:: #mythtv

Daily chat history

Current users (77):

aarcane_, aloril_, amessina, Anssi, Beirdo, brfransen, caelor, Captain_Murdoch, Chutt, clever, coling, Cougar, dekarl, dgeary2_, eharris, ElmerFudd, fetzerch, ghoti, Gibby, gregL, GreyFoxx, Guest85778, jams_, jarle, jarryd, jheizer, jheizer_, joe____, johanbr, joki, jpharvey, jst, jwhite, jya, kc, kurre2, kwmonroe, laga_, MavT, moparisthebest, MythBuild, MythLogBot, nephyrin, Nothing4You, nvzn_, nyloc, peper03, poptix, purserj, rhpot1991, robink, rsiebert, Seeker`, seld_, Sharky112065, sjennings, skd5aner, sl1ce, SmallR2002, sphery, sraue, stuarta, stuartm, superm1, TandyUK, taylorr, tgm4883, Tobbe5178, toeb, tonsofpcs, tris, wagnerrp, wahrhaft, wolfgang1, XDS2010, xris, _charly_
Thursday, October 17th, 2013, 00:03 UTC
[00:03:30] nvzn: jya: do you know if it is simple enough to use dolby-truehd and dts-hd with a soundcard and analog speaker setup in linux?
[00:04:21] wagnerrp: didn't that get answered in the other channel a couple days ago?
[00:05:09] jya: nvzn: I've answered that question in mythtv-users
[00:05:20] jya: i answered you I believe
[00:05:51] wagnerrp: looks like... ~19 hours ago
[00:05:57] nvzn: i see it
[00:05:59] nvzn: thanks
[00:06:05] nvzn: sorry for asking in the wrong channel
[00:06:13] nvzn: i'll move over to users if i have more q's
[00:06:16] nvzn: cheers
[00:29:29] nvzn_ (nvzn_!~nvzn@unaffiliated/nvzn) has joined #mythtv
[00:32:28] nvzn (nvzn!~nvzn@unaffiliated/nvzn) has quit (Ping timeout: 240 seconds)
[00:48:23] skd5aner: my first patch in a very long time, #11908 – as simple as it is, I hope it helps (and is correct) :)
[00:48:23] ** MythLogBot **
[00:50:56] nascar (nascar! has joined #mythtv
[00:52:43] rsiebert (rsiebert! has joined #mythtv
[00:56:04] rsiebert_ (rsiebert_! has quit (Ping timeout: 260 seconds)
[01:35:37] Captain_Murdoch: jya, forbidden with wget on that URL. if you can get a couple network captures of a succesful and failed download I'll take a look at them.
[01:35:56] jya: let me redo a URL for you
[01:36:42] jya: how do you do your network capture ?
[01:37:40] jya: hmm.. that url still works as of now.. do you have an iphone by any chance?
[01:37:55] Captain_Murdoch: just tcpdump to a file. use "-s 2048 -w outputfile.pcap"
[01:38:17] jya: I have a pretty busy network activity; it may be hard to see what you want...
[01:38:22] Captain_Murdoch: yeah, have a 4 but can't troubleshoot now, getting kids ready for bed and noticed this message.
[01:38:57] jya: if you just use the youtube add; and attempt to play the video via airplay.. (it will fail)
[01:39:01] Captain_Murdoch: put an entry in your hosts file for and then use "tcpdump -i yournetworkinterface -s 2048 -w outputfile.pcap host theIPyouputinyourhostsfile and port 80
[01:40:06] Captain_Murdoch: can you get that url via chrome/firefox on the desktop?
[01:40:10] Captain_Murdoch: s/you/i/
[01:40:14] jya: yes...
[01:40:39] Captain_Murdoch: ok, will have to try from another machine a little later, this one doesn't show anything about airplay.
[01:40:40] jya: hmmm.. I wonder if that works because wget, chrome etc.. uses IPv6, while the myth would likely use ipv4
[01:40:55] jya: did you enable airplay on you frontend?
[01:40:55] clever (clever! has quit (Ping timeout: 265 seconds)
[01:41:07] Captain_Murdoch: haven't updated past 0.25 yet.
[01:41:17] Captain_Murdoch: and haven't used it's limited functionality there.
[01:41:26] jya: ah... won't work then.. you need http live streaming support
[01:44:17] Captain_Murdoch: behind the curve again on Qt version on my master backend, need to update that before I can update MythTV.
[01:45:02] Captain_Murdoch: moved most server functionality off the box already so will just use something less 'stable' than CentOS this time. :)
[01:45:36] jya: what's the syntax to set the host and port ?
[01:45:43] jya: tcpdump -i en0 -s 2048 -w wget.pcap host port 80 ?
[01:45:47] Captain_Murdoch: in tcpdump? "host IP and port PORT"
[01:45:55] Captain_Murdoch: need the and
[01:46:03] Captain_Murdoch: rest looks good
[01:46:10] jya: in quotes?
[01:46:13] Captain_Murdoch: no
[01:46:25] clever (clever! has joined #mythtv
[01:46:26] Captain_Murdoch: sorry
[01:48:22] jya: ouch, 49 packets with wget; 558 with myth :)
[01:49:36] Captain_Murdoch: probably retries
[01:49:58] Captain_Murdoch: not sure, capture will tell
[01:52:49] Captain_Murdoch: jya, stepping away for a bit to get the kids ready for bed.
