Wednesday, 2014-06-18

*** tpb has joined #timvideos00:00
mithromparuszewski: it is likely I'll be around tomorrow morning earlier, I'm going to try and make some very early meetings at work but haven't had much success doing so in the past.00:00
ayush3504mithro: another thing, for the measurements, I'm not able to find any photographs of the digilent boards that have been taken from top. They're all taken from an angle that makes it impossible for me to measure that way. So the only way I see is to ask people and Digilent support.00:03
mithroayush3504: I don't think the angle makes much difference00:03
mithroat the size, perspective doesn't make a huge difference?00:04
ayush3504mithro: I don't know depends on what we consider as huge. The ratios of the sides would change unless it's a front/top view.00:06
mithroif you have the right dimension within +/- a couple of millimeters that should be a good start00:07
ayush3504mithro: will try00:07
mithroayush3504: do your best and include information on your errors00:08
ayush3504mithro: yep00:09
ayush3504mithro: for the bidirectional thing, how about giving it as an option, because we always have rx, tx pairs anyway and even the drivers come in pairs. Should we let it be an option to populate (or connect through wire) the board with an optional input connector?00:11
mithroayush3504: so, first I think we need to think about the idea of daughterboards working stand alone, versus requiring the motherboard to work00:17
mithroayush3504: if we go the route of requiring all daughterboards to work stand alone accepting RS232 input then the RX/TX for MIDI + DMX is going to require a microcontroller on the daughterboard to do the translate from RS232 to MIDI/DMX00:18
ayush3504mithro: agree00:19
mithroayush3504: otherwise the microcontroller could "sit inside" the FPGA00:19
ayush3504mithro: RS232 to MIDI doesn't need translation00:19
ayush3504mithro: IR needs, if not using the IrDA00:20
ayush3504mithro: for midi<->rs232, take a look at this
mithroayush3504: does that actually work? IE are the voltages actually spec compliant or only "kinda works" ?00:23
ayush3504mithro: the only difference is midi operates at some fixed baudrate and uses slightly different current/voltage levels00:23
mparuszewskimithro: I am back. :)00:23
ayush3504mithro: have to check that, that's not a circuit i'm treating as final00:23
mithroayush3504: okay00:24
ayush3504mithro: you mentioned about midi having some other connector besides xlr-3?00:25
mithrobe back in 5, need to reboot computer00:26
mparuszewskimithro: Can we talk about all things tomorrow 10pm UTC? :) I feel really tired, sitting at a computer trying not to sleep.00:31
mithromparuszewski: lets try and do that, I can't guarantee I'll be around then00:33
mparuszewskiIf you have time now, we can also talk now. I thought that you are talking with ayush :)00:33
mithromparuszewski: we can try and catch up tonight too00:34
mithromparuszewski: when will you be back?00:34
mithromparuszewski: I can multitask :), but if your tired I think it's better to post pone it00:36
mithroayush3504: search for "midi sound card connector"00:36
tpbTitle: MIDI soundcard - keyboard - AllPinouts (at
mparuszewskiGreat, thank you. :) I can talk tomorrow from 3 pm utc to 11-12 pm00:38
ayush3504mithro: Ah, I know that one. I think we should not worry about the connectors in midi (and dmx) because we can solder them. I could not find any particular model of Right angled, PCB mount XLR-3 connectors that are well-stocked. So we will be manually soldering the connectors anyway.00:39
mithromparuszewski: what is your timezone again?00:40
mithroayush3504: the MIDI soundcard connector seems semi-standard for a PCB mountable connector?00:40
mithroayush3504: would it fit on a daughterboard at the moment?00:40
ayush3504mithro: No, because my daugherboard width is limited (as of now) with DB9 connector width which is 31 mm. DB15 is going to be wider for sure.00:41
mparuszewski+2 UTC, its 2.40 am now :)00:41
mithromparuszewski: is that Warsaw?00:42
