#qi-hardware IRC log for Monday, 2010-12-27

qi-bot[commit] Xiangfu Liu: gmenu2x files: add emacs icon http://qi-hw.com/p/openwrt-packages/c0002c501:47
kristianpaulwolfspraul: do you know what the Jz4750L Ingenic board is, and what i have an FPGA ?05:05
kristianpauli just discover it reading some lines from u-boot changelog05:06
wolfspraulkristianpaul: don't understand your question05:11
wolfspraulI think the 4750L is the low-cost (=cob) version of the 475005:12
kristianpaulwolfspraul: i found this line in the U-boot source code05:13
kristianpaul2009.05.1405:13
kristianpaul* Add Jz4750L platform: F4750L board(which is a FPGA board).05:13
kristianpauland the FPGA word call my atention, no more :-)05:14
wolfspraulhmm05:16
wolfspraulI can only guess.05:16
wolfspraulso the 4750 is a 'normal' chip, very much like 4720/40 just with some improvements05:16
wolfspraul'l' is the low-cost version05:16
wolfspraulnow, inside the Ingenic office I regularly see _a lot_ of different boards05:17
wolfspraulthey are building all kinds of small-volume prototype boards for development, for customers, with customer, by customers, etc.05:17
wolfsprauland I have seen fpgas in combination with Ingenic chips on some of those boards05:17
wolfspraulvery much like SIE05:17
kristianpaulI see05:18
wolfspraulprobably because Ingenic, and/or their customers, are prototyping or developing something05:18
wolfspraulthe support for all these different boards in the u-boot, or Linux kernel, or other kernel sources is messy05:18
wolfspraulbecause a lot of these boards are only made in a few units, and are superseded by something else quickly, and/or the customer or Ingenic aborts the development, etc.05:18
wolfspraulso maybe that line is connected to such a board05:19
wolfspraulalso notice that it says "F4750L"05:19
wolfspraulF = fpga?05:19
kristianpauldont know05:19
wolfspraulbut even then, you will most likely not find information about this board anywhere05:19
wolfspraulthey make 10, for whatever reason, learn something from it or not, and move on05:20
kristianpaulyes i realized that before asking you :-)05:20
kristianpaulThey move fast :-)05:20
wolfsprauloh sure05:20
wolfspraullots of projects always going on05:20
wolfsprauleverything fast. quickly started, quickly abandoned.05:20
zrafawolfspraul: all: some guy was talking me several times because he uses jlime and he was very interested on the wikireader thing and other features... Few days ago he was explaining me that he, and his students (uni) are trying to write a nice GUI, very easy to use (imagine like just one search bar like goolge, or few buttons), to convert the nanonote as a device similar to this : http://www.literacybridge.org/talking-book/ .. But, because nn is a lot more05:22
zrafawolfspraul: He already has around 30 nns!.. which is a really nice05:23
zrafaif he is really interested on to do that thing.05:23
zrafaI think that his idea is really nice as final target for a device like nn,05:23
zrafaso I am trying to tell him to join community (like this channel :) ) in order to do the whole idea working05:24
kristianpaulwow 30 sounds like a pilot coming?..05:24
kristianpauls/pilot/small deployment05:24
wolfspraulzrafa: wow he has 30 nanonotes?05:26
wolfspraulsure it sounds good05:27
wolfspraulI think we will see lots of great apps next year...05:27
Action: kristianpaul :D05:27
wolfspraulI am currently hacking up a little script that will let me download ogg music files from wikimedia commons :-)05:28
kristianpaulwow05:28
zrafawolfspraul: yes, he told me that he and his students has around 30 nns. And maybe they would need more05:28
zrafawolfspraul: downloader script : nice :) if you do I would add the GUI :)05:29
wolfspraulwell they only have classical music there, but it's a start. I also saw the 'spoken wikipedia' project, which made me realize to look for audio books...05:29
kristianpaulwolfspraul:  audio books > a lot of then in the public domain at librivox.org05:30
zrafawolfspraul: does your script would list on stdout the list of music files available?05:32
kristianpaulhad you tught in Gutenberg proyect, i think their format is better to port on nanonote than wiki-* also wikireader startert to incorporate? (maybe people get tired of reading just wikis..)05:32
kristianpauls/tught/tought05:32