[01:53:22] jya: ok.. creating a ticket with the capture
[01:57:41] jya: Captain_Murdoch: #11909
[01:57:41] ** MythLogBot **
[01:58:32] jya: hmm... I see an error about : [http @ 0x10a5306e0] HTTP error 403 Forbidden
[01:59:00] jya: that indicates to me that an attempt to download the file via ffmpeg was attempted just before... wonder where
[01:59:17] jya: that could be the problem: it thinks there's too many attempts, too quickly
[02:00:42] jya: ah yes... it's the TestForHTTPLiveStreaming that uses ffmpeg before to test if a stream is HLS or not
[02:14:10] jya: hmm... even without that it doesn't work
[02:14:41] jya: going to redo a tcpdump; with the network access done with a direct call to QNetworkAccessManager
[02:16:02] nyloc (nyloc! has joined #mythtv
[02:20:09] _nyloc_ (_nyloc_! has quit (Ping timeout: 272 seconds)
[02:37:13] dmfrey (dmfrey! has quit (Quit: Ex-Chat)
[02:41:09] peper03 (peper03!~peper03@mythtv/developer/peper03) has quit (Ping timeout: 272 seconds)
[02:41:24] peper03_ (peper03_!~peper03@mythtv/developer/peper03) has joined #mythtv
[02:41:31] peper03_ is now known as peper03
[02:50:00] jheizer: skd5aner, talking to you earlier got me to work on MobileMyth tonight :)
[02:58:32] Captain_Murdoch: jya, you around still?
[02:58:37] jya: yep
[02:58:53] Captain_Murdoch: those 2 captures are for the same URL right?
[02:59:07] jya: yes
[02:59:34] Captain_Murdoch: ok. the issue is that mythtv is double-encoding the % sign.
[02:59:50] jya: that's with the patch I provided. which basically, bypass all existing code for that particular URL and download via Qt network manager
[03:00:08] jya: i thought it was something like that...
[03:00:19] jya: had the feeling yesterday :(
[03:00:29] Captain_Murdoch: in the wget one, you have this in the URL: "/sparams/id%2Cip%2Cipbit" and in mytthv you have "/sparams/id%252Cip%252Cipbit"
[03:01:16] wagnerrp: those shouldn't be escaped, should they?
[03:01:18] Captain_Murdoch: at least it's simple.
[03:01:20] jya: i see...
[03:01:34] Captain_Murdoch: yeah. %2C becomes %252C
[03:02:00] Captain_Murdoch: so not double-encoding, just encoding when it shouldn't. :|
[03:02:03] wagnerrp: ah, you did just say that... :)
[03:02:16] jya: it's the URL I pass to QNetworkRequest
[03:02:21] ** wagnerrp stops making comments without reading the immediate backlog **
[03:02:38] Captain_Murdoch: wagnerrp, when the rest of us stop, you can feel free to do so as well. :)
[03:02:48] jya: it's already encoded ; so I guess it shouldn't encoded at all
[03:03:06] jya: or is there a way to tell QNetworkRequest that a url is already encoded
[03:03:08] Captain_Murdoch: yeah, may need to un-encode before passing it. not sure if you can tell it that it's already encoded.
[03:03:42] wagnerrp: anyway, what i meant was that they should not even be encoded the first time, should they?
[03:03:59] wagnerrp: shouldn't it actually be passing the '&' and '='?
[03:04:05] Captain_Murdoch: it's not a %, it's whatever 2C is.
[03:04:48] wagnerrp: looks like 2C is a comma
[03:04:57] jya: ok.. so I need to decode the URL pass in the AirPlay playlist first
[03:05:45] joki (joki! has quit (Ping timeout: 272 seconds)
[03:07:59] Captain_Murdoch: yeah, I think so. I see other people having the same issue and they're told to unencode first by doing something like: QUrl downloadurl = QUrl::fromEncoded(theencodedurl); to convert it.
[03:09:17] jya: i guess when can state at an API level that all URL passed to MythDownloadManager shouldn't be encoded
[03:09:58] Captain_Murdoch: yeah, I'd think so.
[03:10:14] joki (joki! has joined #mythtv
[03:15:28] eharris (eharris! has quit (Ping timeout: 240 seconds)
[03:17:34] eharris (eharris! has joined #mythtv
[03:33:31] dgeary2 (dgeary2! has joined #mythtv
[03:45:58] jya: Captain_Murdoch: are the headers that come with a http body always url encoded?
[03:47:13] Captain_Murdoch: not that I know of.
[03:48:28] fetzerch (fetzerch!~quassel@unaffiliated/fetzerch) has quit (Ping timeout: 264 seconds)
[03:48:43] eharris (eharris! has quit (Ping timeout: 260 seconds)
[03:49:41] fetzerch (fetzerch!~quassel@unaffiliated/fetzerch) has joined #mythtv
[03:50:08] eharris (eharris! has joined #mythtv
[03:51:54] NightMonkey (NightMonkey!~NightrMon@pdpc/supporter/professional/nightmonkey) has quit (Quit: Body blow! Body blow!)
[03:55:54] jya: hmm... how could I detect if something is url encoded or not then?
[04:02:11] Captain_Murdoch: not sure if there's a sure-fire check. url decoding shouldn't mess up something that isn't encoded already unless it has text that looks like %-codes in it. if people put real % signs in their URLs they deserve to have them broken. :)
[04:03:28] jya: fair enough.
[04:04:39] jya: airplay working again...
