [09:11] MacSlow: as said, i do surface buffering :) it's only my testcase that has changing text every buffer (drawing the current time) [09:11] it shouldn't install 3 decoders for the same same thing [09:11] ie for a mp3 is installs the mad, fluendo and ffmpeg binaries [09:11] yes, it should only select the one with the highest priority and let users choose to install the others and not the one with the highest priority [09:13] slomo, ok... then you should not have an issue using pango_cairo_layout_path() and cairo_stroke() [09:14] slomo, actually guess you would want to do pango_cairo_layout_path()... cairo_fill_preserve() in white and then cairo_stroke() in black [09:16] MacSlow: for subtitles it's no problem, yes, but for text that changes every frame it is. that's what i'm trying to explain :) and i really don't want pango_cairo_layout_path(), fill_preserve + stroke... it's slower than show_layout() + layout_path() + stroke [09:16] MacSlow: see https://bugzilla.gnome.org/show_bug.cgi?id=598695 for the context [09:16] Gnome bug 598695 in gst-plugins-base "[timeoverlay] use much CPU ressource compare to previous version" [Normal,New] [09:17] slomo, only solution is a real OpenGL-backend for cairo if you need heavy-duty text-rendering with cairo [09:19] slomo, if you can live with only the typical western set of ASCII characters you could do a "texture-map" character lookup and use that [09:19] slomo, but that would kill all unicode-sexiness [09:19] but that would be fast as hell [09:23] mac_v: thanks for testing the synaptic ppa version - what card are you using when the problem appears? I wonder if its driver dependant. i get about ~8% or so xorg cpu, nothing abnormal high [09:23] with nvidia [09:30] mac_v: is it a intel card/driver as well? [09:31] mvo: ATI x1400 [09:33] kenvandine: let me know [09:33] luckily bindwood wasnt put in main ;) [09:49] davidbarth, hey, any clue about bug #438664? [09:49] Launchpad bug 438664 in empathy "empathy should use same status icons like fast-user-switching applet" [Wishlist,New] https://launchpad.net/bugs/438664 [09:49] davidbarth, not sure if using different icons in the session menu is a design choice [09:50] seb128: we're using icons provided explicitely for this menu; the integration of these icons in IM apps was not defined in a way or another [09:51] ok, not sure what to do with the bug [09:52] is something you want to consider? [09:52] should somebody be subscribed to comment on it? [09:56] ArneGoetje: did you run the last po2xpi run yet? [09:57] or is export running atm? === proppy1 is now known as proppy [10:26] BBL, need to repair a friend's jaunty server [10:26] pitti: before you leave ... did the last langpack run already happen or is export still running? [10:26] (sorry) [10:57] mpt: do you think the changes to the button-have-icons and menus-have-icons should be mentioned in the Ubuntu Karmic release notes? https://wiki.ubuntu.com/KarmicKoala/HumanReleaseNotes [10:58] mpt: thanks for the notice about my blog being blocked btw, seems I had a really old wordpress install, so it got hacked [10:59] * mpt wonders why that URL has "Human" in it [11:00] andreasn, I think that would be a good idea [11:00] who do I ask for that? [11:00] andreasn, it's a wiki page ;-) [11:01] well, I don't want anyone to slap my fingers... ;) === ccooke_ is now known as ccooke [11:06] kenvandine: i pushed the gwibber threading changes i had [11:06] kenvandine: we just got rid of all threads, but the worker thread [11:06] think should be all clean now ... though you cannot say for sure ;) [11:07] gwibber is always good for surprises [11:07] i found a dead function in frontend code that also starts a thread .... havent removed it, but it should definitly go away [11:07] asac: back [11:07] asac: I think it's scheduled for today 2200 UTC (finish) [11:08] asac: so I'll start building them tomorrow [11:09] great [11:10] pitti: does arne do a test run before uploading? [11:10] asac: ArneGoetje is off sick, I'll do them [11:10] asac: I usually do a local build of the German before [11:12] pitti, I though we agreed to standard language to try was french? ;-) [11:12] mpt: looks sane now? [11:13] ok. all ok then. [11:13] andreasn, yep, thanks [11:13] pretty much copied the notes from gnome 2.28 release notes and changed