#milkymist IRC log for Monday, 2011-04-11

wpwraklarsc: (weird bugs) perhaps there will be a few early retirements soon after llhdl hits the streets :)00:13
wpwraklarsc: or quite a lot of night shifts00:13
xiangfulekernel: http://en.qi-hardware.com/wiki/Milkymist_One_accessories#keyboard01:06
sh4rm4just found that: https://www.opensparc.net/opensparc-t2/index.html01:38
sh4rm4an opensource ultrasparc cpu ^^01:38
sh4rm464bit even01:38
sh4rm4and verilog01:41
xiangfuwhere is the "rtems-milkymist " upstream. I cann't access "https://github.com/fallen/rtems-milkymist" anymore01:55
xiangfuso this one "git://github.com/milkymist/rtems-milkymist.git" is became the upstream now right?01:57
wolfspraulthere is also this one https://github.com/kristianpaul/rtems-milkymist02:00
xiangfuI saw the new code about ethernet all goes to git://github.com/milkymist/rtems-milkymist.git02:10
xiangfuseems all git repo under lekernel/*, now switch to new URL. git://github.com/milkymist/*02:11
xiangfugit://github.com/lekernel/* -- changed to --> git://github.com/milkymist/*02:11
Action: xiangfu will include all files under "https://github.com/milkymist/datasheets-m1" to data partition. 02:13
xiangfufor new the data partition include all patches "https://github.com/milkymist/flickernoise/tree/master/patches" and documents "https://github.com/milkymist/datasheets-m1" and only one wallpaper: http://www.milkymist.org/wiki/index.php?title=File:Comet.png02:15
wolfspraulxiangfu: you mean you want to put the datasheets onto the m1 flash? why that?02:15
wolfspraulI would think most are PDF, but we will not build the PDF viewer right now as it is far too slow in bootup time. And then - why would someone want to read datasheets on the m1?02:16
xiangfuyes. for fill the data partition.02:16
xiangfuwolfspraul: my plan is build flickernoise and flickernoise-pdf. by default the reflash_m1.sh will only flash the flickernoise (no pdf build-in)02:18
xiangfuthose datasheet maybe not a good idea for end user.02:18
wolfspraulI think they are completely useless on the m1.02:20
wolfspraulI also wouldn't bother building the PDF image, because with a startup time of 45 seconds or so that is way too much, it's only a prototype/alpha feature 'because we can'. other than that it has no value right now, imho.02:21
wolfspraulI'd rather focus on improving the non-pdf version, and of course keeping the datasheets out unless I totally misunderstand how someone might want datasheets on his m1...02:21
kristianpaulI wasnt able to run a patchwith last firmware, i mean, gui is okay but when i just run the patch there was nothing on screeN03:13
kristianpaulI was using a video projector,03:13
kristianpauli dont have the brand or model, was a quick presentation for flisol03:14
xiangfukristianpaul: you mean this one "http://www.milkymist.org/msd/msd-apr2011.tar.bz2" or last commit?03:15
xiangfukristianpaul: the msd-apr2011.tar.bz2 works fine here.03:21
xiangfuI am using normal LCD monitor.03:21
kristianpauli dont, well,  may be was the video projector03:24
kristianpaultomorrow i will finally have a meeting with some "VJ" related guys here,so i hope this will work there03:25
xiangfuthe Video-in Patch is cool.03:25
kristianpaulas i said, i just lost image when rendering the patches i choosed03:25
kristianpaulvideo-in oh yes03:25
kristianpauli did some tests, but i could set the videocamera so ijust showed some milkdrops patches and people just said,:o wow03:26
xiangfu(btw: the memcards leave Beijing at "2011-04-08  00:59:12")03:30
xiangfukristianpaul: I am try to build all images in my pc. base on last commit.03:31
kristianpauloh well, i'm out home for some days, i'll let you know as soon something arrives03:31
kristianpaulxiangfu: did you worked ona quick way of re-build rtems isnt?03:32
kristianpaulgood, i'll try out soon, :-)03:32
xiangfukristianpaul: http://projects.qi-hardware.com/index.php/p/m1s/source/tree/master/build03:33
xiangfuneeds some more command if build from scratch  (git clone all source code under github.com/milkymist/*)03:34
wolfspraulkristianpaul: I wouldn't expect that letter to arrive until late April, early May.03:39
kristianpaulsure, i'm relaxed about that :-)03:41
xiangfuhttp://downloads.qi-hardware.com/software/images/Milkymist_One/latest/  :)07:04
xiangfuwolfspraul: http://en.qi-hardware.com/wiki/Milkymist_One#Screenshots  :)07:53
wolfspraulxiangfu: nice! about the GUI/render screen resolutions, what are your current binaries doing now?07:58