[04:05:32] jya: still have the issue of the first playback order not working... don't get why
[04:16:56] dgeary2 (dgeary2! has quit (Quit: Ex-Chat)
[04:23:34] stichnot (stichnot!~stichnot@mythtv/developer/stichnot) has joined #mythtv
[05:36:29] stichnot (stichnot!~stichnot@mythtv/developer/stichnot) has quit (Ping timeout: 272 seconds)
[06:07:28] FabriceMG (FabriceMG! has joined #mythtv
[06:10:43] SteveGoodey (SteveGoodey! has joined #mythtv
[06:15:29] SteveGoodey (SteveGoodey! has quit (Client Quit)
[06:20:30] MythBuild: build #109 of 0.27-f19–32bit is complete: Failure [4failed git_1] Build details are at . . . t/builds/109 blamelist: Jean-Yves Avenard < >
[06:22:04] cbovy (cbovy!~cbovy@ has joined #mythtv
[06:32:09] skd5aner (skd5aner! has quit (Ping timeout: 252 seconds)
[06:32:39] skd5aner (skd5aner! has joined #mythtv
[06:59:50] stichnot (stichnot! has joined #mythtv
[06:59:50] stichnot (stichnot!~stichnot@mythtv/developer/stichnot) has joined #mythtv
[06:59:50] stichnot (stichnot! has quit (Changing host)
[07:00:34] natanojl (natanojl! has joined #mythtv
[07:18:15] cbovy (cbovy!~cbovy@ has quit (Ping timeout: 260 seconds)
[07:22:55] nascar (nascar! has quit (Ping timeout: 260 seconds)
[07:38:49] natanojl (natanojl! has quit (Ping timeout: 272 seconds)
[08:00:04] stuarta: MythBuild: force build 0.27-f19–32bit now please
[08:00:05] MythBuild: build forced [ETA 12m54s]
[08:00:05] MythBuild: I'll give a shout when the build finishes
[08:12:06] MythBuild: build #110 of 0.27-f19–32bit is complete: Success [3build successful] Build details are at . . . t/builds/110
[08:17:21] stuartm: wagnerrp: Captain_Murdoch can say for sure, but I don't think adding 'ChannelIcons' to the special groups is correct, it's already in the built-in groups and no current built-in groups are also in the special groups list
[08:18:35] stuartm: in fact I believe built-in groups are editable, which we want to avoid for channel icons as some parts of the code (and third party scripts) still work directly with the ~/.mythtv/channels directory
[08:20:14] stuartm: skd5aner: ^^
[08:20:51] stuartm: seems like the correct fix would be to exclude StorageGroup::m_builtinGroups() from the space calculations instead
[08:23:24] stuarta: or fix broken code / scripts?
[08:26:26] sl1ce (sl1ce! has quit (Ping timeout: 256 seconds)
[08:26:53] stuartm: still wouldn't be correct for the second reason I gave, a storage group shouldn't be in both the 'special' and 'buildin' lists
[08:27:31] stuartm: and I don't see any value in letting users specify a different storage location for channel icons
[08:28:54] stuartm: if I had a time-machine I'd even go back and make the fanart/artwork storage groups built-in too, present the user with the fewest configuration steps as possible
[08:32:03] rhpot1991 (rhpot1991!~rhpot1991@ubuntu/member/rhpot1991) has quit (Ping timeout: 260 seconds)
[08:32:38] tgm4883 (tgm4883!~tgm4883@ubuntu/member/tgm4883) has quit (Ping timeout: 264 seconds)
[09:03:42] stuartm: after fixing the GetChannelIcon service API mythweb should be displaying icons again, but it's not :(
[09:19:04] cbovy (cbovy!~cbovy@ has joined #mythtv
[09:53:05] dekarl1 (dekarl1! has joined #mythtv
[09:55:39] dekarl (dekarl! has quit (Ping timeout: 248 seconds)
[10:13:12] sl1ce (sl1ce! has joined #mythtv
[10:20:26] cbovy (cbovy!~cbovy@ has quit (Ping timeout: 245 seconds)
[11:51:22] Merlin83b (Merlin83b! has quit (Ping timeout: 265 seconds)
[11:54:52] warped_ (warped_! has joined #mythtv
[11:57:52] Merlin83b (Merlin83b! has joined #mythtv
[12:20:43] MavT (MavT! has joined #mythtv
[12:20:59] MaverickTech (MaverickTech! has quit (Read error: Connection reset by peer)
[12:37:39] Merlin83b (Merlin83b! has quit (Read error: Connection reset by peer)
[13:00:02] warped_ (warped_! has quit (Quit: warped_)
[13:04:02] cbovy (cbovy!~cbovy@ has joined #mythtv
[13:32:53] nascar (nascar! has joined #mythtv
[13:42:19] nascar (nascar! has quit (Ping timeout: 260 seconds)
[13:52:34] Jordack (Jordack! has joined #mythtv
[14:02:22] Captain_Murdoch: stuartm, built in groups shouldn't be in the disk calculations because they shouldn't normally exist in the storage group table and that's what's used to get the directory list. The built in Streaming group is probably the only one where we might want people to override it, we could probably set it up so that the rest could't be added via setup since they are declared for convenience not configurability.
[14:03:14] Captain_Murdoch: excluding m_builtinGroups from the disk space calcs would be good as well though if people have configured one such as the HLS streaming group.
[14:39:17] cbovy (cbovy!~cbovy@ has quit (Ping timeout: 268 seconds)
[14:53:47] stuartm: Captain_Murdoch: so the issue in #11908 is bogus then? channel icons were not included to begin with, so they don't need to be included?