some wording [11:14] is there some files system option that would prevent flock from working? [11:19] andreasn, though once again we're suffering from a problem that we don't know who the target audience is for the release notes [11:19] ivanka, we should figure that out one day [11:20] mpt: yes, indeed [11:35] is gwibber supposed to send replies to facebook messages? because it doesn't seem to work at all for me [11:39] rodrigo_: only if you have it set up to do so in the settings [11:39] czajkowski: I have [11:40] it just fails silently, I write the reply, press ENTER, and wait, but my reply never shows up [11:40] rodrigo_: check if it shows up in fb [11:40] rodrigo_: ah ok I've mine set not to reply to fb, ohnly post otherwise would be noisey [11:41] davmor2: it doesn't [11:41] czajkowski: hmm, I only have 'Receive' and 'Send', no 'post/reply' [11:42] maybe I'm using an old version... [11:42] asac: we just tried to configure a manual eth connection here (no DHCP); I added a new connection type, but that won't actually appear in nm-applet's dropdown list [11:42] asac: I'll just file a bug now, unless you say that's already known? [11:42] (I didn't quite find a related one) [11:42] not known [11:42] i tested various options and didnt get any such feedback [11:43] pitti: what does "i added a new conneciton type" mean? [11:43] asac: s/type// [11:43] yeah [11:43] pitti: killall nm-applet -> nm-applet -> appears? [11:43] asac: connection editor -> edit connections -> cable -> add -> add a static one [11:43] asac: lemme try [11:43] pitti: jockey is showing my nvidia gfx anf bcm wifi on live cd but on installed system I get nothing displayed says there are no drivers for this system. [11:46] asac: confirmed, it's there now [11:46] and even working [11:46] asac: want a bug report for that? [11:47] pitti: please file a bug, yes [11:47] there is something racy [11:47] i always ha the suspicion there is [11:48] \# [11:48] could also be a glib-dbus bug (e.g. signal not delivered) [11:55] bbl [12:16] asac, which function? [12:17] kenvandine: on_oldsearch [12:17] thx [12:17] or something. ... that spawns a thread ... the whole code of this should go away imo [12:18] if its used we want to see that ... as we definitly dont want to run a thread like that ;) [12:18] but i couldnt find any call to on_old_search [12:18] kenvandine: maybe just check if thats still used. i can remove the parts then. [12:18] not sure with all the other on_ ... functions [12:18] maybe those should be reviewed too [12:19] guess not for karmic both though [12:20] seb128, about https://bugzilla.gnome.org/show_bug.cgi?id=593086 which LP bug attached the log file? [12:20] Gnome bug 593086 in General "empathy crashed with SIGABRT in empathy_tp_chat_acknowledge_message" [Major,New] [12:20] kenvandine: i was too dump to figure out where the hell the image tags/css are set [12:20] i didnt look for long [12:20] but thats a nasty thing [12:21] yeah [12:21] imo it should just use the http:// urls [12:21] and let webkit do the caching [12:21] unless webkit doesnt have a cache on its own of course [12:21] but then i couldnt find anything related to those images [12:21] the image cache definitly isnt used anymore ;) [12:21] i removed all images in there and it stays empty [12:21] cassidy, bug #408530 [12:22] Launchpad bug 408530 in empathy "empathy assertion failure: empathy:ERROR:empathy-tp-chat.c:1391:empathy_tp_chat_acknowledge_message: assertion failed: (m != NULL)" [Medium,Triaged] https://launchpad.net/bugs/408530 [12:22] cassidy, it the bug which has been original forwarded and is listed in the first bugzilla comment [12:22] kenvandine: what i found was ./gwibber/microblog/identica.py: self.image = user["profile_image_url"] [12:22] kenvandine: but thats not set anywhere [12:23] seb128, thanks [12:23] cassidy, you're welcome [12:24] asac, that might be the old code which isn't used anymore [12:24] yes. i get the feeling that half of the code is old ;) [12:25] kenvandine: but what code is used? does it use a custom url handler? [12:25] or just webkit ? [12:25] i dumped the html snippet for message to the terminal and it didnt have any sign of a image [12:25] then i stopped looking as i clearly don't understand whats done there ;) [12:29] are you guys going to be pushing any more