[00:02] fta: tbird is now officially not xul ready for 3.0 :( [00:02] well ... guess no real news ;) [00:03] I told you so already, didn't I [00:03] <[reed]> not enough time [00:03] i'd say not enough motivation [00:03] there's been a bug for that for years [00:04] mailnews appears to be a huge-pile-of-pita ;) [00:04] (in regards of properly xulifying) [00:05] but agreed, its a priority thing of course [00:05] but tbird has to do a bunch of new features which are even hard enough to do in a reasonable amonut of time [00:05] so we probably have to accept that kind of compromise [00:06] fta: have you subscribed to the xul list on mozdev? [00:06] i think you were on CC list as well ;) [00:06] not yet [00:06] xulcentral [00:06] :) [00:06] yep [00:06] mozilla Bug 449228 [00:06] Mozilla bug 449228 in Composer "SVG authoring" [Normal,Unconfirmed] http://bugzilla.mozilla.org/show_bug.cgi?id=449228 [00:07] Keywords: ... helpwanted [00:07] ;) [00:07] probably a heavy weight task [00:08] but it's Product: SeaMonkey [00:08] yep [00:08] most likely it could deserve a standalone produce ;) [00:09] product [00:09] fta: what features were missing before you would try NM again? [00:11] ? [00:12] fta: you dont use NM right [00:13] what features were missing that made you abandon NM? [00:13] on my desktops, i don't. no need, it's all static [00:13] on my laptop, i'm using it [00:13] fta: ok. you can also do static setups now in NM 0.7 [00:14] in case you care ;) [00:14] it was possible before too, but caused me too many problems, included having all my MX blacklisted, hence i've lost a lot of emails [00:14] is your lapptop on intrepid? [00:15] yes [00:15] fta: in 0.6 no static ips were supported iirc [00:15] only dhcp [00:15] ? i've used static with nm since edgy [00:15] static what? [00:16] static IPs? [00:16] yes [00:16] that doesnt work [00:16] you probably used network-admin to configure it [00:16] and then network manager didnt manage anymore [00:17] i don't want anything using /etc/network/if-down.d/sendmail [00:17] fta: why isnt it used in your current setup? [00:17] arent you using ifupdown? [00:17] (which would call that on ifdown) [00:18] hmmm btw, i expect a reconnect to happen any minute :-D [00:18] hmm 11minutes left ;) [00:19] bug 99459 [00:19] Launchpad bug 99459 in sendmail "[feisty] sendmail unable to automatically determine domain name" [Undecided,Confirmed] https://launchpad.net/bugs/99459 [00:20] right ... by why isnt that script run in your current setup? [00:21] because i no longer run anything to manage my network. it's statically configured and i'm happy like that [00:21] debian bug 375787 [00:21] Debian bug 375787 in sendmail-base "sendmail-base: bad dialup.m4" [Important,Closed] http://bugs.debian.org/375787 [00:45] hmm... anyone up who could help me with a client/server related, programming problem? (and sorry for offtopic :)) === asac_ is now known as asac [00:46] depends [00:46] shout [00:46] oo [00:47] I need help with unique identification system... client connects to a server, then sends a job to the server, and it can disconnect after [00:47] he can connect later, and ask for the results... [00:47] So, I'm not really able to use client's host:port, since it's port is mostly random [00:48] So, I came up with two solutions [00:48] 1.) get ID from server, and give it to the user, so he can use it later [00:48] yep, that's a cookie [00:48] 2.) trust client and let him provide some sort of unique username [00:49] a simple cookie :) [00:49] hmm... actually, i haven't thought of keeping if in a file, so I can get it when client connects later [00:49] s/if/it/ [00:49] either the server passes a cookie to the client so it could come back, or you need a auth mechanism. [00:50] I thought of letting user type his id if he knows it... [00:50] hmm, I'll think I'll go for cookies [00:50] ...and now I'm hungry. off to get some icecream, too :) [00:50] thanks, fta :) [00:50] cookie is easier, if you have permanent storage [00:50] (or if it's still in memory) [00:51] I think I can use it. the project spec doesn't mention it, so I can use whatever I like in