mithromparuszewski: 3pm utc seems to be 11pm for me, so if I'm going to catch the early meetings tomorrow I should be in bed by then00:43
mithromparuszewski: lets aim for sometime between 6am (10pm your time) and 9am (1am your time) tomorrow?00:44
mparuszewskiOk, I will be available then for sure. :)00:45
ayush3504mithro: and I added 10 mm more to the 31 mm so that we have some extra clearance (5mm each side) for the cables. There's always some extra plastic 2-3mm at the sides of the DB9 connector in cable. And there needs to be some space for using the thumbscrews.00:45
mithroayush3504: how much wider would the daughter board be to fit that MIDI connector?00:45
mithroayush3504: what about the good old PC parallel port?00:45
mparuszewskiSee you tomorrow, have a nice day, bye! :)00:46
mithromparuszewski: sleep well!00:46
ayush3504mparuszewski: bye :)00:47
ayush3504mithro: parallel port? that's like 3 times a serial port00:47
techman83mparuszewski: cya!00:47
mithroayush3504: could we have "double width" daughterboards?00:48
mithroayush3504: (IE daughter boards which cover two "slots") ?00:48
ayush3504mithro: hmm, we sure can. But i wish to stick with soldering for another reason. The PCB mount connectors are more expensive than the panel mount type.00:49
mithroayush3504: when you say "soldering" you mean cables from the holes to the connectors?00:50
ayush3504mithro: i'm talking about connectors like these:
ayush3504mithro: these would need to be soldered to the daughtere board through a ribbon cable00:51
mithroayush3504: I'm confused?00:51
ayush3504mithro: so if we go with these type of connectors instead of PCB mount ones, we may not need to consider increasing width of daughterboards00:51
ayush3504mithro: if you see the back pins of these connectors we need to solder wires on it. It isn't designed to be soldered on pcb.00:52
mithroayush3504: okay, I think for anything apart from the XLR connectors we probably want to go PCB mount?00:55
JoelwSlightly less annoying than solder-type DB connectors is using IDC connectors, e.g.
ayush3504mithro: another point, when you panel-mount a connector, any stresses during cable connector would be handled by the chassis. But in can of PCB mount, it will be the daughterboard PCB handing the stress. If you want a PCB connector that is also fastened to the chassis it will require us to have tight tolerances of dimensions.00:56
ayush3504mithro: two typos above: connector->connection, can->case00:58
mithroayush3504: I think we probably still want to go the PCB mounted, we have pretty tight tolerances for the Atlys part of the case if using a model where the connectors are exposed directly rather than broken out.01:00
mithroayush3504: it also gives us the option of doing non-direct by using short cables01:02
ayush3504mithro: another point: the daughter board isn't directly fastened to the chassis. We're relying on the soldered headers for the mechanical support. So it's sensitive to stress. I've seen that over time, header pins can slip against their soldered joins no matter how nicely they have been soldered.01:03
ayush3504mithro: yes, IDC connector looks good. I wonder if it's not too costly.01:03
ayush3504thanks Joelw01:07
JoelwIDC connectors shouldn't be a lot more expensive, and they're a bit easier to put together - I just crimp them in a vice, which is loads faster than soldering them up.01:09
JoelwYou can probably get them pre-made as well.01:10
mithroayush3504: so, where is your table of mechanical sizes?01:10
mithroayush3504: can you port your table in the serial expansion board to a spreadsheet so you can add a bunch more detail? IE We should include info on the that MIDI connector and parallel ports01:11
ayush3504mithro: ok, shouldn't we use comment section for this. coz we won't be adding more rows to this. Just saying, coz I like to have things in one place01:14
mithroayush3504: we will be definately be adding more rows to this01:14
mithroayush3504: plus more details01:14
ayush3504mithro: ok01:14
mithroayush3504: again, these are "living documents" they will change / adapt / grow as we get more information01:15
ayush3504mithro: ok.