gwibber fixes into karmic? :) [12:31] Ng: please test trunk [12:31] Ng: bzr branch lp:gwibber; cd gwibber; (stop gwibber UI); killall gwibber-daemon; ./bin/gwibber-daemon + ./bin/gwibber [12:32] Ng: not sure what your issues are though ;) [12:33] kenvandine: btw, you can now ctrl+c gwibber-daemon ;) [12:34] let me check the quit thing now [12:34] asac: bug #383759 is the most immediately annoying bug I have atm :) [12:34] Launchpad bug 383759 in gwibber "bug with font size as int" [Medium,Triaged] https://launchpad.net/bugs/383759 [12:35] Ng: no hangs of daemon? [12:35] Ng: ok that can be taken [12:35] asac: I have been seeing hangs, but this one stops the UI from starting [12:37] asac: \o/ [12:37] Ng: stops the UI? [12:37] wow [12:37] does that throw an exception? [12:38] yeah, I just added a comment to the bug with the traceback === MacSlow is now known as MacSlow|lunch [12:39] asac: I'm running trunk now, will leave it going and see how it goes :) [12:40] Ng: where do you set that font size ... in gnome? [12:41] just wonder if they get a similar problem if you onfigure font size like 12.2px ;) [12:41] rather than point [12:41] asac: yeah, in gnome, so it's a point size. you just type in the size rather than select one from the list [12:42] hmm [12:42] so most likely pixel sized font sizes will make this choke too ;) [12:42] anyway this one should be good [12:42] for its part of the bug [12:43] are pixel sized font sizes legal in a fontconfig world? [12:44] I have no idea how one would configure gnome to use such a thing ;) [12:45] Ng: committed [12:45] asac: sweet, thanks [12:45] i will check that noww with px [12:46] now I just have to persuade MacSlow|lunch to love floating point font sizes in notify-osd :D [12:47] Ng: i have this /apps/gwibber/preferences/default_font [12:47] in gconf-editor [12:47] is that the font in question? [12:47] Ng, I've some local fixes partly based on your branch [12:48] hmm. how broken is notify-osd with font sizes? [12:48] just does "rounding" ... or fails completely? [12:48] asac, can get tiny... but does not crash or anything like that [12:49] asac, but I've to put the change to thorough testing and need to talk to davidbarth too [12:51] MacSlow|lunch: tiny like "rounding" ... or "nulling out"? [12:51] just curious [12:57] asac: I think it relates to /desktop/gnome/interface/font_name [12:58] MacSlow|lunch: I appreciate it :) [12:58] asac: rouding, I think it parses 6.5 as 5 [12:58] as 5 [12:58] ok [12:58] but I suck at coding, so I could have read that wrong [12:58] Ng: it splits by . and uses the second particle? [13:00] asac: I'm not sure, I didn't pay too much attention because my position is that the parsing is unnecessary, but like I said, I suck at coding. The clue being that I'm not paid to write code and you guys are ;) [13:02] Ng: ok. but it parses? [13:02] asac: yeah [13:03] kenvandine: isnt the idea that if you quit gwibber, it usually still runs in background and spits out notifications/indicator stuff? [13:03] just wondered if the -quit is really what we want [13:10] Ng: http://paste.ubuntu.com/299006/ [13:10] thats with 13.333px set manually in gconf [13:11] /gnome/interface/font_name == Sans 13.333px [13:11] ok guess this needs more fixage [13:12] asac: is that a legal value for font_name? [13:12] there seems to be a bit of a lack of exact specification around this stuff, but AIUI they're basically just strings that pango can parse [13:14] yes [13:15] Ng: font_name is just a pango_description [13:15] Ng: try to change it ... your font on desktop should adjust ... some apps are buggy, but all apps usually start [13:16] let me check what websettings can accept for values [13:16] unfortunately there is nowhere a real api doc for webkit ... at least not in my search engine horizon [13:16] * asac getting source [13:20] ok webkit has default-font-size really as an integer for point sizes only [13:21] "Martin Pitt is the heart and sole of the desktop team." <- I don't find Martin leathery at all [13:28] asac, that was the original idea, but there have been complaints about there not being a way to quit the daemon at all [13:28] and people keep saying that should be wired up to the client quit [13:28] we need to think that through in the UI for lucid [13:28] for example, people don't want to be