that case :) [01:28] Jazzva: you could also use "client-certs" [01:28] Jazzva: or simple http auth [01:29] (if you dont want to use session cookies ... which makes sense if your service is stateless) [01:29] asac: well, it has certain states [01:29] what kind of? [01:30] server keeps track of jobs on workstations and their statuses... [01:30] and status of the workstations... [01:30] but I think I don't need that kind of info in a cookie [01:31] Jazzva: imo, the server keeps that state indpendent from a session cookie [01:31] asac, huh? :) [01:31] it has to. cookies usually expire, and should be re-newable [01:31] fta, these won't... unless the user decides to delete them :) [01:32] Jazzva: without knowing any details i might be wrong, but from what i understand [01:32] asac, http://blog.vlad1.com/2008/07/30/no-browser-left-behind/ is guess that's what you referred to yesterday [01:32] you are looking for a auth mechanism to post jobs [01:32] and to query server state [01:32] the server itself does all the job management [01:32] yep... [01:32] which isnt related to the client server communication at all [01:32] e.g. why would a server stop processing a job when the client isnt connected? [01:33] fta: yes [01:33] server shouldn't stop processing the job. (well, workstations, which are behind the server, but that's another part of the project :)) [01:34] client will connect back, and it will ask the server for the results. the server will send them to the client, if they're ready [01:34] right [01:34] i thought that cookie should be simple enough [01:34] for what? [01:34] server would keep track of next avail ID [01:34] jobid? [01:35] client ID [01:35] (I need to know the jobs that one client sent) [01:35] ok [01:35] so I need to keep track of the clients, and provide a unique IDs for all of them [01:35] but client ID is not the same as session cookie [01:36] and you dont need authetntication? [01:36] nope [01:36] so it will be something like cookie - client gets his unique ID and uses it later to get job results. after that, it can reconnect and get some other ID if it wants to [01:36] and client doesnt know about any ID in the beginning? [01:37] you can use cookies ... sure ;) [01:37] nope... client will get his id from the server, and then will use it when it tries to connect again [01:37] yay for cookies :) [01:37] thanks :) [01:37] yeah. ersistent cookies is right then [01:38] ... but that constraints you to HTTP [01:38] which shouldnt be a big problem i guess [01:39] can't I make some sort of my cookie (which will keep an integer as ID :))? [01:39] fta: yes. canvas was the example they mentioned for that [01:40] and then just read that int from a file and send it to the server [01:40] Jazzva: you could use uuidgen ;) [01:40] which hopefully will be unique [01:41] but without client/server negotiation you usually cannot be 100% sure that something is unique [01:41] but then, without authentication it doesnt really matter ;) [01:41] anyone can snoop and steal your id ;) [01:41] right... but I don't think that's the problem in this project :) [01:41] Jazzva: what is the problem? [01:42] make a client/server/workstation programs, client sends a job, server does the job mgmt, workstation gets the job, returns the results to the server, which are returned to the client [01:43] just generate a hash on the server, md5, sha1 or hmac, that kind of thing. collisions are highly improbable, especially for a school project ;) [01:43] so, it's mostly concentrated on a synchronization between all threads, and distribution of work [01:43] fta: if he generates a hash on the server, he can also just use a integer ;) [01:44] Jazzva: i'd give all participants a client certificate ;) [01:44] :) [01:44] which can then be used to do everything [01:44] asac, will google up, as I'm not sure what exactly do you mean :) [01:46] Jazzva: will the workstation be a server itself or will it poll for jobs [01:46] poll for jobs... the server will send it to the one of the workstations [01:49] ok [01:50] if its server centric, then use the "acquire id from server approach" [01:50] otherwise