tpbTitle: Connector specifications for serial expansion board - Google Tabellen (at
mithroayush3504: please fill in the details01:19
ayush3504mithro: sure, why do we need rj12?01:20
mithroayush3504: it is an alternative option for RS422/RS48501:21
mithroayush3504: you should include all the different connector options in that spreadsheet too01:21
ayush3504mithro: have you seen it used somewhere?01:21
mithroayush3504: IE the PCB mount verses IDC connector verses wiring01:22
mithroayush3504: not really, but it can't hurt to have the information in the spreadsheet01:22
ayush3504mithro: there's also a spreadsheet for other components, make sure we don't have to duplicate stuff in both of the spreadsheets
tpbTitle: Cost spreadsheet for Serial Expansion Board - Google Tabellen (at
mithroayush3504: great! You could combined the two spreadsheets if you want and just use two different sheets01:26
ayush3504mithro: ^ i will be filling this in detail when I start doing the schematics.01:27
ayush3504mithro: good idea, I forget about using sheets every time01:28
ayush3504mithro: GPIO and parallel ports look very similar, should we skip one of them or combine maybe?01:29
mithroayush3504: no, too much information is better then too little information01:29
ayush3504mithro: didn't get the context01:30
ayush3504mithro: i was talking about combining the daughter boards, we can get a parallel port to work as gpio right?01:32
mithroayush3504: I think they might be similar but the parallel port is going to be too wide I think01:32
ayush3504mithro: what would you be using parallel port for? because it's too wide01:32
mithroayush3504: GPIO type things, but as I said we might want double width daughterboards01:33
ayush3504mithro: for GPIO type things I think the screw headers are more convenient01:34
mithroayush3504: possibly but I don't want to rule out the parallel header just yet.01:35
ayush3504mithro: anything that works for standard parallel port may not work either, because it's all emulated through a serial port i guess.01:36
mithroayush3504: mayb e01:40
shenkimithro: re: fx2, yeah, if i had infinite amounts of free time i'd re-write the firmware on a cortex m001:46
shenkiayush3504: hey01:46
shenkiayush3504: mithro said you're looking through the linux usb serial driver01:46
shenkiayush3504: what are you trying to figure out?01:46
ayush3504shenki: hey :)01:46
mithroayush3504 / shenki: we need to figure out if the usb-serial driver in Linux already supports multiple serial ports on a single end point01:47
mithroshenki: I'd like ayush3504 to look into it and ask you questions about it01:49
shenkiim trying to figure out what that would look like01:50
ayush3504shenki: we don't seem to have much endpoints on the FX2, so we want some way to multiplex these serial ports over a pair of In/Out end-ponts01:51
mithroshenki: I'm guessing the data we send to the USB controller with some type of "send to this actual port" prefix01:51
shenkiwe shouldn't be looking to create yet another usb serial device. usb cdc is the way to go01:53
mithroshenki: usb-cdc doesn't support that idea01:55
ayush3504shenki: yes but the usb cdc as specified in specs takes 3 endpoints - Int In and Bulk In/Out and we don't have that many left on fx2. So we're looking for a workaround.01:55
mithroshenki: so I think we want see if there is a device supported by the usb-serial driver that we could emulate01:56
ayush3504shenki: like this one -
tpbTitle: kernel/git/torvalds/linux.git - Linux kernel source tree (at
mithroayush3504: how is those mechanical information going?02:00
ayush3504my net died a few minutes ago. will be back in a while.02:04
mithroayush3504: Google Docs have offline support if you enable it02:04
ayush3504yep thanks, but i need to surf web to fill in the info02:05
*** CarlFK has quit IRC02:11
CARAM_mithro: I have flumotion docker image mostly done if you'd like to take a look or give feedback02:17
CARAM_just some minor stuff left for watchdog / register02:18
mithroCARAM_: I probably won't get a chance to review it until the weekend at the earliest02:18
mithroCARAM_: so you should put instructions on your blog and then link it here02:18
mithroCarlFK might give it a try02:19
mithroaps-sids might also be another person to look at it (he is doing the flumotion porting to latest gstreamer)02:19
CARAM_mithro: sounds good.02:23
CARAM_here are the current instructions:
tpbTitle: Running a Flumotion Docker Image - Max Stritzinger (at
CARAM_as stated, the current use of pushconfigs is undesirable, so my next step is to cut that out02:56