[14:53:47] ** MythLogBot **
[14:54:05] stuartm: excluded, don't need to excluded
[15:06:14] FabriceMG (FabriceMG! has quit (Quit: FabriceMG)
[15:15:51] cbovy (cbovy!~cbovy@ has joined #mythtv
[15:19:51] stuartm: dekarl1: am I right in thinking that there's no standard format for the xmltv grabber configuration file? I was thinking of a simple solution to populating the xmltv grabber config with the channels it needs to grab but that's a non-starter if the configs don't all follow the same formatting
[15:20:57] stuartm: there are other solutions, but they would all require changes/updates to the xmltv spec that might take months to happen (and longer to reach end-users)
[15:22:49] stuartm: e.g. the much talked about, but so far non-existant lineups support
[15:27:01] rhpot1991 (rhpot1991!~rhpot1991@2001:4968:202:3:e4ae:1a6b:7ae8:eb97) has joined #mythtv
[15:27:02] rhpot1991 (rhpot1991!~rhpot1991@2001:4968:202:3:e4ae:1a6b:7ae8:eb97) has quit (Changing host)
[15:27:02] rhpot1991 (rhpot1991!~rhpot1991@ubuntu/member/rhpot1991) has joined #mythtv
[15:32:11] stuartm: skd5aner: any luck reproducing your deadlock?
[15:40:26] dekarl-is-not-wo (dekarl-is-not-wo!51c8c614@gateway/web/freenode/ip. has joined #mythtv
[15:41:26] dekarl-is-not-wo: stuartm: yes, there is no standard configuration format, but there is a standard configuration api format :) (the latter needs extensions, too. mea culpa, I've been pushing that out of scope for some time now)
[15:41:55] dekarl-is-not-wo: see
[15:43:11] dekarl-is-not-wo: lineups is for automatic configuration (at least the part that I'm after is)
[15:46:41] stuartm: dekarl-is-not-wo: yeah I know about the configuration API, I've been promising to support that for years now and finally I actually hope to keep my word (leveraging the new setting stuff)
[15:47:15] stichnot (stichnot!~stichnot@ has joined #mythtv
[15:47:15] stichnot (stichnot!~stichnot@mythtv/developer/stichnot) has joined #mythtv
[15:47:15] stichnot (stichnot!~stichnot@ has quit (Changing host)
[15:47:45] stuartm: but there's still a disconnect – several grabbers ask users to pick the channels they want to grab data for and then a few minutes later we ask users to do it all again by mapping the xmltvids to channels in the database
[15:48:21] stuartm: and that needs to happen each time a user adds or removes a new channel
[15:49:05] dekarl-is-not-wo: stuartm: shall we extend the list of parameter types as we go? e.g. add a way to tell mythtv "this is a password style field and should be *******", "this is a directory / file, so use a file selector", "this is scratch space, feel free to automatically fill in /tmp"
[15:49:24] stuartm: a simple hack might have taken the xmltvids found in the mythtv database and supplied those to the xmltv grabber at runtime – it's at best a half solution, but it would avoid the need for users to modify two places instead of one
[15:50:01] dekarl-is-not-wo: can't we shows to lists "scanned channels" and "available guide channels" and let the users mix and match them? Then turn them on/off via the configuration api
[15:50:22] dekarl-is-not-wo: s/to lists/two lists/
[15:51:28] dekarl-is-not-wo: (the next step would be to allow for multiple grabbers per video source, or different OTA guide sources to pick from)
[15:51:58] dekarl-is-not-wo: but in the end you(MythTV) would use the configuration API to turn channels on/off automatically
[15:52:58] stuartm: dekarl-is-not-wo:
[15:53:23] stuartm: 4th March 2009
[15:55:10] dekarl-is-not-wo: as far a I know there are 0 (none at all) users of the API, so if you spec out what we need I can hack it into the Configuration module
[15:55:39] stuartm: dekarl-is-not-wo: the current configuration API is a 'question' and 'answer' API, designed basically as a pass-through the old command line configuration, there isn't (or wasn't) stuff like 'AddChannel' methods
[15:56:24] stuartm: the script provides a question, the format of the answer(s) and we just display this on-screen to the user with our own widgets/UI
[15:57:10] dekarl-is-not-wo: stuartm: yes, but the last stage is always the "tick all channels that you want from this list" iirc. We can extend that if you like.
[15:57:43] stuartm: if we can extend it to do what we think we need, then I'm all for that :)
[15:59:23] dekarl-is-not-wo: that's what I was thinking. No users. We control both ends, so lets move it forward. Once we reach the linups we need to include Nick
[16:03:59] stuartm: just remembered that the configuration API allow required that we write the configuration file – we don't hand the answers back to the script but write them out, so the config file does need to be standardised
[16:05:59] SteveGoodey (SteveGoodey! has joined #mythtv
[16:09:27] stuartm: of course it makes more sense for the script to write the answers to the configuration file itself, that way we don't need to know or care about the format (or update for changes)
[16:09:38] stuartm: so that's the first thing we'd change
[16:10:11] NightMonkey (NightMonkey!~NightrMon@pdpc/supporter/professional/nightmonkey) has joined #mythtv
[16:11:06] dekarl-is-not-wo: stuartm: hmm . . . ;view=markup looks like there is a standard, if you use the configuration module instead of implementing your own stuff.