able to hide notifications [13:29] i was thinking that would be cool to do if you set your IM status to away or busy [13:29] something like that [13:29] and make it configurable [13:29] so then to make it "quieter" you just set to busy or away [13:30] asac, but just an idea [13:30] asac, so my quit branch makes the quit button and quit menu item quit the backend as well, but closing the window doesn't [13:31] kenvandine: how about adding that to the indicator ;) [13:31] quit? [13:31] yeah [13:31] ;)( [13:32] i bet there would be resistance to that :) [13:32] hehe [13:32] yeah. i think configuration would be good [13:32] not sure how to best stop gwibber daemon [13:32] asac, anyway, that is a situation we need to give some thought too [13:32] if they dont want it to be in indicator it probably needs to be in tray [13:32] asac, on my list for the next refactoring :) [13:33] and easier now that i understand what interests are in the indicator [13:33] asac, so interesting observation with your threading fixes [13:34] ordering is whacky for multiple services [13:34] for example i have a facebook post that says 11h ago, followed by a flickr image 6d ago followed by a dent 2h ago [13:35] really [13:35] yeah [13:35] and i confirmed that didn't happen in r467 [13:35] asac, it looks like everything from the same service is in the correct order [13:36] but intermingling them isn't right [13:36] like all my dents are in the right order but not properly aligned with tweets and suck [13:36] such [13:36] kenvandine: sureits a regression? i dont see it in my messages [13:36] have twitter/facebook and idnetica [13:36] all properly ordered [13:36] most noticable with flickr since those aren't posted often [13:36] asac, no i just reverted to r467 and they ordered properly [13:37] updated again and they are out of whack again [13:37] thats the last one? [13:37] yeah [13:37] with your curl fixes [13:37] kenvandine: you reverted all? [13:37] hmm [13:37] thought just one commit [13:38] my dents, tweets, and facebook are all nearly in sync but looking closely i see they aren't ordered right [13:38] i mean ... i cannot think of why reall yold stuff should be different [13:38] i wouldn't have noticed if it wasn't for this whacky flickr feed [13:38] the recent tweets, fb, and dents are wrong too i just wouldn't have noticed if it wasn't for flickr [13:38] hmm. my order is a strict order [13:38] just went through all i see [13:38] nothing goes back [13:38] weird [13:39] kenvandine: can you just revert the lightweight OperationManager commit? [13:39] sorry [13:39] operationhandler [13:39] 471 [13:39] no worries [13:39] * kenvandine does it [13:40] asac, oh this might be noteworthy [13:40] i do have one message getting dropped for a bad time [13:40] Dropping message with bad time attribute: {'username': 'Ken VanDine', 'account': , 'client': , 'protocol': 'facebook'} [13:40] Traceback (most recent call last): [13:40] File "/home/ken/src/gwibber/gwibber/microblog/facebook.py", line 39, in __init__ [13:40] sender = profiles[data["actor_id"]] [13:40] KeyError: '41911143546' [13:41] def get_messages(self, filters, count): data = list(self.filter_messages(filters)) data.sort(key=operator.itemgetter("time"), reverse=True) if count == 0: return {"count": len(data), "total": len(data), "messages": data} else: return {"count": count, "total": len(data), "messages": data[:count]} [13:41] not sure if that could mess up the whole order [13:41] thats the code for get_messages [13:41] so its alaways sorted before sending over the wire [13:41] maybe the timestamps are a bit off [13:41] that looks sane [13:41] i wonder if the client is doing anything weird [13:42] * kenvandine thinks there is way too much logic in the client [13:43] kenvandine: if not hasattr(message, "time"): [13:43] print "Dropping message with bad time attribute:", message.