kristianpaul"The software on the Talking Book is the heart of the device. Because it is a computer, we can work with our partners to customize the software to precisely fill their needs. "05:34
kristianpaulwell i dint expected something better..05:34
kristianpaulzrafa: i guess they are using your wikipedia-gnudict plugin?05:36
zrafakristianpaul: about talking book: just a note: he is not wanting to use that software if I understood him corretly. It is just a link to show the idea05:43
kristianpaularggg i need a good ccc low bandwitch streaming link..05:43
kristianpaulzrafa: (link) sure it points the idea05:44
zrafakristianpaul: I have no idea exactly which thing they use, but from what he told me he wants to understand how we did the wikireader, no to use it. And he is not going to do a wikireader either, they want to do a new thing easy to use05:44
kristianpaulsure05:45
zrafaso they would not use the current stuff. They want to understand how to play music, video; how to show graphics, pictures, how to do some engine of knowledge, things like that05:45
Jay7wikireader + festival :)05:56
wolfspraulzrafa: yes sure, browse categories and list files. 10 lines php :-)05:59
wolfspraulit's not finished yet, need to find a few hours so that it can download as well. I just want an easy way to download music to my nano.05:59
wolfspraulalso for it to work on the NanoNote, the current script uses the HTTP_Request class which I doubt it installed right now.06:00
wolfspraulof course there are other ways like invoking wget, but I like to stay in php06:00
kristianpaulwpwrak: i finally discover the cpu is reading tos fast the ram-like-fpga-setup, so i'm implementing a delay in software.. to avoid repeated data07:57
kristianpauls/tos/too07:59
kristianpaulhmm i dint work as i spected..08:04
kristianpaulI need an IRQ (glup!)08:05
kristianpaulI found a "IRQ" signal, but it is also LCD_D15 pin.. :/09:36
wpwrakkristianpaul: no free IRQs on SIE ? :-(10:13
wpwrakwolfspraul: (kicad patches) hmm, eeschema-bom-only-mode.patch has a change to eeschema/CMakeLists.txt where the context contains cmp_library_keywords.cpp. my bzr revision 2358 doesn't have this.10:23
wolfspraulwpwrak: he one we 'standardized' upon was 2448?10:30
wolfspraulthat was the one you told me you are using...10:30
wpwrakhmm, maybe i got lost myself :) lemme check ...10:31
wolfspraulupticking kicad is another thing, I will not make the first step there, no need right now (for me)10:31
wolfspraulbut if you uptick it, feel free to just comment out my -scripted patches, and I am standby to uptick them as well, right away10:32
wpwrakright, 2448. funny. all my notes say 2358. never trust documentation ;-)10:32
wolfspraulwell but why do you get the context problem?10:32
wolfspraulyou are actually using 2358? but all this time you thought it was 2448?10:33
kristianpaulwpwrak: i just sent a mail to qi (cc: carlos camargo) exmplaining my situation10:33
wpwrakno, i'm using 2448 but the places where i have written down the number (e.g., docs/GETTING-STARTED in gta02-core) say it's 235810:33
wolfspraulbut why do you get the context problem then?10:34
kristianpaulwpwrak: (IRQ) well not free, but i still asking my self why the pin in sharef with LCD, or is just a label..10:34
wolfspraulmy 2 patches apply cleanly against 2448, in the quilt series as uploaded10:34
wpwrak(2448) and i lost that kicad source tree with the disk loss a few weeks ago. so i restarted from documentation ... and thus ended up with 2358 :)10:35
kristianpaulwpwrak: (well not free) > not sure if free10:35
wolfspraulas for KiCad upticking, that's a task I can definitely offer to work on, if you want to get rid of it and suspect that some improvements might be worth the work.10:35
wolfspraulI am slowly getting more and more familiar with KiCad, so I can already 'test' a newer version to a certain degree.10:35
wolfspraulit's up to you...10:35
wolfspraulof course I may uptick it to xxxx, and then you find some major problem with xxxx that I just wasn't aware of10:35
wpwrakneed to check what the latest version offers. there are a few bugs in 2448 that i wouldn't mind seeing gone.10:35
wolfspraullike I said, I offer to work on this, if that helps you10:36
wolfsprauleither you explain to me what the 2448 bugs are, so I can test whether they are gone, or you try yourself. you need to decide what costs you less time...10:37