wolfspraulis the GUI in 1024x768, and then it switches to 640x480 for rendering, and back to 1024x768 when it goes back to the GUI?07:59
xiangfuby default it's 1024x76807:59
wolfspraulor is it always staying in 640x480 ?07:59
wolfspraulmy projector can only support up to 800x60007:59
wolfspraulnot sure what happens if m1 displays a 1024x768 signal, probably the projector will just stay dark07:59
wolfspraulwhen we are in the GUI, can we have some key-press to iterate through resolutions?08:00
xiangfuwolfspraul: yes. it's switch to 640x480 when rendering08:03
xiangfuwolfspraul: yes. we can switch 640x480, 800x600, 1024x76808:03
xiangfuthere are three buttons in "System Settings"08:04
wolfspraulyes but if my projector can only do 800x600, I will have a hard time switching if it boots in 1024x76808:04
wolfspraulbecause I cannot see anything :-) (need to try, but that would be what I expect)08:05
wolfspraulso either we boot in 640x480 and let people increase manually, or if we think most people will have support for 1024x768, then we can boot in that, but there should be a keypress to go down for poor souls like me08:05
wolfspraulin fact the keypress should just iterate through all available/hardcoded resolutions, imho08:06
xiangfugood idea, keypress iterate.  what keys you recommend?  F11 or Ctrl + F1108:08
xiangfuthe big video-in 360x288 make m1 very slow.08:09
xiangfuI have to move the mouse very carefully08:09
xiangfuI am working on add a switch button on video-in08:10
rohwolfspraul: i am just ordering the aluminium sheets... do we want the thick (0.3) EN AW-Al Mg 3 or the thin (0.2)  EN AW-Al 99,5 ones?08:10
wolfspraulforgot which one we preferred again08:11
wolfspraulroh: the harder one I guess, 0.3 almg308:12
wolfspraulroh: did you see the feedback about ST ANDBY? and the lack of labels for the 3 buttons?08:12
rohdo you remember how much you payed?08:12
rohthe labels on the buttons were not there on purpose.08:12
roh the ST ANDBY i can't explain08:13
rohits not there on the ones i checked08:13
wolfspraulok then. I can see it on my wood one too. it looks like the T is shifted a bit up and leftwards.08:13
rohthe labels were ommitted since they either are different to the ones on the pcb or the numbers go from right to left, which is counterintuitive08:13
wolfspraulthat doesn't sound like a good argument to me.08:14
wolfspraulso because we are confused about the labels ourselves, we leave out any way to identify them? :-)08:14
rohits sw-defined anyhow.. left, middle, right?08:14
rohin the dark nobody sees buttons anyhow08:14
wolfspraulactually I thought render/on/standby are the labels for the buttons08:15
wolfspraulI don't think I'm the only one.08:15
wolfspraulso I press the middle button, because I think that's ON08:15
wolfspraulwhich actually works08:15
wolfspraulthe buttons definitely need some type of label, imo08:15
wolfspraullet's check with Sebastien08:15
wolfsprauljust checking on the PCB, yes the right one says "SW1", the middle one is "SW2", the left one is "SW3"08:16
wolfspraula simple 1 2 3 would do it for me, maybe in a circle? or you could write the 1 2 3 onto the actual button itself?08:18
wolfspraulroh: at gemmel, I paid about 30 EUR I think08:21
wolfspraulcutting (149x124) was included08:21
rohok. we'll see what they answer.08:23
wolfspraulthat was including VAT08:27
wolfspraullekernel: I think we need some idea for labeling the 3 buttons08:28
wolfspraulnot labeling them at all is not an option, it will create big confusion later08:29
wolfspraulI can tell you 80% of people will think render/on/standby actually is for the buttons, the left one is 'render', the middle one 'on', and the right one 'standby'08:29
xiangfuthat is what I'm think at begin :-)08:38
wolfspraulxiangfu: definitely. and the only reason we don't hear feedback is because people will then press the middle button which they believe is 'ON', and m1 will turn on! :-)08:41
wolfsprauland most people will not go beyond that point in terms of the buttons or leds, so we hear no complaints about the button labels.08:41
wolfspraulroh: another feedback - it seems sometimes after assembly and tightening the screws, one finds out that a button is continuously pressed down09:42
wolfspraulso you cannot boot the m1 until you have identified that 'stuck' button09:42
rohhm. that shouldnt happen at all.09:44
rohatleast it worked for me when i didnt glue where it shouldnt be glued (so yes, i encountered it ;)09:44