generate a unique id on client [01:50] which would make this more peer-to-peer like [01:51] i think client certs are overkill for something like that [01:51] anyway ... i am off ;) [01:51] time to sleep for me [01:51] night! [01:51] night, asac [01:51] thanks for the help ;) [07:37] asac: Jazzva are we still packaging chatzilla since now there is another extension? [08:02] fta: i thought you had said you were dropping tbird-3.0 and ff-4.0 from your list of packages, atleast that is what you said a month or so ago (before your vacation) [09:04] hi ! === jt1 is now known as jtv [09:47] !rar [09:47] rar is a non-free archive format created by Rarsoft. For instructions on accessing .rar files through the Archive Manager view https://help.ubuntu.com/community/FileCompression. There is a free (as in speech) unrar utility as well, see !info unrar-free [10:05] hi XioNoX [10:08] hi asac ;) [10:09] still trying to detect the tab switch with mCurrentBrowser.... [10:09] :D [10:13] asac, i can do a alert(gBrowser.mCurrentBrowser); [10:13] and i get a [object XULElement] [10:13] but I don't know what fields there are in it [10:20] XioNoX: i found a good resource ;) [10:20] cool [10:21] http://developer.mozilla.org/en/docs/Code_snippets:Tabbed_browser [10:21] does that help? [10:21] I've to read it first :) [10:22] Detecting tab selection is the interesting part I think [10:23] XioNoX: yes. http://developer.mozilla.org/en/docs/Code_snippets is nice as well [10:23] XioNoX: just read aroujnd [10:23] ;) [10:24] this is a cool thing [10:25] XioNoX: i think that gBrowser.selectedTab.myUsedPlugins could be used [10:25] to store the plugins used on that tab [10:25] well [10:25] you get the point [10:26] of course not only for selectedTab [10:27] I don't understand exactly how the full thing should work [10:29] When we have 2 tabs with flash content [10:29] does it work like we have 2 firefox windows ? [10:31] And when we switch tabs, the js should detect it, and run a function to know what is the mimetype of the content ? [10:32] and by the way show the statusbar icon ? [10:36] or when the tab is load for the first time, save all mimetypes present on the page in a global array (with all tabs content) with the tab ID, and when we switch tab, the function Detecting tab selection select the good content modify the statusbar icon to pass the good mimetype to the opening window ? [10:36] asac ? [10:45] XioNoX: no ... whenever a page loads we get events from our xbl [10:45] so we remember which tab has which plugins [10:46] .... the only thing we do while switching tabs is to check whether our memory for that tab has a plugin at all [10:46] if it doesnt we disable the menu et al [10:46] ok [10:46] XioNoX: i think you ment something similar [10:46] in your last sentence [10:46] so it is what i've write last ? [10:46] ok [10:46] i think so [10:46] good [11:00] does the tabs have a unique ID ? [11:00] if yes, how can I get it ? [11:00] I find nothing in code snippets [11:04] XioNoX: why do you need a unique ID? [11:06] to store the tab id with the mimetype [11:07] but maybe it is useless [11:07] XioNoX: try to attach it to the tab directly [11:08] (not really sure if that works) [11:08] like: [11:08] var tab = .... [11:08] tab.usedMimeTypes = .... [11:08] if we run Detecting tab selection and then getPluginInfo(aEvent.target);, it will get the plugin info of the curent tab or not ? [11:09] but what "tab" refer to ? [11:09] depends of what getPluginInfo does [11:09] XioNoX: you probably get the tab through target [11:09] the other way is to: [11:09] var mytab = gBrowser.selectedTab [11:11] and attach an array on it ? [11:11] like var mytab.mymetype[] = ........ ? [11:11] mytab.myArray = [11:11] yes something like that [11:12] but maybe attach PluginInfo [11:12] thugh for now it doesnt matter ;) [11:12] just try [11:12] how far you get [11:13] ho! by the way, tomorrow is a off day [11:13] in france [11:15] lucky you [11:16] héhé :) [11:23] var gBrowser.selectedTab.mimeType = mimeType; [11:23] alert(gBrowser.selectedTab.mimeType); [11:23] don't work [11:23] but alert(mimeType); work [11:31] XioNoX: why would alert("" + mimeType"); work? [11:31] ah [11:31] hmm [11:31] alert(mimeType); return the mimetype [11:31] XioNoX: ok so maybe we have to manage them on our own [11:32] or all the mimetypes present on the page beacause it is in a loop [11:32] XioNoX: you can use selectedTab as the "key" in a map [11:32] ok [11:32] but take care that we have to listen for "tabs getting closed" [11:32] a kind of tab id ? [11:32] and remove it from the map [11:32] I know [11:32] XioNoX: yes. the object is a good enough tab id [11:32] it should be unique throughout the lifetime [11:32] of the tab [11:33] but maybe calling getPluginInfo when tab switch could work [11:35] XioNoX: i dont understand where getPluginInfo comes from? [11:35] what does it do? [11:35] browser.js [11:35] and what does it do? [11:35] ultimage function who retun the mimetype :D [11:36] nobody know [11:36] I is Magic I think [11:36] asac, see here : http://pastebin.mozilla.org/513810 [11:38] XioNoX: for that you need to find the plugin element first ;) [11:39] XioNoX: i dont want to search the full dom tree for plugin elements everytime you select a tab [11:39] especially since we could remember which plugins are on which tab [11:39] I call it with getPluginInfo(aEvent.target); [11:40] so maybe the aEvent of the tab change will be enough for him ? [11:41] XioNoX: sorry. in what context are you? [11:42] XioNoX: what do you want to call getPluginInfo when tab switch? what purpose? [11:43] to get mimetype of plugin in the curent page :) [11:45] if it work it will be easier than managing a king of database [11:46] but it is heavier for the browser [11:46] i think [11:49] XioNoX: imo its a performance issue [11:50] ok [11:52] so let's go for the array [11:52] but first, I need to eat ! [11:52] c u === `6og is now known as Kamping_Kaiser [12:35] back! [13:53] asac, what do you think of gBrowser.currentTab.setAttribute("myextension-myattribute", "xyz"); ? [13:53] can we give him an array ? [14:04] and what happend if we open a new tab who have flash in the background ? [14:23] XioNoX: you cannot use attribute for other things than strings afaict [14:28] XioNoX: what you can do is implement a xbl destructor [14:29] and unregister it there [14:29] if there is no entry left you could remove the key (tab) from the browser as well [14:30] makes sense? [14:33] people on xulfr [14:34] sayd that we can do ton_onglet.mon_tableau = ... [14:35] with var ton_onglet = gBrowser.selectedTab; ? [14:35] XioNoX: i thought you tried that [14:35] (that was the initial idea) [14:35] I thought too [14:36] 12:23 < XioNoX> var gBrowser.selectedTab.mimeType = mimeType; [14:36] 12:23 < XioNoX> alert(gBrowser.selectedTab.mimeType); [14:36] yes [14:36] exactly [14:36] XioNoX: maybe try again ... but since tabbrowser is a xbl object and not a "normal" javascript thing, it might indeed not ork [14:36] XioNoX: [14:37] paul> XioNoX: tu peux attacher un objet JS à ton noeud DOM. [14:37] paul> XioNoX: you can attach an object JS to your DOM node [14:39] XioNoX: i didnt look closely [14:39] the first line of the two is bogus [14:39] find out on your own [14:40] ;) [14:40] its syntactically non-sense [14:41] ? [14:41] look closer [14:42] even : var activeTab.mimeContent = "test" ; [14:42] [14:42] alert('foo'); [14:42] don't work [14:42] yes [14:42] thats normal [14:42] its not javascript [14:42] the fifrst line is wrong [14:42] i dont know why you think its right ;) [14:42] asac: still having problems with the localization. Tried solution 1, added additional "locale branding" lines to the en-US.manifest but this did not redirect the branding after the langpack-xx-xx@firefox-3.0.ubuntu.com extension gets installed. Seems the extension always wins. As for the 2nd solution you proposed, I wasn't sure which manifiest file to modify. en-US.manifest? [14:42] I don't know [14:43] what is not javascript ? [14:43] you will figure [14:43] look around in the code [14:43] there are other places where something similar is achieved [14:43] why do you think you need a "var" ? [14:44] I've removed it before you sayd it :) [14:44] and it work [14:44] var is for local variabled [14:44] makes no sense to use it there [14:44] i get it [14:44] good ... then go back to the initial idea [14:44] just add a new field like that with mime-type to store the array [14:45] XioNoX: if you can access the getPluginInfo thing you could also get the plugin info for your element and store that ... but for now lets only use the mime-type to get things going [14:46] ok [14:47] fta, teatime should bbe fixed now in Debian... [14:51] Jazzva: nice [14:51] in testing? [14:51] if not ask for a freeze exception [14:51] in debian-release [14:51] asac, not sure... let me re-read the mail :) [14:52] (i just got up, after 10.5 hours of sleep. that's bad) [14:53] Jazzva: if it just entered sid [14:53] waait a few more days and get confirms that the RC bug was fixed [14:53] then ask for exception to enter testing [14:53] aka lenny [14:54] crimsun: the problem with your alsa-lib branch is that it would break kde (that dont have pulse) [14:54] crimsun: is there a way to configure a fallbacdk [14:55] e.g. if pulse is not detected fallback to whatever was used by fdefault [14:55] asac: I know a brute force way of fixing but it's ugly: modifying the language-pack-xx (as that contains the ff lang extension) and insert the correct branding [14:56] asac: much rather figure out how to override the langpack extension branding files if possible though [14:58] bfiller: try the second solution [14:58] it doesnt matter which manifest you modify [14:58] any should be ok [14:58] ( as long as its read by xulrunner/firefox) [14:59] asac: so browser.manifest or classic.manifest would be ok [14:59] the override should always win [14:59] i just havent used it for locales [15:00] yes [15:00] for testing thats ok [15:00] later we can figure where to put that [15:00] asac: ok, you said to add the following: override chrome://branding/locale/brand.dtd chrome://mybranding/locale/brand.dtd. Where do I put the "mybranding", in the associated .jar file? [15:00] asac, it's still not built in any. [15:01] bfiller: use chrome://ubufox/locale/brnad.dtd [15:01] for now [15:02] you can put the override in ubufox chrome.manifest [15:02] ifrst test if you can open chrome://ubufox/locale/brnad.dtd [15:02] then add the override and see if [15:02] asac: I'll give it a try thanks [15:03] chrome://branding/locale/brand.dtd opens the same file [15:03] if thats the cas then chances are high that it works ;) [15:04] asac: trying it now, I'll let you know [15:16] hi [15:16] hi fta [15:16] hey fta [15:17] Jazzva, is that 2.8.0-5 ? [15:17] fta, yes [15:17] cool, rebuilding it.. [15:17] Jazzva: when it built properly we might try to sync it to ubuntu [15:17] (i assume that its not in ubuntu yet) [15:17] asac, I wondered if it's too late to be synced [15:17] no, it's not in. I thought it could be just synced [15:17] Jazzva: well. i think its a different thing wheteher a package is just synched [15:18] or whether the package is maintained by a ubuntu contributor [15:18] well, this should be just synced [15:18] it's not maintained by an ubuntu contributor [15:18] not? [15:18] hmm [15:18] is it in the archive atm? [15:18] Ian Wel... I forgot the surname [15:18] if its broken in ubuntu and a synch would cure that, we can certainly sync it [15:19] good :) [15:19] Jazzva: just file a bug against teatime about the bug that makes it unusable [15:19] then present the sync a solution [15:19] mhm... ok :) [15:19] and ask hobbsee (i think she still does universe synchs) [15:19] I'll wait for it to get built in debian [15:19] sure [15:20] testing it... [15:20] well, I'm off to redesigning GUI for client and server for my project. now to actually send a job :), instead of test message :) [15:20] hehe [15:21] i dont think the syn needs to happen to day [15:21] take your time [15:21] asac, it's not a time-consuming thing, so I'll check periodically :) [15:21] yeah, fixed [15:22] thanks for testing :) [15:22] fta, I'll ask you to report that in a bug report, once I open it... I think that would be good :) [15:22] Jazzva: hehe [15:23] Jazzva: you can subscribe to package tracking system [15:23]