[16:11:39] cbovy (cbovy!~cbovy@ has quit (Ping timeout: 260 seconds)
[16:11:46] warped_ (warped_! has joined #mythtv
[16:11:47] dekarl-is-not-wo: stuartm: aye, I'd prefer to encapsulate the interal file formats via an API
[16:20:09] dekarl-is-not-wo is now known as dekarl-work
[16:20:29] dgeary2 (dgeary2! has joined #mythtv
[16:34:50] stuartm: as far as the matching of xmltv to channel, I'm wondering if that's best handled by using --list-channels to return the all the possible names/xmltvids, then in the current channel editor when you edit a channel populate a select box with those options (no more copy/paste)
[16:35:35] stuartm: once we're done editing the channels we can give a list of all those xmltvids which are actually used back to the grabber which updates the configuration
[16:37:37] stuartm: we can add in a little auto-detection which works like the channel icon grabber – matching up xmltvids with channels where there are identical matches between channel names, and only prompting users for those channels where we couldn't figure it out
[16:37:53] dekarl-work: stuartm, maybe add some hints to the channel list so mythtv can suggest likely matches first
[16:39:03] stuartm: in fact, a while back I had the idea of using the data we've already collected for the icon services – we send networkid, serviceid and xmltv for channels back to already, using that we could auto-populate xmltvid for 99% of users/channels
[16:39:16] stuartm: dekarl-work: yeah
[16:40:05] dekarl-work: aye, but then you go down the rabbit whole of reworking lots of stuff, ending up with questions like "why don't we all set up private instances of atlas for our channel/guide needs?" :)
[16:40:26] dekarl-work: s/whole/hole/
[16:40:27] stuartm: I think a combination of all of the above would be a vast improvement – auto-detection would handle most cases, and for the rest the users just have to select from a pre-populated list with the best matches listed first
[16:42:26] dekarl-work: where do you intend to put the configuration file? in mythbuntu you run mythtv-setup with the frontend user but mythbackend with the backend user :/
[16:42:37] stuartm: dekarl-work: the data is already there, we've been collecting it for years, a bit of perl glue on the backend and a minor addition to the existing icon handling code ... we'd have a working solution by the end of next week as opposed to waiting for atlas or whomever to sort out their implementation
[16:43:04] dekarl-work: stuartm, I still need to figure out how to checkout services :/
[16:43:33] stuartm: dekarl-work: that's a little broken, but we can workaround that scenario using the storage group code – we can write files to the backend configuration directory even from the frontend
[16:44:22] dekarl-work: stuartm, so you want to get rid of mythtv-setup and put everything in the frontend? (because the backend won't be running in parallel to setup right now...)
[16:45:04] dekarl-work: ohh, so you call backend functions from setup without the backend running? How do you bridge the user id divide?
[16:45:13] stuartm: dekarl-work: no, my point was simply that we can write files to the backend configuration directory from any client, even the frontend
[16:45:27] stuartm: even a remote frontend
[16:45:45] stuartm: someone added that capability in the last year
[16:45:59] dekarl-work: hmm, I've been thinking about that point multiple times in the past years and could not come up with a simple solution
[16:46:40] dekarl-work: but how does that work when the backend is stopped to let go the capture devices?
[16:46:48] stuartm: mythutil uses it for the --copyfile feature
[16:55:10] stichnot (stichnot!~stichnot@ has joined #mythtv
[16:55:10] stichnot (stichnot!~stichnot@mythtv/developer/stichnot) has joined #mythtv
[16:55:10] stichnot (stichnot!~stichnot@ has quit (Changing host)
[16:56:06] stuartm: dekarl-work: oh, well obviously it wouldn't ...
[16:56:08] warped_: jpabq: I'm curious – did You find any interesting in samples I provided for #11882?
[16:56:54] stuartm: if it was the case that we still haven't removed that restriction by 0.28, then it's mythbuntu's problem to solve
[16:57:11] stuartm: we've always required that mythtv-setup ran as the backend user in the past, so ...
[16:58:02] stuartm: it must already run mythfilldatabase (the manual run it suggests) as the backend user, as that's where the xmltv configs are
[16:59:18] stichnot: warped_: I'm trying to download your sample but it fails. Is it still available?
[17:00:39] warped_: stichnot: are You trying to download files with URLs in ticket ?. At bottom of ticket there is bunch of links….
[17:01:17] warped_:
[17:01:18] dekarl-work: stuartm: can of worms :) your pragmatic solution is fine with me
[17:03:40] dekarl-work (dekarl-work!51c8c614@gateway/web/freenode/ip. has quit (Quit: laters)
[17:22:05] Captain_Murdoch: stuartm, unless he's created his own Channel Icons SG overriding the built-in one, then channel icons SG should not be included in the stats. MainServer::BackendQueryDiskSpace gets the list of special groups and excludes that in the SQL to get a list of all non-special SG's defined in the DB. Mainserver doesn't know about the builtin groups, they are only internal to the StorageGroup class to allow us to use StorageGroup code
[17:22:05] Captain_Murdoch: to find/use files. If he didn't manually override the Channel Icons SG then we might have to dig deeper to see why it was included in the disk stats
[17:23:22] stuartm: Captain_Murdoch: ok thanks, I'll revert the 'fix'
[17:24:25] stuartm: or let wagnerrp do it if he prefers
[17:24:36] Captain_Murdoch: :)
[17:27:47] stuartm: dekarl1: you're still unable to clone the services repo?
[17:34:17] warped_ (warped_! has quit (Quit: warped_)
[17:44:32] dgeary2_ (dgeary2_! has joined #mythtv
[17:45:36] dgeary2 (dgeary2! has quit (Ping timeout: 245 seconds)
[17:51:01] tgm4883 (tgm4883!~tgm4883@ubuntu/member/tgm4883) has joined #mythtv
[17:54:06] skd5aner: stuartm: when I work up today, mythbackend wasn't running, but I haven't seen the deadlock recenty... it was always very random. Sometimes daily, sometimes a few weeks in between it showing itself
[17:55:10] skd5aner: s/work/woke
[17:58:22] skd5aner: stuartm: if the patch for #11908 is incorrect, please feel free to revert. However, the root issue is still true – my channel icon sg shows up in the disk calculation section without it
[17:58:22] ** MythLogBot **
[17:59:14] skd5aner: Seems this is the section where it queries which directories to calculate and which ones to exclude: . . . er.cpp#n4445
[18:01:36] stuartm: skd5aner: DELETE FROM storagegroup WHERE groupname='ChannelIcon';
[18:02:02] skd5aner: why would it be in that table though?