rohmaybe it gets better with the more precise cut i hope to get this time09:44
wolfspraulif you can, please try to glue the buttons on your side09:45
rohthats mostly a question of time and labour.09:45
wolfspraulgluing the buttons requires some precision, so it's best done by whoever is closest to the actual production09:46
rohwe are in midst of moving atm, so the next 14 days will be fun anyhow09:46
wolfspraulno rush09:46
wolfspraulas for buttons being stuck, I have encountered that several times09:46
wolfspraulbut of course you are right - the buttons were in various states of 'glue quality'09:46
rohalso i havent heard back from the laserguy, so maybe it could take 2-3 weeks to get everything rolling and done09:46
wolfspraulwe should use a glue film or so, to get a consistent glue thickness09:47
rohif you find a really thin, good doublesided film, please holler09:47
rohi tried some and they all were quite thick (0.2-0.3mm) and made the buttons wobbly to the side and not gave a proper button feel09:47
xiangfuthe rescue mode will first try to load image from tftp.09:50
wolfspraulxiangfu: reflashed your latest set. when I click on Start!, my projector goes black09:53
xiangfuhow about switch resolution in "system setting"?09:54
wolfspraulalready switched to 640x480, same09:54
wpwrakwolfspraul, roh: (buttons) if they're sw-defined and have "strange" numbering, maybe label them A, B, C or X, Y, Z ? that way, the letters can be in the usual left-to-right order. the mapping would be A-3, B-2, C-1.09:54
wolfspraulwhen I'm in another resolution, I have a lot of flickering before the projector goes black09:55
wolfspraulah, tried another patch now it works09:55
xiangfuyes. there are some patch have bug. in new flickernoise.09:56
Action: xiangfu back online later09:58
wpwrakbtw, the "patch" term. so "patch" is synonymous to "presets" ? is "presets" used in the sense of "parameters to a given process/procedure" or does it have a special meaning in mm1 terminology as well ?10:07
wolfspraulxiangfu (in absentia) - since rendering switches back to 640x480, and all dialogs in the GUI will be designed to work well in 640x480 (for now?), I don't understand why the default GUI resolution should be higher than 640x48010:10
wolfspraulI think it should be 640x480 by default, the others can be in the settings of course10:10
rohwpwrak: patch means 'flickrnoise program'10:10
wolfspraulonly if we start to design the dialogs for something higher, say 800x600, I would change the default to that as well10:10
rohit comes from that winamp-plugin foo.. and yes, the term 'patch' for that is as well stupid as confusing10:11
wolfspraulbut then we need to clearly say whether 640x480 is still supported, or works 'somewhat', or what10:11
rohatleast misleading10:11
wolfspraulI've had a discussion about that 'language' with Sebastien recently, and the conclusion was that it's called 'Flickernoise Patches'10:11
rohbut in the end it wasnt chosen within th scope of the project10:12
wolfspraulroh: according to Sebastien, it was in fact especially chosen for the scope of the project10:12
wolfspraulI don't care so much as to which word we use, as long as we all call it the same way10:12
wpwrakroh: (patch) confusing indeed. can you have multiple patches on the same screen ?10:12
rohwell.. _everybody_ including musicians i know does see patch as either a cable or some 'diff' form textblock10:12
rohand not 'a file with program code in it'10:12
wolfspraullet's hope Sebastien is back, there is something in the backlog about it somewhere10:13
rohwpwrak: i dont think so (multiple patches)10:13
wolfspraulwe all need to be on the same page10:13
wolfspraulit's Flickernoise Patches for now...10:13
rohwpwrak: check out some patch.. its like code for a weird vm which basically does math10:13
wpwrakwolfspraul: unusual "official" terminology means that people will often add a "non-official" translation. e.g., think of kicad "components" or "modules". those terms are only understood in a very narrow context.10:14
wolfspraulthat's fine but I will still go with the 'official' one10:14
wpwrakroh: (not multiple) so they're not even like real-life patches ;-)10:14
lekernelsh4rm4: yeah opensparc is cool. but enormous and not appropriate for FPGAs. when you have an FPGA, you'd rather keep the CPU small instead of having all the bells and whistles of OpenSPARC and take advantage of the reconfigurability to implement custom accelerators10:14