wolfspraulif we move kicad-patches to eda-tools, I can commit right there and it should be quite efficient for me10:37
wpwrakthanks ! it's mainly a question of running the new stuff on a design you know well and see if anything appears unusual10:37
wolfspraulsure10:37
wolfspraulthat's why I think I am only 20% ready for the job now, slowly increasing10:37
wolfspraulmaybe by the time we manufacture Xue I'm at 75%? :-)10:37
wpwrak(2442 bugs) they're things that happen occasionally. like kicad running into an internal error if you resize areas. or "create a corner" only being available sometimes.10:38
wolfspraul244810:38
wpwrakerr yes, 244810:38
wpwrak2448 patched cleanly. building ...10:39
kristianpaulMauroRojas: hola10:43
MauroRojasHola!10:43
wpwrakMauroRojas: welcome back among the living ! ;-)10:43
MauroRojasthanks :-)10:43
wpwraknice ... pcbnew --help  ... now it looks like a real program :)10:45
wpwrakwolfspraul: having -print-layers and --print-layers is a bit confusing. if the short option is the same at the long option, there probably shouldn't be a short option at all10:47
wolfspraulyes, eeschema doesn't have that yet, but when I go back to eeschema I'll clean it all up there, and merge your --plot patch as well.10:47
wolfspraulyes but the wx layer has its system10:47
wolfspraulI'm trying to follow KiCad as much as possible, and will do even more so then I start to cleanup10:47
wpwrakwolfspraul: also, the help for ps-pads-drill is confusing. --ps-pads-drill-ops is clear, but what does -ps-pads-drill do ? (supposedly the same as --ps-pads-drill-opts)10:48
wolfspraulhttp://docs.wxwidgets.org/trunk/classwx_cmd_line_parser.html10:48
wolfspraulyes sure, same10:49
wolfspraul-ps-pads-drill is a random 'short name' for the long name10:49
wolfspraulthe wx... class has this concept of 'short' and 'long'10:49
wolfspraulI can try to see what happens if I leave the short string empty10:49
wolfspraulmaybe it will only display the long one then, that would be better, I agree10:49
wolfspraulthese are pecularities of the wxCmdLineParser class I haven't gotten around to yet10:50
wolfspraulwx is support to work well on Mac and Windows too10:50
wolfspraulso they also have this concept that the separator can be != '-'10:50
wolfspraullike / on Windows10:50
wpwrakhmm yes, i see. i think the idea for a "short name" is for it to be just a letter10:50
wolfspraulI have two options in all this. I can try to follow the KiCad classes, design, etc.10:51
wolfsprauland that will produce some 'quirks' when you look at it from a pure Linux/Unix perspective10:51
wolfspraulor I can just do it all Linux-style, but then bringing it to Mac or Windows would be a lot of work10:51
wolfspraul(and most likely never happen)10:51
wolfspraulyes that's the idea, 'just a letter'10:52
wolfspraulsomething short10:52
wolfspraulI can try to see whether the class accepts an empty/missing short name, haven't tried that yet10:52
wpwrakthat would also work on unix10:52
wpwraki think if you can just change the "short" names - that are actually quite long - to single letters, that would be sufficient10:53
wolfsprauldon't you think that's even more confusing?10:53
wolfspraulsure we can do that...10:53
wpwrakthe usefulness of a ton of cryptic one-letter abbreviations is somewhat questionable, but at least they wouldn't get in the way10:53
wolfspraulmaybe let's see how the parameters evolve first10:53
wolfspraulI suggest to completely ignore the short names right now10:53
wolfspraullet's focus on making the long ones better10:53
wolfspraulI will check whether the class allows disabling of the short ones.10:54
wpwrakyep, if there's a choice for not having short names at all, i think that would be better. we're not exactly optimizing keystrokes here anyway ;-)10:54
wpwrakwolfspraul: since we discussed consistency ... a netlist generation option in eeschema would also be nice10:59
wpwrakwolfspraul: it would also let one run the full build process from files in the repository to the fabrication data11:00
wolfspraulyes definitely, will add netlist to eeschema11:03
wolfspraultogether with erc11:03
wolfsprauland merging the current --plot and --bom patches into something cleaner11:04
wolfspraulwpwrak: did you run some of the pcbnew cmdline stuff successfully?11:04