[18:02:15] stuartm: no clue how that happened
[18:02:21] skd5aner: i dont recall manually creating it??
[18:02:22] stuartm: but it shouldn't be there
[18:02:45] stuartm: skd5aner: which distro? Could be that one of them has incorrectly created it
[18:02:56] skd5aner: i compile from git
[18:03:09] stuartm: then I don't know, sorry
[18:03:32] stuartm: maybe there was a commit where it was briefly a special group instead of a built-in group
[18:05:23] skd5aner:
[18:07:21] stuartm: well that's ... wrong :/
[18:07:40] stuartm: and deleted
[18:08:59] stuartm:
[18:09:56] stuartm: so from day one there has always been a built-in storage group for channel icons
[18:10:11] skd5aner: ok, well... now I feel dumb :S
[18:11:14] stuartm: looks like the actual creation of the group goes right back to the addition of storage groups – or thereabouts
[18:11:46] stuartm: when we actually started using the group I changed the path so that it used the pre-existing location for channel icons
[18:12:26] stuartm: i.e. users wouldn't need to move their icons, they would, theoretically at least, just work
[18:12:58] skd5aner: so...
[18:13:14] stuartm: I wonder how many people followed the advice on that wiki page (or saw it repeated elsewhere)
[18:13:45] stuartm: maybe we need a schema update to remove it
[18:13:58] stuartm: not that it does much harm ...
[18:15:01] skd5aner: just a min... brb
[18:20:44] bill6502 (bill6502!~bill@ has joined #mythtv
[18:23:06] skd5aner: if I just delete the entry from the table, that'll "fix" it? How does the channelicon sg location get defined then?
[18:23:22] stuartm: it's hardcoded
[18:23:44] bill6502: stuartm: I wrote the Wiki: , so I'll undo it. Is putting the icons under ~mythtv/channels is the proper method?
[18:23:45] stuartm: along with several other 'minor' storage groups
[18:24:03] stuartm: bill6502: yes
[18:24:22] skd5aner: so, it always expect it to be in that location on the MBE?
[18:24:27] skd5aner: expects
[18:24:41] stuartm: that's where both mythfilldatabase and the channel icon downloader in mythtv-setup write them
[18:24:41] skd5aner: damn laptop keyboard sucks
[18:24:43] stuartm: skd5aner: yes
[18:24:57] skd5aner: k, I'll move it back there then and remove that entry
[18:25:00] skd5aner: from the db
[18:52:52] cbovy (cbovy!~cbovy@ has joined #mythtv
[18:56:40] warped_ (warped_! has joined #mythtv
[18:59:49] stuartm: beer for anyone who figures out why mythweb is failing to load channel icons, it's supposed to use the services API, and that's now working fine but mythweb (at least here) is failing to write the icons to it's cache or display them
[19:01:09] stuartm: not sure why the cache required either, it could just use the services url directly in the <img> tag – the browser will cache the image
[19:01:17] natanojl (natanojl!~jonatan@mythtv/developer/natanojl) has joined #mythtv
[19:07:01] bill6502: Argh, should have read my email. I just stomped on top the changes by GBee, sorry stuartm. Mostly the same, fixed MY incorrect SHA1 (used master's not 0.27.) And, allowed for users that start the BE as other than mythtv.
[19:32:37] skd5aner: stuartm: I can get mythweb to load icons with a symlink
[19:33:33] stuartm: yeah, it should work without that
[19:34:14] stichnot: warped_: thanks, I guess I was trying the wrong link.
[19:34:20] stuartm: the code calls the services API url, saves the data to data/tv_icons – or it's supposed to
[19:34:36] stuartm: seems broken but I'm not seeing an obvious reason
[19:35:24] warped_: stichnot: if there is anything I can do to help You with 11882 – I'll be more than happy to do so…
[19:36:04] Chutt (Chutt! has quit (Ping timeout: 260 seconds)
[19:36:31] Chutt (Chutt! has joined #mythtv
[19:38:11] sjennings_ (sjennings_! has joined #mythtv
[19:39:00] sjennings_ (sjennings_! has quit (Client Quit)
[19:39:24] stuartm: Gary, no beer I'm afraid, permissions issues were the first thing I checked
[19:39:50] sjennings (sjennings! has joined #mythtv
[19:41:05] stuartm: decided that instead of wasting time fixing mythweb, I'll start working on integrating mythweb functionality into the internal http server ... more in line with my personal goals :)
[19:41:22] ** skd5aner claps **
[19:42:58] stuartm: starting with the guide, since that's at least 75% of mythweb
[19:43:46] stuartm: if the beer lasts I aim to have something to show by the end of the weekend
[19:46:11] stuartm: probably won't be fancy, javascript will be kept to a minimum, but I think emulating (copy/paste will be involved) the mythweb guide should be achievable in that time-frame
[19:46:46] warped_ (warped_! has quit (Quit: warped_)
[20:00:42] stuartm: dblain: so qsp qjs, is that documented anywhere?
[20:01:14] skd5aner: stuartm: scheduling recordings by search (and managing existing recording rules) is what I use mythweb for 80% o the time
[20:01:36] skd5aner: off to a funeral :(
[20:01:39] stuartm: skd5aner: I was talking in terms of the code rather than usage
[20:01:55] skd5aner: gotcha
[20:02:34] stuartm: skd5aner: sorry to hear that
[20:03:53] bill6502: mythweb is checking to see if the icon exists, and if it does, it uses it. If I remove the icon, THEN mythweb sends ...Guide/GetChannelIcon?ChanID=1021. And, fails because the file/icon isn't there. 404 error in the BE log with upnp logging on.