wolfspraulor rather I need to know what 'official' is. If we don't even know, maybe there is nothing official :-)10:14
wpwraki wonder if common parlance "effect" (as in "visual effect") could be synonymous with Flickernoise "patch"10:15
lekernelimo opensparc wouldn't even fit in the MM1 FPGA (perhaps S1 Core would) or be slower, while our SoC only uses ~45% of the LUTs...10:16
lekernelyeah everything is on github.com/milkymist now10:16
lekernelkristianpaul: what resolution were you operating the GUI at? 1024x768?10:18
rohwpwrak: i'd simply call them 'programs'10:19
rohor scripts10:20
rohlike in a script for a play10:20
wpwrakroh: "program" sounds more like /bin/ls and such. similar with scripts.10:20
lekernelwolfspraul: label 1/2/3 yeah10:21
lekernelmaybe add "on" on the middle button10:21
wpwrakbut yes, "scripts" seems to be closer, because it also includes the procedural element10:21
rohlekernel: the buttons can rotate.. so on the button is a bad idea.10:21
rohlets add triangle, sqare and circle and get sued by sony... greeat marketing!10:22
wpwrakroh: use symbols, [], *, and O ? :)10:22
wpwrakheh ;-)10:23
lekernelwolfspraul: (640x480 by default) then kristianpaul's screen wouldn't work at all...10:23
rohbtw.. does the vga port do ddc?10:23
rohsome beamers need rather.. special timings... like in... special people10:24
lekernelyeah it does ddc, but software doesn't10:24
rohah. nice. so the hw-part is tested?10:24
lekernelthere's a function to read the EDID which works, but nothing beyond that10:24
rohit was just 'read some page via i2c' right?10:24
lekernelthe factory test program attempts to read EDID and check its header10:24
rohwow. gemmel wants .55 euro per sheet.10:27
rohand thats not drilled and doesnt have a sticker10:27
lekernelaw_: wolfspraul: roh: do you have github accounts?10:33
wolfspraullekernel: [screen res] oh, didn't know about kpaul's screen10:41
wolfspraullekernel: [screen res] so what do we do then?10:41
wolfspraulwe must absolutely focus on one default setup, and optimize for that10:41
wolfspraulif the default res is 800x600 or 1024x768, we have two problems right now:10:41
wolfspraul1) the dialogs are not optimized for such a high resolution, especially at 1024x768 everything looks too small10:42
wolfspraulyou would even want to use the zoom feature of some projectors, reversing the whole thing... :-)10:42
wolfspraul2) if we increase the dialogs, or settle on a higher 'default' resolution, we may as well remove the smaller ones because dialogs that don't fit on screen are really bad10:42
wolfspraul3) rendering will go back to 640x480, with a lot of flickering it seems10:42
wolfspraul3 problems10:43
wolfspraulI suggest 640x480 to be the default GUI res, and there can be a keypress to iterate over resolutions10:43
wolfspraulif kpaul's projector/screen doesn't work at 640x480, how about rendering?10:43
wolfspraulroh: so we label then 1 2 3?10:45
wolfspraulwith circle?10:45
wolfspraul[buttons can rotate] - good point, so not on the button for sure10:45
wolfspraulseems lekernel left too early for the 'Milkymist Patches' thing10:48
rohwolfspraul: if i label 1 2 3 its completely different to the pcb and schematics labels 3 2 110:53
rohand nope.. i dont have github accounts10:55
wolfspraulroh: we can change the pcb for rc310:55
wolfspraulI doubt anywhere in the code we have a reference to 'sw1', 'sw2', 'sw3'10:55
roh(after theit last fuckup i was laughing to hard)10:56
wolfspraullekernel: are you OK with swapping the silkscreen that says 'sw1' and 'sw3' on rc3?10:56
lekernelwolfspraul: yeah10:56
lekernelroh: could you create one?10:56
wolfspraulroh: ok, settled. we label them 1 2 3 from the left, and the silkscreen in rc3 will be adjusted as well10:56
rohlekernel: i will think about it. to be fair.. i am not really interrested in even more spam10:57
lekernelbeen on github since august 2009, never received a spam from them10:57
rohi get most my spam via accounts on google code, sourceforge and similar stuff.10:57
rohlekernel: well.. not from them directly.. but from all the mindless headhunters scraping all these platforms10:57
lekernelthis happens everytime you have a visible project10:58
lekernelwolfspraul: add an "ON" label too, no?10:59
wolfspraulwhat are the other two used for?10:59
rohlekernel: havent had it before the big github and similar platforms explosion. in the end: i dont believe in 'too big' userbase infrastructure anymore at all. just makes the fail bigger, not the problems less or smaller ;)10:59
wolfspraulthe middle one is 'ON'10:59
