[00:41] Hey people [00:42] do you think it is worth the pain to have an app allowing to browse through an historical of notifications ? Just a GUI parsing the .cache/notify-osd.log file ? [00:44] I think it would be, yes. [00:45] I know I would use it. Not often, but I would. [11:06] An idle moment of thought brought me this question: Would it ever be useful or interesting to be able to trigger an action when a notification matching certain parameters arrives? [11:07] I mean, in many ways what's actually being *implemented* here is a human-computer event interface :-) [11:16] if an action should be done automatically, the application sending the notification could manage it more accurately [11:16] if the action is meant to be triggered by the user after he read the notification : 1) its not compulsory and we shouldnt disturb him more by adding an action to the notification [11:17] 2) its compulsory and he should have a dialog with a compulsory action instead of a notification [11:19] I'm thinking more user-defined or cross-package things [11:21] ah for user defined, i wish it was there, indeed ^_^ [11:21] but i think it wont happen. Canonical is right when they say notifications should have no actions. It's a bad design, and they dont want to introduce it again, imo [11:22] A couple of random examples being, say: An app that can run a command whenever a certain string appears in an IM message [11:23] Or a fun one: When battery is critical, copy some important files to a backup [11:24] basically, I'm wondering if it's useful for package A to be able to say: Whenever a message matching arrives, I want to be triggered so I can do . [11:28] (Oh, and how was UDS?) [12:00] great :) [12:00] ccooke, i think Dbus / DCOP / gconf are more adaptated for this kind of usage [12:01] and having the app sending the message which you want to exploit be aware that package A wants to do this or that helps making sure for the app devs that its a wanted behaviour [12:01] and if they do it the proper IPC way they can throw it more efficiently and not relay on libnotify [12:25] Right [12:26] in which case, are messages repeated onto dbus? === mpt_ is now known as mpt === mpt_ is now known as mpt === mpt_ is now known as mpt === mpt_ is now known as mpt === agateau_ is now known as agateau === mpt_ is now known as mpt