[20:06:14] bill6502: would we really want a minimum of 10 TCP messages( SYN, SYN-ACK, ACK,PSH,ACK,PSH,ACK,FIN-ACK,FIN-ACK,ACK) at a minimum for every icon if the file is already there...
[20:08:43] dblain: stuartm: No wiki or official docs yet. I did write a quick sample page that explains the basics of it that's in the backend already. (serverside.qsp)
[20:09:17] stuartm: dblain: got it, thanks
[20:10:00] dblain: Basically any of the extensions you listed allows for embedded javascript that will run on the server. All service API classes are accessible as script objects (no http calls required)
[20:10:51] dblain: I don't remember the state of the translation support. We played with different ways to allow for translation, but I don't remember how we left it.
[20:11:51] dblain: stuartm: I'm interested in an embedded "MythWeb" like interface. Let me know if I can help with anything.
[20:12:46] stuartm: dblain: fantastic
[20:13:34] stuartm: what is this country coming to, they've edited out the violence from an episode of the Simpsons?!?
[20:14:23] stuartm: I mean, that's unheard of
[20:26:57] bill6502: stuartm: About that beer: typo in: modules/tv/classes/Channel.php: httpRequest('Guide/GetChannelIcon', array('ChanID'. It's ChanId<-----------. Works A-OK.
[20:27:50] stuartm: bill6502: for that, I'll throw in a second pint
[20:29:23] bill6502: probably needs to be a virtual beer, shipping charges to Chicago etc.
[20:29:32] stuartm: that means it's been broken since it was converted to use the services API, and no-one spotted the problem ... sadly sums up the level of attention mythweb has received in the last couple of years
[20:29:58] ** dblain wonders if it would be better to make all query strings case-insensitive...? **
[20:30:05] ** dblain wonders if it would be better to make all query strings case-insensitive...? **
[20:30:13] dblain: oops.
[20:30:56] stuartm: dblain: it would be less prone to errors
[20:32:07] jheizer: As a services consumer it hasn't been bad. I have only been tripped up on it a time or two.
[20:32:17] jheizer: Much more likely that I just typo it completely.
[20:32:24] stuartm: can't really see a downside to case insensitivity considering that the services API is meant to be accessible to as many people as possible
[20:32:38] jheizer: Also very true.
[20:32:53] jheizer: Ex: 15 minutes last night wasted over a missing 'e'
[20:33:57] dblain: Might be able to expand on the error to explain what's wrong as well. I'll have to take a look at it.
[20:34:07] stuartm: if it was just for internal use, well I'm pedantic enough to prefer an API that requires the extra care and attention, but an api used by third parties could afford to be less rigourous
[20:34:29] bill6502: i'd suggest it's more of a consistency issue, GetChannelInfo is the only one one using ID rather than Id
[20:35:00] jheizer: Ahh yeah. Rest are Id.
[20:35:34] stuartm: it's using ChanId, except mythweb was using ChanID
[20:35:52] stuartm: or is – haven't fixed it yet
[20:36:24] jheizer: bill6502, hmm, I'm 100 miles south of you.
[20:36:41] dblain: The service's exposed as a whole could use a consistency check. It's grown organically with each developer adding what they need. (not always checking to see how things were named in other services)
[20:37:38] stuartm: dblain: that would be good
[20:38:35] stuartm: bill6502: sorry, I should have credited you in that commit message
[20:38:37] dblain: I guess, if we removed the case sensitivity, then changes like ChanID -> ChanId wouldn't break existing applications.
[20:38:56] cbovy (cbovy!~cbovy@ has quit (Ping timeout: 245 seconds)
[20:39:22] dblain: Can anyone think of any security issues with the proposed change?
[20:40:36] bill6502: stuartm, np, liked your solution much better. I wrote it so that either full paths or filenames only would work. Yours forces folks to use the correct format.
[20:40:42] stuartm: dblain: nothing comes to mind, for ascii at least
[20:42:55] stuartm: dblain: are you proposing the change for the arguments only, or the method names as well?
[20:43:09] natanojl (natanojl!~jonatan@mythtv/developer/natanojl) has quit (Ping timeout: 272 seconds)
[20:44:18] dblain: stuartm: I was thinking of method names as well, but not set on it. (Part of me wants even clients to use the correct case/name... but I don't feel stronge enough about it to not make the change)
[20:44:55] dblain: Since technically the RFC for url's states that they are case sensitive (other than hostname)
[20:46:22] stuartm: I was initially think arguments, since it's not unreasonable to confuse something like ChanID with ChanId – ID is usually capitalised
[20:46:26] Chutt (Chutt! has quit (Read error: Connection reset by peer)
[20:47:20] stuartm: seeing people using getchannelicon on the other hand would be hard ...
[20:47:32] dblain: okay. Either way, I'm going to see how hard it would be to return meaningful errors... Something like "Error: Parameter 'ChanId' not valid"
[20:48:11] dblain: That's just ugly!
[20:50:05] stuartm: not just me then :)
[20:52:04] bill6502: dblain: from testing on the MythTV Android Frontend project, I can tell you that the error text isn't read. But if your idea also appeared in the BE log, probably with VB_UPNP, that would be great!
[20:56:54] dblain: bill6502: definitely can log it, but I'm thinking having the error in the reply body will help during development, and packet captures.
[20:59:41] bill6502: dblain: i've done that, but it required shutting off gzip to be useful. not a big deal, just a new download to the tablet/phone. i used to have an environmental variable in the MythTV code e.g. DONTGZIP=1...
[21:01:07] dblain: bill6502: If you're debugging and have control of the application, it's just a matter of not specifing gzip in the header.