wolfspraulI agree with roh about github, but it's definitely good to have more milkymist stuff in one place (even though I think that should be milkymist.org or so)11:00
wolfspraulfor qi-hardware, I want to keep copyleft hardware (manufacturing) related stuff together11:00
wolfspraulbut other than that I like the new milkymist group on github11:00
wolfspraullekernel: how about the default GUI resolution? not sure you saw my points above11:01
wolfspraulI think this is important. we should have one default, and things should look good on that default.11:01
lekernelwolfspraul: currently the other buttons are used to a) boot in rescue mode b) boot in SDRAM calibration mode11:01
rohwolfspraul: well.. its not that bad as other hosted s*it ... atleast one has a proper backup ;)-11:01
wolfspraulif kpaul's screen doesn't work at 640x480, how about rendering for him? I don't get it...11:01
lekernelwolfspraul: next SoC versions will support rendering in any resolution11:02
lekernelbut right now, it switches back to 640x480 for rendering a patch11:02
wolfspraulbut right now, I think the default GUI res should be 640x48011:02
wolfspraulI don't see any reason, to be honest, for a higher one except for showing that we can :-)11:03
wolfspraulthe dialogs are not designed for higher res, rendering will switch back (including flickering)11:03
lekernelfirst showing that we can is good since everyone asks about that :)11:03
lekerneland imo the dialogs look good in 102411:03
wolfspraula bit small, no? :-)11:04
lekerneland yeah there's this little flickering problem... but will be fixed at the same time as support for arbitrary resolution rendering is added11:04
wolfsprauldoes this mean we should increase their size and make 640x480 unusable? or should we support multiple 'zoom states', font sizes, etc?11:04
wolfspraulare you aware of any projectors/screen that cannot do 1024x768 ?11:05
lekernelyeah. and we'll want to support PAL resolution (TV-Out) as well.11:05
wolfspraultv-out ?11:05
wolfspraulover the vga connector?11:05
wolfspraulalright then, let's not waste time. you say the default GUI res should be 1024x768 right now?11:06
lekernelI don't care11:06
wolfsprauldialogs look good, we will not increase them to not fit on 640x480 anymore, and rendering will switch down to 640x48011:06
lekernelif it's 640, showing off 1024 is only two button clicks away11:06
wolfspraulyes. and if we add more features, we can move everything up.11:07
wolfspraulI just try to make it a really good out of the box experience.11:07
wolfspraulsmall dialogs are not, flickering is not11:07
wolfspraulso with the current feature set (as of today, literally), I would keep the default at 640x48011:07
wolfspraulbut I'm flexible11:07
wolfsprauljust trying to understand11:07
wolfspraulfor the 3 buttons, well sure we can label the middle one 'ON'11:08
lekernelok, let's use 640 then11:08
wolfspraulbut the left one 'RESCUE', that sounds scary11:08
wolfspraulnot to mention 'SDRAM calibration' on the right one :-)11:08
wolfspraulso maybe '1', 'ON', '3' ?11:08
wolfspraulalso strange11:08
wolfspraul'1', 'ON', '2' ? confusing11:08
wolfspraulyou know the product best, you tell me11:08
wolfspraulLEFT, ON, RIGHT11:09
wolfspraulL, ON, R11:09
wolfspraulhow about that?11:09
CIA-48flickernoise: Sebastien Bourdeauducq master * r533c99c / src/sysconfig.c : Make default resolution 640x480 - http://bit.ly/forcli11:09
lekernelL/ON/R looks good11:09
wolfspraulroh: L, ON, R it is11:10
lekernelbtw in the GUI I'm thinking about using them as random patch switch (for demos) / power off (long press) / reboot (long press)11:10
wolfspraulmaybe on the silkscreen we make it SW_L, SW_ON, SW_R then11:11
rohwolfspraul: i'll wait for the feedback from the laser-guy first11:12
wolfspraulroh: you mean about L/ON/R ?11:14
wpwraklekernel: or free R by having  ON long -> reset, ON even longer -> power off11:14
wolfspraulhow about this 'Patches' thing?11:15
wolfspraullekernel: you saw above that roh and wpwrak think that's not a good term. To cut it short - until I hear from you about a new one, I will continue to call it Milkymist Patches...11:15
wolfspraulother proposals above were presets, programs, more?11:16
wpwrakwolfspraul: i think "presets" is also misleading, because it suggests something static11:17
wpwrakwolfspraul: (more) "effects", "scripts"11:17
lekernelpatch is good. writing equations is just like connecting stuff together, just in a different representation.11:17