wolfspraulwithout crash - did it produce usable files?11:04
wpwrakwolfspraul: drill file so far. worked flawlessly11:29
wolfspraulnice!11:30
wolfspraulthat makes for a happy flight to Berlin now, thanks for letting me know :-)11:30
wolfspraulso you gave your green-light already for moving the patches to eda-tools, and doing more work on the eeschema side (including merging your --plot patch)11:30
wolfspraulI will try to do this soon, to get over the whole KiCad thing, maybe the next week or so11:31
wpwrak(happy flight) ;-))11:31
wolfspraulfirst need to get 27c3 behind me, not sure how much spare time I have there11:31
wolfspraulso maybe I'll finish the kicad cmdline stuff when I'm back from there11:31
wpwrak(27c3) you'll probably sleep negative hours :)11:31
wolfspraulwhat I need from you is feedback on whether the current cmdline options work, and whether you need more, or whether you have some ideas for re-layout of the parameters11:32
wolfspraulthose kind of things11:32
wpwrak(patches and eda-tools) in fact, i have them sit there already. they're just a git add, commit, and push away :)11:32
wolfspraulbut we can do that slowly as we realize more clearly what we need...11:32
wolfspraulhe, nice11:32
wpwrak(option names) maybe change --print-layers to --list-layers, considering that we;re in the midst of plotting and printing here11:34
wolfspraulsure11:35
wpwrakalso, the help text for -l should say what that list is for :)11:35
wolfspraulI think the options will evolve over time11:35
wolfsprauldon't know :-)11:35
wolfspraulit lists the 'enabled' layers11:35
wolfspraulI think it's the same as those that are checked by default in the dialogs11:35
wpwraklet's try somthing harder now ... the postscripts11:36
wolfspraulunless you overwrite those defauls somehow (KiCad seems to save options and option overrides in several places)11:36
wolfspraulI try to make the cmdline options as deterministic as possible11:36
wolfspraulthey should not be influenced by anything but the .sch and .brd files11:36
wpwrakthe whole options systems is borderline incomprehensible :-(11:36
wpwrak(the one of kicad i mean)11:36
wolfspraulbut it is quite possible that they still are influenced by stuff saved locally in KiCad somewhere11:36
wolfspraulif so, that's a bug11:36
wolfspraulI tried to write the codepath in such a way to bypass those 'local' options and option overrides11:37
wolfspraulbut we need to see over time, I expect some surprises...11:37
wolfspraulso this --list-layers lists the 'enabled' layers, whatever that is exactly11:37
wolfspraulI hope it's a deterministic on/off switch in the .brd file.11:37
wolfspraulalso you may need those strings if you want to specify layers for the svg or plot options11:38
wpwrakmaybe it's the ones you select in  pcbnew: Design Rules > Layers Setup ?11:38
wolfspraulyes11:38
wolfsprauland those are the same strings used in the comma separated lists11:38
wolfspraulin --layers=a,b,c11:39
wpwrakaha. the area fill problem hits11:43
wpwrakinterestingly, --drc doesn't affect it. --drc also seems unusually fast.11:45
wpwrakthe report look good, though. strange.11:46
wpwrakmaybe you run just the report generator without recalculating the checks ?11:46
wpwrakrun time: 5 s for manual DRC (time during which the dialog is "running", so without pcbnew startup an such). 0.6 s for --drc and --plot11:48
wolfspraulwhich --plot are you running ?11:50
wolfspraulwhich .brd file?11:50
wpwrakruning pcbnew, filling the zones, and saving it with the zones filled solves the problem of them not getting plotted. it's a nasty trap, though. (could be something that's fixed in a more recent version, though. it's something that can bite you also without --plot)11:50
wpwrak--plot=ps_a411:50
wpwrakatusb.brd11:50
wpwrakpcbnew  --plot=ps_a4 -l Front -mirror --ps-pads-drill-opt=real atusb.brd11:50
wolfsprauland you like to mix short and long options, probably to torture me :-)11:51
wolfspraulthe short ones are really horrible...11:51
wolfspraulso this plot when run on the cmdline produces a different result from running in the plot dialog?11:52
wolfspraulwhen you say "running pcbnew, filling the zones, and saving it with the zones filled solves..." what do you mean exactly in KiCad?11:53
wolfspraulwhat are the steps...11:53