[21:01:30] Chutt (Chutt! has joined #mythtv
[21:01:54] ** dblain is heading home **
[21:05:42] Jordack (Jordack! has quit ()
[21:09:21] stuartm: bill6502: there is also HTTPREQUEST_DEBUG=1 which dumps the entire response to stdout
[21:09:30] stuartm: before it's compressed
[21:13:35] bill6502: i saw that (in a recent commit?) and tried. perhaps because I rarely start the BE in a terminal session it didn't show anything. I'll give that a try, THANKS!
[21:24:48] stuartm: master only atm
[21:25:26] bill6502: np, i'm on it
[21:25:33] stuartm: and yeah, only useful if you can actually see the output :)
[21:26:37] stuartm: I found it useful for debugging some upnp stuff, it was easier than grabbing the raw packets in wireshark
[21:28:15] bill6502: I'm updating the Wiki by the way (for channel icons) right now
[21:46:53] jheizer_ (jheizer_! has joined #mythtv
[21:48:14] SteveGoodey (SteveGoodey! has quit (Quit: Konversation terminated!)
[21:48:24] stuartm: dblain: is there any logging for errors in the scripting?
[21:51:28] stuartm: dblain: nm
[21:51:37] dblain: stuartm: Nothing that I introduced. I'm not sure what's available from Qt. Most Qt objects/functions are available in scripts
[21:52:45] ** dblain installing windows 8.1... rebooting **
[21:52:47] stuartm: there's a basic 'parse error' with a line number, nothing more specific currently implemented, but I wasn't even interpreting the number given as a line number – assumed it was an error number
[21:53:06] stuartm: line number at least points me in the right direction
[21:53:43] dblain (dblain!dblain@mythtv/developer/dblain) has quit ()
[21:53:59] wagnerrp: stuartm, Captain_Murdoch: have at it
[21:56:35] dekarl1 is now known as dekarl
[21:59:07] dekarl: stuartm: if you can come up with a good way to signal "channel is off air" that would come in handy for stuff like #11624 – filtering "SIGN OFF" from new movies
[21:59:07] ** MythLogBot **
[22:00:52] dekarl: stuartm: and no, no sevices for me :/ (I'm just guessing at the repo urls atm)
[22:01:48] stuartm: well there are different types of 'off air', there's "channel is not broadcasting anything" and there's "channel is broadcasting a holding page/mheg because there is currently no programming"
[22:31:21] stuartm: great, just crashed the backend through QTScript
[22:33:34] stuartm: throw( "StartTime is invalid" );
[22:34:16] jheizer_: It is a known response that a Services API call to GetFileList on a non-existant Storage Group returns all recordings?
[22:35:51] jheizer_: I got mixed up on which backend I was calling against and was confused by the results haha
[22:35:52] stuartm: no idea, but it doesn't seem right – providing a response even to an invalid request just masks programmer errors
[22:36:24] jheizer_: Agreed. And it is only recordings files, not videos, etc.
[22:40:03] jheizer_: Hmm, actually even if it exists but is "non standard" maybe, it returns recordings.
[22:40:37] jheizer_: I realized I could dev the new gallery into MobileMyth with .27 as all the calls are there already as long as I created the storage group like master.
[22:41:14] jheizer_: At first I was just hitting the .27 vs master backend, but now that i created it it still returns recordings.
[22:41:21] stuartm: I really don't like exceptions, miss one try ... catch and you kill your app, for what is rarely a fatal error
[22:41:34] stuartm: and in the backend it's just dumb
[22:52:51] knightr: but when they are correctly implemented (like in other languages) they can do the exact reverse...
[22:59:52] bill6502: jheizer_: mythbackend --setverbose file will show you it's looking in the Default group: Unable to find storage group 'NameThatDoesntExist', trying 'Default' group
[23:00:37] bill6502: was the log for: http://mc1:6544/Content/GetFileList?StorageGr . . . tDoesntExist
[23:07:11] stichnot (stichnot!~stichnot@mythtv/developer/stichnot) has quit (Ping timeout: 272 seconds)
[23:15:57] bill6502 (bill6502!~bill@ has left #mythtv ()
[23:17:22] jheizer (jheizer! has quit (Read error: Connection reset by peer)
[23:17:50] jheizer (jheizer! has joined #mythtv
[23:41:50] knightr (knightr!~Nicolas@mythtv/developer/knightr) has quit (Read error: Connection reset by peer)
[23:44:57] stichnot (stichnot!~stichnot@mythtv/developer/stichnot) has joined #mythtv
[23:51:44] knightr (knightr!~Nicolas@mythtv/developer/knightr) has joined #mythtv
[23:54:54] stuartm: anyone know much about ecma script? Can't get it to parse an ISO date string, yet it will happily parse something like "Wed, 09 Aug 1995 03:14:05 GMT"
[23:55:41] stuartm: the ecma docs are clear that all implementations must handle ISO strings at a minimum, yet the QT one just fails consistently
[23:57:11] stichnot: I'd really like to design a nice way to make in-playback adjustments "stick" after I stop playback and resume later. E.g., playback position, timestretch, audio sync, subtitle track, subtitle delay, fill mode, aspect ratio, picture levels, etc. The bookmark serves for playback position, but the rest have to be readjusted each time. Playback groups kind of work for timestretch if you can...
[23:57:12] stichnot: ...think ahead. But it's not clear what the interface to control this should look like, or which attributes should be global versus per-frontend.
[23:57:32] stuartm: about 5 minutes short of ditching the QTScript and just implementing it in plain C++
[23:59:16] stuartm: stichnot: adjustments per recording, or applied to all recordings?
[23:59:36] stichnot: per-recording

IRC Logs collected by BeirdoBot.
Please use the above link to report any bugs.