lekernelscript/program is for nerds11:17
wpwrak"effect" ?11:18
wpwraklekernel: (writing equations) you mean verilog ? or something labview-ish ?11:19
lekernelverilog? oh, no :)11:20
lekernelthe milkdrop-like language... I don't know labview but I'd guess yes11:21
rohwolfspraul: about how he needs the files, pricing etc11:21
wpwrakspeaking of labview, they use "program" and "dataflow". "flow" ?11:21
wolfspraulroh: don't understand. how does that relate to L/ON/R ?11:21
wolfspraulL/ON/R is settled, right?11:21
rohwolfspraul: that it would need changes i need to test.. and i can only test a thing at a time.. no worries.. i'll get to it soon enough11:22
wolfspraulAdam will adjust the silkscreen on rc3.11:22
rohnow i just need to decide if i order shielding sheets for just this batch or for more11:22
wolfspraulroh: oh sure. but let's say L/ON/R is in the queue then :-)11:22
wolfspraulit cannot be that hard to add it, unless I completely misunderstand the process, since there is quite a bit of text there as well11:23
wolfsprauldon't order for the next batch now, that's too risky.11:23
wolfspraulI want to sell those 80 quick, of course. but a lot will depend on the launch and whether we can find strong marketing/retail/sales partners.11:23
rohyou mean we have a real chance to drop the shield soon?11:23
wolfspraulI didn't think about it that way.11:24
rohi got an offer for 200 sheets for 110E now.11:24
rohand still got 50 left. (mixed thick and thin)11:24
wolfspraulwhat's wrong with gemmel?11:24
wolfsprauljust use those 50, mixed no problem11:24
rohnothing.. thats their offer11:24
wolfspraulget another sheet from gemmel11:24
wolfspraulshould be about 30 EUR11:24
rohthose 30 would be near what 'some thinner' and in total around 70-80 sheets would cost11:25
rohjust abbreviating.11:25
wolfspraulI think the 0.3 almg3 is better - stiffer11:25
wolfspraulI suggest you use up the 50 you have, plus another 0.3 almg3 sheet from gemmel11:26
rohok.. we'll see how much more they want per sheet if i take much less (like only 50 instead of 200)11:26
rohthese 'classic' companies are weird.. they asked me to fax something *g*11:27
wolfspraulthey don't work in how many pieces you want, they work in sq/m11:28
wolfspraultheir roll has a certain width11:28
rohi see.11:28
wolfspraulforgot how much, maybe 1m, or 1.4m or so11:28
wolfspraulthen you cut off some amount, 50cm, 1m11:28
wpwraklekernel: (language) looks like a script ;-) but that's the representation, not the function. i guess the function would still be something like "effect" i common terms. (i wonder what they called those little demos back in the amiga days. just "demos" ?)11:29
wolfsprauljust say 'the smallest they sell' (that's what I bought)11:29
wolfspraulthen they should cut that for you into 149x124mm pieces11:29
wolfspraulthrow away the leftover11:29
wolfspraulif you go there, they will cut it right there (30 minutes or so). of course they have to have the roll in stock, which last time was the case for 0.2mm al 99.5 and 0.3mm almg311:30
wolfspraullast time I got about 40 pieces or so out of each of the two 'smallest size' sheets I bought11:31
wolfspraulthat would fit perfectly with that is missing for you11:31
wolfspraulask them how much 1 meter of the almg3 roll is, I believe it was 1.4m wide11:33
wolfspraulI think back then I got half a meter (times 1.4m width)11:33
wolfspraulwould that fit? let's see... 50*140=7000 cm211:34
wolfspraul15*12.5=187.5 - 7000/187.5=3711:34
wolfspraulyes, looks right11:34
wolfspraulmaybe when you order online, they will not offer 0.5m of the roll, maybe the minimum is 1m? they seemed to be a little more flexible with on-site customers11:35
wpwrak(amiga) yeah, "demo" or "intro". http://en.wikipedia.org/wiki/Amiga_demos11:35
wolfspraulroh: ask them for a quote of 0.5m of the 1.4m wide 0.3mm thick almg3 roll, cut into 149x124mm pieces. should be around 30 EUR, including VAT.11:36
rohwolfspraul: i ask directly for pieces. i dont need to care how big their roll is11:43
rohgoing there would be much more expensive than shipping11:44
wolfspraulroh: but that difference may explain the higher prices you get :-)11:44
wolfspraultheir computer system works in these rolls, and how many meters you want from the roll11:44
wolfspraulthe cutting is the second step11:44
wolfspraulthat's what I tried to explain11:44
wolfspraulso of course you can ask for '40' pieces11:45
rohwolfspraul: there is no difference. you bought something like 40sheets of each thickness.11:45