Action: wpwrak hates brown-outs12:04
kristianpaulUPS?12:04
wpwraki had one - just made the brown-outs worse12:07
wpwrakand another one :-(12:13
wolfspraulwpwrak: I found a bug wrt default line thickness, uploaded a new pcbnew-scripted.patch13:37
wolfspraulI was able to reproduce that with the atusb.brd and options you gave me13:37
wolfspraulso that's good, that's definitely an improvement13:37
wolfspraulI also renamed --print-layers to --list-layers13:37
wolfspraulupdate is here http://downloads.qi-hardware.com/people/wolfgang/kicad_patches/pcbnew-scripted.patch13:38
lekernel_27C3http://www.youtube.com/watch?v=6AesY6V-qto (that was shot in early april)14:03
kristianpaulargg you french is hard to catch ;-)14:19
kristianpauloh well seems LCD have free pins after all, just was a bit tricy as they keep labeled as LCD and the new functions too.17:48
kristianpaulandres-calderon: the only thing i see sram example is missing is a busrt mode for data transfer from FPGA to Xbusrt ;-)18:22
kristianpaulIs some kind of frustating at first then funny see a fpga as ram :-)18:27
andres-calderonkristianpaul: its just a starting point. There must be room for a lot of optimizations (timing, burst?, dma?..  I do not know ).18:28
kristianpauldma??18:28
kristianpaulreally...18:29
kristianpaultiming still tricky but even if it get optimized (cpu cylces, etc..) i think is mainally oriented to be efficient both with buffer like setup or with not hihg demanding IO transfer setup :-)18:31
kristianpaulandres-calderon: i was checking dma for jz4725 also wpwrak , but seems there is no documentation or feature avaliable for EMC18:32
kristianpaulso far i just realized DMA capabillties with the nanonote and the mmc controller18:32
andres-calderonSome really nice SoC support external DMA.18:33
kristianpaulany hint from your side is very wellcome, o just met SIE when ti was renamed to that name... surelly you had lots of tetchnical discuss with carlos camargo about this topic i guess18:33
kristianpaulandres-calderon: is that really nice SoC the Xbusrt JZ4725? :-)18:34
kristianpaulas i said any help with dma in SIE is apreciate :-)18:34
andres-calderonkristianpaul: nice price18:35
kristianpaulhmm'18:35
kristianpaul?18:35
kristianpaulah i you meant Xbust chips are cheap, i totally agree :-)18:36
wpwrak(dma) the xburst don't seem to have it :-( the only one where i found a mention in the documentation is the 4720, and there, it's an empty chapter.18:37
kristianpaulbtw are you p:anning get a M1? i already order one will be nice have more people in Colombia working around this nice and free SoC  :-D18:37
wpwrak(dma) there are also no pins marked for DMA request/ack and such. nor are there register values that suggest they could be used for selecting external dma.18:38
kristianpaul(argg tomorrow interesting talks at 27c3  are too early..)18:38
andres-calderonwpwrak: Hi Werner, What do You think about  this chip? http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=296-19652-1-ND18:41
andres-calderonwpwrak: too expensive?18:43
kristianpaulwpwrak: (dma and 4720) hopefully that should be asociated with mmc..18:43
kristianpaullooks easy to use chip :-)18:44
andres-calderonkristianpaul: very easy, i like it18:47
wpwrakandres-calderon: i think low design risk trumps low cost for now :)18:49
andres-calderonwpwrak: i agree18:50
kristianpaulwolfspraul: hi, how looked the M1 case?? (blue one)18:52
webpowerhi all18:53
kristianpaulhi wolfspraul18:53
kristianpaulhi webpower18:53
webpoweri want to know the differences between LGA 1156 and LGA 136618:53
webpowersocket18:53
wpwrakthe chip looks friendly.18:53
wolfspraulhe18:53
wolfsprauljust arrived in Berlin after an adventurous flight and landing, then Sebastien said he was too tired and now I'm at his place going to crash...18:54
wpwrakwolfspraul: (line width) excellent ! i ran into this, too. you fixed it before i could mention it :)18:54
kristianpaulok18:54
wolfspraulbut the great thing I saw during the 10 minutes I was at 27c3 was the m1 happily crunching away in the lounge, entertaining people18:54
wpwrakwolfspraul: (flight) what happened ?18:54
wolfspraulvery nice18:54
wolfspraulif it can handle that for 4 days I feel much better about the quality of the whole thing...18:54
wolfspraul[flight] delays delays delays18:55