*** mparuszewski has quit IRC03:11
*** CarlFK has joined #timvideos03:34
*** ChanServ sets mode: +v CarlFK03:34
*** Niharika has joined #timvideos04:35
mithroayush3504: how is it going?05:31
ayush3504mithro: just signed in now from another place05:31
ayush3504mithro: can we talk about the daughter board stacking idea?05:32
mithrodaughter board stacking idea?05:32
ayush3504mithro: you mentioned about stacking the daughterboard layout on the slot?05:32
mithroayush3504: you mean including the idea of including RS232 daughterboard on the mainboard so it can be used without any daughterboards?05:33
ayush3504mithro: there wasn't any daughterboard for rs232, it's was always there as a default option.05:34
mithroayush3504: we talked about this last night, we should design rs232 as if it was a daughter board, then just copy the daughterboard onto what would have been the blank space for the daughter boards05:35
ayush3504mithro: but then, will have to send rx tx ttl signals straightaway to the db9 footprint. So db9 footprint wouldn't make sense05:37
mithroayush3504: we also talked about that - the idea of just having TTL signal into the daughter boards and making the daughterboards compatible with TTL or RS232 signals05:38
ayush3504mithro: and that wasn't clear to me, and still isn't. so you said you'll make a diagram to explain05:39
mithroayush3504: what is not clear? The signals from the board motherboard to the daughterboards being TTL?05:40
mithroayush3504: that also solves the duplicate RS232 drivers05:40
ayush3504mithro: i have an easier solution to the duplicate rs232 driver problem, it's just to short the rx, tx pins through jumper resistors05:41
ayush3504mithro: with the line driver ic pads unpopulated05:41
mithroayush3504: but then the motherboard needs to be modified depending on the daughterboards which are connected05:42
mithroayush3504: special cases are bad05:42
ayush3504mithro: but when sending ttl to the daughter board it doesn't make sense to have DB9 connectors. And I don't know how to make the daughterboard compatible with rs232 as well as ttl.05:43
ayush3504mithro: ^ using the same rx. tx lines05:44
*** mparuszewski has joined #timvideos05:44
ayush3504mithro: can we have a dedicated coupling header on the daughter board to to accept ttl signals? and leave the db9 to accept rs232?05:46
mithroayush3504: that is an option05:46
mithroayush3504: another option is to do some research and figure out the RS232 to TTL conversion methods05:46
mithroayush3504: you should be able to think up a circuit to do the 12V -> 5V and -12V -> 0V05:51
ayush3504mithro: yes but doesn't the line driver do this?05:51
ayush3504mithro: my concern is supporting TTL and RS232 logic levels on the same pins05:51
ayush3504mithro: which looks difficult and optional provided we can use another header for TTL stuff.05:52
ayush3504mithro: i don't see any harm in putting a dedicated ttl header. and that solves the issue of customizing motherboard05:54
mithroayush3504: putting a dedicated ttl header is an option, but how do you keep your daughterboards should also work independently?05:55
ayush3504ayush3504: by populating DB9 female and line drivers in the daughterboard.05:55
ayush3504mithro: e.g. for a rs485 daughter board the daughterboard would have RS232 port>rs232 line-driver>-->rs485 line-driver>rj45 jack>05:58
ayush3504mithro: between the two line drivers the signals are ttl and will be also connected to the ttl header05:58
ayush3504mithro: so if we use the daughterbaord with motherboard, we can skip the rs232 port and line driver05:59
mithroayush3504: that sounds reasonable05:59
mithroayush3504: so a "rs232" daughterboard would be just a TTL->RS232 driver then?06:00
ayush3504mithro: yes, that can be an option. or you can just populate a DB9 male on the front footprint on motherboard and populate the line driver on the motherboard06:01
ayush3504mithro: by front footprint i mean the one that is outwards06:02
mithrodo any of the daughterboards need the line driver apart from the rs232?06:03
ayush3504mithro: midi06:04
ayush3504mithro: ir, if you use IrDA protocol06:04
ayush3504mithro: err , no not the ir06:04
mithroayush3504: this information should be in connecter specification06:04
mithroayush3504: as I see at the moment, it's rs232 and midi?06:05
mithroayush3504: so it seems better to think of them as part of the daughterboards then?06:05
ayush3504mithro: you mean combining these two?06:06
mithroayush3504: no06:07
ayush3504mithro: didn't get you06:07
mithroayush3504: I'm saying we shift the rs232 line driver to be part of daughter boards rather then the main board06:08