wolfspraulbut if a 0.5m slice of the 1.4m roll gives you 37 pieces, they may just charge you the next total increment - I'm not sure.11:45
rohwolfspraul: i asked for a price for 200 of the thick ones (heavier than a mix of thick and thin ones)11:45
wolfspraulanyway I gave you the background data, you can do what is best for you.11:45
wolfspraulwhat I bought was very cheap. maybe it was even 30 EUR for all 75 sheets.11:46
rohi now ordered 25 of the thick ones.11:46
wolfspraulincluding cutting, and including VAT11:46
roheh 50.11:46
wolfspraulmaybe they charge you a full meter of the roll then11:46
rohi dont think so. that would be unserious.11:46
wolfspraulthey are not charging for the cutting, interestingly. at least not when you are there.11:47
rohthey do. once per order11:47
wolfspraulwell those are all very small amounts, so the most important thing is efficiency. they are setup for some things, and not for others.11:47
wolfspraulwhen I was there, she gave me a price. then we added the cutting later (we just had that idea later), and the price stayed the same.11:47
rohwolfspraul: there are 15euro 'nebenkosten'11:56
wolfspraulI think the Performance/Done./Close dialog can be removed (in flickernoise)13:37
wolfspraulsay if I click on the mouse, why is that dialog even popping up - no need imho13:38
Fallenouis it possible to go back from performance mode to flickernoise GUI mode ? (without power cycling, or rebooting the software) ?13:39
wolfspraulsure, just press the mouse button, for example13:39
wolfspraulkeyboard probably too?13:39
lekernellol: http://2.bp.blogspot.com/_bDJ_hte7vS8/S_UIR6XXsaI/AAAAAAAAACo/cMd-zL4O5Zw/s1600/FTD0040-DB9-USB-RS232_Module_s.jpg13:40
lekernel(by ftdi)13:40
Action: kristianpaul never take care about close the Performance/Done dialog13:40
lekernelwolfspraul: when you have several patches in your performance set, this dialog shows a progress bar13:40
lekernelalso reports error13:40
wolfspraulkristianpaul: that's why I would just remove it entirely.13:40
wolfspraulalright then :-)13:40
kristianpaulhmm error is important13:40
kristianpaulbut is part of the same code for this dialog?13:41
lekernelif you have only one patch, yeah, it's useless13:41
kristianpaulah ..13:41
lekernelbut try making a set with ~40 patches or so13:41
Action: kristianpaul never loaded multiples patches at once13:41
xiangfulekernel: thanks for import the m1s.git to 'milkymist/scripts.git'14:10
xiangfulekernel: I have two small patch for flickernoise:14:11
xiangfuhttp://pastebin.com/DTuF5NPs : release '0' with 'SC_RESOLUTION_640_480' in fb.c14:11
xiangfuhttp://pastebin.com/tpJFXYay : add Ctrl+F1 for iterate switch screen resolution14:12
xiangfulekernel: if you agree. I will direct commit to flickernoise.git :)  (since I have write access now)14:12
lekernelok, commit them14:12
lekernelhm wait14:12
lekernelin switch_resolution() please use sysconfig_get_resolution() instead of res = SC_RESOLUTION_640_48014:13
xiangfulekernel: now. I am working on increase the Video-in preview . I will try to add two switch button. 144x180 and 288x360.14:13
lekernelbtw the current RTEMS API to switch resolutions is quite of a hack, so no need to spend time on fixes like patch 114:14
lekernelthe current API will disappear when we have the new SoC which supports scaling (and rendering in any resolution)14:14
xiangfulekernel: ok. then no patch 1. just patch 2.14:15
lekernelyou can commit patch 1...14:15
lekernelnow that you wrote it14:15
xiangfupatch 2: it's 'static' can not get value from function14:15
lekernelbut no need to review the code for consistency there14:15
lekernelhm? remove static then14:15
xiangfupatch 1: ok , understand now.14:16
lekernelres = get_resolution();14:16
lekernelif(res >  SC_RESOLUTION_1024_768) res = SC_RESOLUTION_640_480;14:16
xiangfuoh. yes. I forget the 'set_resolution',14:17
lekernelalso, what happens when you press ctrl f1 while rendering?14:18
xiangfutesting now.14:21
lekernelit probably breaks imo :)14:21
lekernelbtw one small improvement for your screen grabber14:24
lekernelcan you remove the unneeded alpha channel? it would cut down a bit the slowness and the file size14:24
xiangfulekernel: yes. I just delay this task. will fix that today or tomorrow14:26
xiangfuwhen rendering. 800x600 == half screen rendering. 1024x768 == black screen14:26
lekernelso please disable ctrl f1 when rendering14:27
xiangfubut we can always go back to 640x480 with out any problem. flickernoise not freeze.14:27