__dict__ [13:43] so does flickr set a time attribute? [13:43] it must [13:44] but the dropped one is flickr, and it looks like it has the time as the key [13:44] kenvandine: i dont see the time in the dump from above [13:46] kenvandine: does self.time = parse_time(data["dateupload"]) mean that its also there as self['time'] ? [13:46] no, that would be weird [13:49] i dont see any place where "time" is set [13:49] except one exception handler in microblog/__init [13:49] which sets it to current time [13:55] guess itemgetter is supposed to look at that field === MacSlow|lunch is now known as MacSlow [14:16] kenvandine: can you dump all messages before the sorting? [14:17] kenvandine: also try the last commit for the lightweight handler ... if thats enough [14:24] asac, i am doing something like that, it looks like the one without the time is nearly blank [14:24] like only my data [14:24] no message data [14:24] not even a sender [14:26] kenvandine: in get_messages in init that message shouldnt be there [14:26] would be interesting to dump all the messages send out there [14:26] to see why the time sorting doesnt work [14:27] ok [14:27] i will do that first [14:27] the traceback was bugging me :) [14:30] what i could imagine is that there is a bogus time somewhere [14:30] that can bust up the whole sorting if it breaks strict order properties [14:33] and if thats the case the order how stuff gets inserted could make a difference [14:33] re [14:33] ok, empathy bug list cleaned now [14:34] which probably happened by the idle_handler refactoring [14:38] asac, ccheney, kenvandine, pitti, Riddell, seb128 - what's the word on the street? [14:39] rickspencer3! [14:39] good morning [14:39] rickspencer3, nothing to signal there I would say [14:39] hey rickspencer3, good morning [14:39] things look good [14:39] all quiet here [14:39] except.. OMG!!! [14:39] quiet ... too quiet [14:39] * rickspencer3 eerie music [14:39] rickspencer3, asac did some nice threading fixes to gwibber :) [14:40] kenvandine, lol - just as you said that I was closing a gwibber apport window [14:40] in mozilla we found out that we compile system sqlite with wrong flags :/ [14:40] arg, so need to rebuilt mozilla? is this a security issue? [14:40] no i told them that we cant do that for release [14:40] needs a carefuly SRU [14:40] grrr [14:40] https://bugs.edge.launchpad.net/firefox/+bug/457791 [14:40] Launchpad bug 457791 in firefox "Build sqlite3 with SQLITE_SECURE_DELETE" [Unknown,Confirmed] [14:41] https://bugzilla.mozilla.org/show_bug.cgi?id=445164 [14:41] Mozilla bug 445164 in Private Browsing "Private data not properly cleared from SQLite database" [Major,New] [14:41] i dont feel comfortable to just pump that into our libsqlite without thorough testing. [14:41] its not really breakage of functionality though [14:41] "just" privacy concerns [14:42] seems people there care about gwibber rather than karmic nowadays [14:42] ;-) [14:43] asac, it looks like replies might be messing up the ordering [14:43] at first glance of the data [14:43] * kenvandine needs more coffee.. bbiaf [14:47] rickspencer3: pleasingly good, we got some fresh faces for the ISO testing round and no problems found that we didn't already know [14:47] Riddell, great news [14:47] good buzz in the blogosphere regarding Karmic Kubuntu [14:53] kenvandine: replies have a differnt time format? [14:54] hello [14:54] hi baptistemm [14:54] Hi asac [14:54] is there a way to remove piding from indicator-applet as I don't use it? [14:54] hi baptistemm [14:54] hi chrisccoulson [14:55] baptistemm, disable the indicator option in pidgin or uninstall it [14:55] hey baptistemm btw [14:55] heya [14:55] or ask tedg [14:57] baptistemm: The command line way is "mkdir -p ~/.config/indicators/messages/application-blacklist ; cp /usr/share/indicators/messages/applications/pidgin ~/.config/indicators/messages/application-blacklist" Long paths. [14:58] ah thanks :) [14:58] tedg, do you have a stock reply for it yet? ;-) [14:59] seb128: heh, actually the first time I've been asked :) [14:59] I just doubt someone could use both but I'm certainly wrong [15:00] I use both [15:00] I'd tend to think using one should remove the other :) [15:00] or rather I try to use empathy but it annoys me I go back to pidgin [15:00] so I'm wrong [15:00] +when [15:01] baptistemm: Yes, but the solution there is that we shouldn't INSTALL both. :) [15:01] tedg, we don't [15:01] tedg, we just don't uninstall on upgrade [15:01] seb128: Effectively we do as I'm sure most of our users are upgrades. [15:01] It is true I upgraded from the previous version [15:01] now empathy is the default client? [15:02] asac, no but looking at the list those are the ones out of order [15:02] after the sort [15:02]