ayush3504mithro: ok, got it. makes sense06:09
mithroayush3504: once we have the daughterboards / motherboards all figured out, we can then look at how to make it so "empty" slots can be directly populated with rs23206:09
ayush3504mithro: for the empty space we can just clone in one of the daughter board layouts, so that will be like a default daughterboard, which is preferable rs23206:12
mithroayush3504: exactly, that is what I have been banging on about06:12
ayush3504mithro: :P06:12
ayush3504mithro: instead of having autodetection for each of these daughter boards, can we just program the EEPROM which we'll be using for serial expansion board?06:15
mithroayush3504: that is an open question06:16
mithroayush3504: how many daughter boards have we got now?06:16
ayush3504mithro: 8 including rs232 and parallel port and gpio06:17
mithroayush3504: with the TTL design, can we do the RC style auto-detection?06:17
mithrohow accurate would we need to be?06:18
mithroif we wanted to make sure we could have up to 32 different daughter boards?06:18
ayush3504mithro: yes, that'll work but why do we need it if we can program our customization in the eeprom.mcu in serial expansion board06:18
*** wanig has quit IRC06:19
*** techman83 has quit IRC06:19
mithroayush3504: I'm talking about the daughterboards on the serial board, what if someone changes a daughterboard and forgets to reprogram the EEPROM?06:20
*** wanig has joined #timvideos06:23
*** techman83 has joined #timvideos06:23
*** sets mode: +v techman8306:23
ayush3504mithro: changing daughter boards won't  be so easy enough to forget though since we'll be soldering them? :p06:25
mithroayush3504: are we? Having dynamically swappable seems like a good option to preserve?06:26
ayush3504mithro: umm, okay06:26
mithroayush3504: we may want to go the solder route, but I don't think we can make that decision at the moment?06:27
ayush3504mithro: well in a future design we can have daughter boards / expansion boards that you can slide in and out, would look kind of fancy ;)06:29
mithroayush3504: the 2.54mm male headers have female versions06:29
ayush3504mithro: yes they have, but i recommended soldering to give some mechanical strength06:30
mithroayush3504: so how many pins are you looking at for the interface to daughter boards now?06:31
ayush3504mithro: 7 pins for the DB9, 10 pins on TTL header - rx,tx,rtc,cts,dsr,dtr,detect,5v_iso,gnd,NC06:32
mithroNC ?06:35
ayush3504mithro: not connected06:35
mithrodetect is?06:35
ayush3504mithro: rc06:35
mithroayush3504: which ones are for rs232, which ones are extra?06:35
ayush3504mithro: the first 6 of the ttl ones and ground06:36
mithroayush3504: so detect is for the auto-detect of the daughterboards?06:36
ayush3504mithro: yes06:36
mithroayush3504: okay, so how many daughterboards could you have with that?06:37
ayush3504mithro: more than 32 easily i guess06:37
mithrohow many pins on the VHDCI connector?06:38
ayush3504mithro: i have to see whether the logic level thresholds change with conditions like temperature06:38
ayush3504mithro: that's a good question. I had posted a question for you in the snippets - do you want h/w control in all of the ports?06:39
mithroayush3504: lets start with yes and then go back if we don't think we have enough ports06:40
ayush3504for each hardware control supported port, we'll need 6 pins06:40
ayush35042 pins for the I2C06:40
ayush35046 h/w controlled ports i guess06:40
ayush3504h/w flow controlled*06:40
mithroayush3504: so? (VHDCI pins - 2) / (6 + ??) == XXX ports?06:43
mithrothis should all be in your spreadsheet06:43
ayush3504mithro: sure06:46
ayush3504mithro: you may want to review the flow control heading in my document
tpbTitle: Serial Expansion board design - Google Docs (at
mithroayush3504: I don't think any of the stuff raised there is that relavent at the moment06:47
mithroayush3504: it will become relevant if we don't think we have enough ports06:47
ayush3504mithro: ok06:47
ayush3504mithro: on the PIC18F which pins should I be using06:50
mithroayush3504: don't worry about the PIC for now06:50
ayush3504mithro: it's a part of schematic right?06:50
mithroyes, but we'll get back to that in a second06:51
mithroayush3504: I'm going to disappear in about 40ish minutes06:53
ayush3504mithro: oh okay06:54
ayush3504mithro: any info about the IR receiver?06:57
mithroayush3504: the IR receiver is really just GPIO with IR LED / Receiver connected to two pins right?07:00
ayush3504mithro: yes, but why do you need IR LED?07:01
mithroayush3504: for remotes possibly, or maybe controlling TV / displays or other things07:02