xiangfulekernel: ok14:27
xiangfulekernel: new patch 2: http://pastebin.com/AfGGJk5314:51
lekernelxiangfu: ok, good14:55
CIA-48flickernoise: Xiangfu Liu master * r5e03616 / src/fb.c : fb.c: replease 0 with SC_RESOLUTION_640_480 - http://bit.ly/ghMfwU15:00
CIA-48flickernoise: Xiangfu Liu master * rd5d459d / (src/fb.c src/fb.h src/shortcuts.c): add Ctrl+F1 for iterate switch screen resolution - http://bit.ly/e6SNP915:00
xiangfulekernel: there is no 'fjmem' in github/milkymist .15:14
lekernelxiangfu: one question: in future, all git repos under 'https://github.com/milkymist' will became milkymist upstream. right?  <= didn't get your question15:32
lekernelgit://github.com/psycho-nico/milkymist-openwrt.git is dead, won't move it15:32
lekernelfjmem... depends whether mwalle wants to use github and/or his own servers15:33
xiangfuI just found the fallen/rtems-milkymist.git is gone. I mean in future the git repos under github.com/milkymist/*  will be the official source code for m1 right?15:34
lekerneli'll mirror fjmem anyway15:34
lekernelfor qemu/urjtag, all/most of it is upstream, no need to create new repositories for this imo15:35
xiangfumaybe. fjmem-m1. other repo are all *-m115:38
lekernelyup, renamed15:44
lekernelthanks for your attention to detail :-)15:45
xiangfulekernel: btw: http://downloads.qi-hardware.com/software/images/Milkymist_One/latest/15:46
xiangfuI build those files for reflash by using 'urjtag'. it's not ready for end users.15:47
xiangfuthe plan is make the build automatic a little.  get all build log.15:47
lekernelyeah, and let's only flash boards that come out with releases15:47
lekernelnot repository snapshots15:47
lekerneli'll try to get another release with the new ethernet core some time before we start flashing the next batch15:48
lekernelso we have time to test and maybe fix last minute bad bugs :)15:48
xiangfuafter I flash the new image. for now the ethernet works fine. but the m1 get freeze sometimes. I will try to correct more info next time it freeze.15:50
xiangfualso some patches not working with new image15:51
xiangfulekernel: do you think I send one email to list about those image is ok? for someone when want test new image?15:52
lekernelwhat do you can 'new image'?15:52
lekerneland please give me the list of the patches that do not work15:53
xiangfuimage under: http://downloads.qi-hardware.com/software/images/Milkymist_One/latest/15:53
lekerneland yeah you can advertise your link on the mailing list. but make clear these are development snapshots.15:53
xiangfulekernel: patches. I will find out them tomorrow. (23:50 now. will goto sleep soon)15:54
lekernelis that with the new ethernet core (minimac2)?15:54
xiangfulekernel: I update all git repo to last. and those images are build today.15:54
lekernelhm, ok15:55
xiangfuyes. it's minimac2.15:55
xiangfu(from the git commit log :)15:55
xiangfutime for sleep. see you16:05
CIA-48flickernoise: Sebastien Bourdeauducq master * r65e17b5 / src/shortcuts.c : Use Ctrl-F2 for screenshot - http://bit.ly/i5Yq9B21:00
terpstra_mwalle, finally got a fully working gdb over JTAG using a 1K debug ROM.21:13
terpstra_works quite nicely :)21:14
terpstra_quite acceptable speed as well21:14
terpstra_i can load firmware via gdb's "load" command at about 3kB/s21:14
terpstra_which, while not super-fast, isn't bad for a bitbanging usb cable21:15
mwalleterpstra_: nice, so you are using the mini rom monitor?21:18
terpstra_i wrote my own21:19
terpstra_the original wasn't relocatable21:19
terpstra_and i wanted to fit it into 1KB21:19
mwallewith ram?21:19
terpstra_1k ROM + 256 bytes RAM21:19
mwalleok nice :)21:20
mwalledoes it support byte/halfword/word access?21:20
mwalle(used eg for flash initialization)21:20
terpstra_no, it does not21:20
terpstra_i might add that later21:21
terpstra_but at the moment i have other goals :)21:21
terpstra_i also don't deal with bus faults21:23
terpstra_which i consider higher priority than flash (which i don't have)21:23
mwallemh the qemu model of milkymists sysctl timers are totally broken :(21:25
terpstra_i got to run now; was a long hackathon21:26
terpstra_bed time!21:26
mwallelekernel: is there a way to know wether the mac address is valid or not (is it included in the rescue bios crc?)22:19
lekernelyes, it's in the bios crc22:20
mwallewe should set a valid default mac address if the crc failed, eg 02:00:00:00:00:0122:22
--- Tue Apr 12 201100:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!