ayush3504mithro: for that well be needing ir receivers right, not leds07:03
mithroIR LED is for sending, IR receiver is for receiving, right?07:03
ayush3504mithro: yep, so what will we send?07:04
mithro--> maybe controlling TV / displays or other things (pretending to be a remote)07:04
ayush3504mithro: HDMI2USB?07:04
mithroayush3504: say the HDMI2USB is connected to a cable box07:04
mithroayush3504: IR would be useful to cause the cable box to change channel07:05
ayush3504mithro: oh ok, and is it worth having a separate dboard for IR? coz we can do it with gpio?07:05
mithroayush3504: maybe not07:06
ayush3504mithro: should i rule out IR then?07:08
mithroayush3504: I think it would still be useful to have, as we have it on the production board07:09
ayush3504mithro: the only practical difference i see is using individual wires instead of a trs connector (for receivier)07:10
mithroand IR receiver generally needs an amplifier (unless it's built into the IR receiver)07:10
ayush3504it;s built into it07:10
ayush3504it's three pins, Vs, Gnd, Dout07:11
mithroit could be a very simple daughter board :)07:11
mithrothe more interesting part is how it mechanically would connect to a faceplate07:12
ayush3504mithro: do you plan to integrate the IR receiver on the dboard itself?07:12
mithroayush3504: ?07:12
ayush3504mithro: i'm assuming that the IR dboard will have connectors and the LED, Receiver will be connected through some cable?07:13
mithroayush3504: I was assuming they would be mounted on the PCB07:14
ayush3504mithro: IR receivers on market are that way07:14
mithroayush3504: that is another option, then the PCB would need a 2.5mm header07:16
ayush3504mithro: no they use 3.5mm trs jack, and that's pretty standard07:16
mithroayush3504: sorry, hear 3.5mm jack07:17
ayush3504mithro: a 2.54 mm screw header could be used for the LED, since there's no standard connector07:17
mithroayush3504: maybe07:18
mithroayush3504: so you still need to fill in all this information into the spreadsheet07:18
mithroayush3504: you've also not mentioned how many serial ports we can get with this current model07:18
ayush3504mithro: and for the dboards with screw header, we would probably want them to be hanging a bit outside the case outline so that we can make connections without opening the case.07:19
mithroayush3504: agreed07:22
mithroayush3504: do you think you have enough to be confident in designing the daughterboard template?07:22
mithroayush3504: then you should be able to do a rough design for each daughterboard07:22
ayush3504mithro: yes, i think so07:22
mithroayush3504: then look at the mechanical design of a front face plate07:22
ayush3504mithro: front face plate? isn't that the back plate?07:23
mithroayush3504: you need to populate all the spreadsheets with mechanical information so we can refer back to it at a later date07:23
ayush3504mithro: will do07:24
mithroayush3504: take a look at ->
tpbTitle: Serial expansion board sketch - Google Zeichnungen (at
ayush3504mithro: and what's the backview?07:29
mithroayush3504: your board doesn't go all the way to the back07:30
mithroI'm about to head off07:30
mithroany last questions?07:30
ayush3504mithro: umm, i'm still a bit confused about this07:31
ayush3504mithro: what's the front and back once again?07:31
ayush3504mithro: please draw a line parallel to the side that you consider "front"07:33
mithroayush3504: I labeled the with arrows07:34
ayush3504mithro: it looks like the right edge is the front, my board isn't on that side?07:35
ayush3504mithro: right edge wrt Top view07:36
*** mparuszewski has quit IRC07:51
*** Niharika has left #timvideos08:44
*** Niharika has quit IRC08:44
*** tija has joined #timvideos08:54
*** mparuszewski has joined #timvideos09:31
*** rohitksingh has joined #timvideos09:53
*** rohitksingh has quit IRC10:28
*** techdragon has quit IRC11:15
*** Joelw has quit IRC11:15
*** techdragon has joined #timvideos11:17
*** Joelw has joined #timvideos11:17
mithroayush3504: I'm just heading to bed any quick questions?11:45
mithroDid you figure out the front / back?11:45
mithroWell I'm going to sleep now. Gnight everyone.12:03
*** thaytan_ is now known as thaytan12:28
*** ChanServ sets mode: +v thaytan12:28
*** FeltonChris has joined #timvideos12:59
*** FeltonChris has quit IRC13:00
*** FeltonChris has joined #timvideos13:00
*** tija has quit IRC13:16
*** mparuszewski has quit IRC13:31
*** Niharika has joined #timvideos13:38
*** mparuszewski has joined #timvideos13:48
*** hyades has joined #timvideos14:06
*** rohitksingh has joined #timvideos14:11
*** slomo has joined #timvideos15:53
*** rohitksingh has quit IRC15:58
*** Niharika has quit IRC16:05
*** rohitksingh has joined #timvideos18:13
*** hyades has quit IRC19:39
mithromparuszewski: ping?20:35
mparuszewskimithro: pong :)20:36
mithromparuszewski: great, give me 5 minutes then we can talk20:38
mparuszewskiOK, I also jeed 5 minutes. :)20:39
mparuszewskiI am ready. :)20:47
mithromparuszewski: I'm just walking into work well be back in 15 minutes21:02
mithromparuszewski: poke?21:18
mparuszewskiReady to talk :)21:19
mithromparuszewski: okay cool21:20
mithromparuszewski: so I wanted to chat with you to understand your progress so far and to discuss some aspects of your project further (and help give you some guidelines to make the correct choices)21:21
mparuszewskimithro: ok, great.21:22
mparuszewskiWhat would you like to know? :) Currently I am working on People on Gettin Started site, I've integrated it with GitHub, some of information are downloaded from GitHub API on client side.21:24
mparuszewskiJust like in Issues21:25
mithromparuszewski: first thing, how close are we to being able to deploy your version of the new site?21:26
mparuszewskiIn my opinion we can do that even now, my version has same content as current deployed on, works correctly and has all functionalities21:27
mparuszewskiToday I've tested GitHub edit interface and works just like and dosen't need any servers or authentication and configuration, works out of the box.21:32
mithromparuszewski: I was going to ask about the edit workflow21:32
mithromparuszewski: how does the embedding work?21:33
mparuszewskiAre you asking about embedding iframe on website?21:35
mparuszewskiI am not sure if I understand :)21:35
mithromparuszewski: I go to and want to edit a web page21:36
mithromparuszewski: what happens now?21:36
mithromparuszewski: having a "nice" edit experience is one part of your project, right?21:38
mparuszewskimithro: Yes, on new website you are able to edit website only via GitHub edit interface or, if we want to have current edit experience with wiki it will need changes in design21:39
mparuszewskibecause every page was splitted into smaller parts (that are in _includes folder)21:40
mithromparuszewski: so what does that look like?21:40
mithroediting via / Github Edit interface is fine, but the person shouldn't need to know anything about those tools21:41
mparuszewskiCurrently if you want to edit website you need to go to GitHub edit interface or, select the file and edit it. But now I understand the problem21:42
mparuszewskiPerson shouldn't think about file which he want to edit, I need to add Edit links on website to make it easy21:43
mithromparuszewski: yeah21:44
mithromparuszewski: ideally the edit link should pop-up a model window inside the website21:44
mithromparuszewski: so you don't even leave the website21:44
mithromparuszewski: it would be nice if we could theme / control the edit interface too - which is why I think is probably the right solution here21:45
mithromparuszewski: editors might also be non-technical21:45
mparuszewskiOk, I will do that, I will display link on hover on element (panel is the element that user can edit)21:45
mithromparuszewski: but not sure if you can embed the github edit interface?21:46
mparuszewskiProse has definitely better preview options, GitHub edit only displays current markdown document21:48
mparuszewskiand its preview21:49
mparuszewskiso yeah, will be better, but I am not sure if it can be configured to work with 'includes', because these are small part and doesn't has layout21:50
mparuszewskithey are included into larger pages21:50
*** lns has joined #timvideos21:50
CarlFKlns: hi21:50
lnsCarlFK, what's up :)21:50
mparuszewskimithro: So maybe I will need to change design a little bit.21:51
CarlFKI dvswitch is great, but..21:51
CarlFKer. lns: dvsw..  but...21:51
CarlFKand I am perhaps one one of the most active users (I use it at least once a month, most others use it one event per year21:52
mithromparuszewski: can you explain the includes better?21:53
CarlFKlns: anyway, I consider dvswitch to be.. legacy ;)21:53
CarlFKlns: it is still being used in production because a replacement hasn't been proven to work (there are a few candidates, but no one has taken the time/risk to use them)21:55
CarlFKand this #chan is way more active than #dvswitch, and given we still use it, you will likely get more from here than #dvs21:55
lnsCarlFK, does "Tim's Videos" stream to Icecast?21:56
mparuszewskimithro: yes. I wanted to have nice design so I needed to split every page into smaller parts. For example, current page 'Summer of code' I split into 8 parts (, you can see 8 panels/boxes with content). Every panel has content loaded from '_includes' folder21:56
mparuszewskiin Jekyll, those small parts are just markdowns and are generated into html21:57
CarlFKlns: icecast has been replaced with something better
tpbTitle: timvideos/streaming-system · GitHub (at
CarlFKwhich has been used in production21:59
CarlFKmore than a few times.  better than icecast in a few ways: mutiple formats (ogv, webm, h2whatever...)  and there is a web site that makes a nice experience for the viewer: like summary of tracks (pycon had 6 tracks)22:01
mithromparuszewski: okay, this design kind of makes sense, but I don't think the structure is correct22:01
CarlFKand an irc client setup for each track, twitter falls... stu like that22:02
mithromparuszewski: can we make includes page level specific rather than having a global site include?22:02
mithromparuszewski: IE we have a and a "Summer-Of-Code" directory which has the include bits inside it for the ?22:04
mithromparuszewski: you still there?22:06
mithromparuszewski: still got lots more to talk about :P22:09
CarlFKmithro: can you whip up some docs for running   (like an example of setting up 1 track)22:23
mithroCarlFK: setup config.json, run pusher.py22:25
CarlFKsetup config.json .... um....22:26
mithroif you only want to push configs to the encoders use --no-collectors22:26
mithrothis file ->
tpbTitle: streaming-system/config.json at master · timvideos/streaming-system · GitHub (at
CarlFKis that setup?22:27
CarlFKI am hoping I need to give hostname/ip of collector and encoder, and everyting else can default to "test" or "main track" or something22:28
mithroCarlFK: everything fulls the information from that file22:29
mithroyou set the IP address or DNS names in the values22:29
mithro                "flumotion-mixer": false,22:29
mithro                "flumotion-encoder": false,22:29
mithro                "flumotion-collector": false,22:29
mithrosee further example in
CarlFKah, that's what that means22:30
tpbTitle: streaming-system/config.private.json.example at master · timvideos/streaming-system · GitHub (at
mithromixer == DVSwitch mixer thing22:31
mithroencoder == flumotion encoder machine running on EC222:31
mithrocollector == flumotion collection machine running in your venue22:31
*** rohitksingh1 has joined #timvideos22:32
*** rohitksingh is now known as Guest5581522:32
*** rohitksingh1 is now known as rohitksingh22:32
CarlFKthat isn't obvious22:32
CarlFKnor am I sure where to document it22:32
CarlFKother than my dump of notes.txt wich is pretty bad22:33
mithroCarlFK: yeah the "flumotion-mixer" is probably a bad name and should be fixed22:34
mithrobut the collector / encoder naming scheme is used throughout the system22:34
CarlFKit also isn't clear what kind of value goes there22:34
mithroCarlFK: see second diagram at
tpbTitle: timvideos/streaming-system · GitHub (at
CarlFKthat doesn't at tell me how to run pusher22:35
mithroCarlFK: you do "python" and it pushes the configs based on your config.json file22:36
*** rohitksingh has quit IRC22:37
*** rohitksingh has joined #timvideos22:38
CarlFKchange :false to : #ip or hostname of collector22:38
mithroIt needs to stay false otherwise the value is not skipped. The config.private.json.example is your guide22:39
CarlFKthat's not a good guide22:39
CarlFKor at least something could be done better22:39
CarlFKthey are good examples, but it doesn't work as a guide that walks someone though the things that need to be done22:41
CarlFK   "channel": "#timvideos",  - someone may want to change it, but it doesn't need to be22:41
CarlFKwhich is the case for all the settings excpet the 3 you mentioned above22:42
*** rohitksingh has quit IRC22:43
CarlFKI'll try to set it up in an hour or so and take notest22:43
mithroCarlFK: true, we need a "Setting Up" section22:48
CarlFKwe have all the materials to build a house, but no blueprints.22:48
mithroCarlFK: patches welcome23:06
CarlFKthere isn't anything to patch :p23:06
mithroCarlFK: sure there is, documentation can be patched23:07
mithroCarlFK: adding new files is patching too23:07
CarlFKbut yeah, I am trying - but I keep getting so lost my scribble_notes.txt I don't even have faith in23:07
mithroCarlFK: dump them into a file in a pull request and I'll correct them if they are wrong23:08
CarlFKoh goodnesss.. but ok23:08
*** CarlFK has quit IRC23:09
*** FeltonChris has quit IRC23:29
mithroayush3504: ping?23:37
mithromparuszewski: I assume you fell asleep?23:37
*** lns has left #timvideos23:40

Generated by 2.12.1 by Marius Gedminas - find it at!