wolfspraulnothing to complain though, lots of snow, runway had to be cleaned, we circled in the air longer than the actual flight, etc.18:55
wpwrakwolfspraul: hah, you should try domestic air travel in argentina ;-)18:55
kristianpaulsnow snow snow everywhere :-)18:55
wolfspraulanyway, arrived. m1 doing a great show in the lounge.18:55
wolfspraulwpwrak: hmm, OK. so the line width was not the only thing.18:55
wpwrak(m1) great !18:55
wolfspraulyou still have the fill zone problem?18:56
wolfspraulI cannot reproduce it, or rather you didn't give me reproduction steps I would understand.18:56
wpwrakwolfspraul: (fill problem) pcbnew: select "Add zones" mode. right-click on a zone. select "Remove Filled Areas in All Zones"18:58
wpwrak(if that item doesn't appear, try again and again)18:58
wpwraknow the filled zones should go from solid to outlines with thin lines18:58
wpwrakthen save the board. then exit pcbnew and plot the board. the postscript won't have filled zones.18:59
wolfspraulhmm. when I click on 'remove filled areas in all zones' - I don't see any visible change. I'm in atusb.brd19:01
wolfspraulor I can only see the difference in the .ps ?19:01
wpwrakmaybe it's already unfilled19:02
wpwrakif you're using exactly the version in git, without running a DRC or such, then it will be unfilled19:02
wolfspraulah yes. when I say 'fill' then it will all become green and yellow etc.19:03
wpwrakyup, that's it19:03
wolfspraulok, so when this gets plotted, in the cmdline version of the plot it doesn't come out right?19:03
wolfspraulin the cmdline version, which one does not come out right - filled or unfilled?19:03
wpwrakthe plot won't have the zones. (neither would it if you plotted manually from pcbnew)19:04
wpwrakthe problem for non-interactive plots is that this problem makes them quite unpredictable. in manual use, you at least have a chance to spot the problem. not that this is a very happy situation either - i've thrown away some stuff where i didn't notice it in time.19:05
wolfspraulwhat is the problem?19:07
wolfspraulthe plot will never have the filled zones, whether they are filled visually in pcbnew or not?19:07
wolfspraulbut you want them to be filled in the .ps plot, when they are filled visually?19:07
wolfspraulor you say the current --plot works fine, it will fill or unfill the zones depending on the settings in the file, but you cannot tell from outside which one it was?19:08
wpwrakyes, you can't tell19:08
wolfspraulin that case, what do you prefer? do you think we should introduce a way to query the fill status from the command line? Or should we automatically do the fill/unfill operation depending on cmdline setting?19:09
wpwraki don't think you'd ever want a plot with the zones not filled19:09
wolfspraulok19:09
wolfspraulin the current version, if it's filled in the saved .brd file, it will also be filled in the .ps plot?19:09
wpwraki think always filling or at least the option of commanding a fill would be highly desirable19:09
wpwrakyes19:09
wolfspraul(when invoked from the cmdline)19:09
wolfspraulah OK19:10
wolfspraulso the only problem is that you don't know from the outside19:10
wolfspraulis the fill/unfill operation always 'safe'?19:10
wpwrakso your stuff works. it's just a pcbnew usability issue that gets worse if you don't have visual feedback19:10
wolfspraulor would you want to manually check the results?19:10
wolfspraulor is it possible that you would want to fill some zones, and leave others unfilled?19:11
wpwrakit should always be safe, assuming your design has passed DRC19:11
wpwrakof course, if you choose to produce something you haven't checked, well ... ;-)19:11
wolfspraulthe fill/unfill setting is always for the whole board? or can it be partially filled/unfilled?19:11
wpwrakyou can also partially fill/unfill. not sure when you'd want to do that. some of this may just be optimizations for faster interactive use.19:12
wolfspraulhmm19:12
wpwraki can't quite imagine that you'd want a partially filled board19:13
wolfspraulso maybe we do this: we leave the current cmdline behavior as it (i.e. we leave the saved .brd settings untouched), and in addition we introduce a new parameter --fill-all-zones19:13
wolfspraulthat will override whatever is in the file19:13
wpwrakalso considering that DRC will fill everything anyway19:13
wpwraksounds good to me19:13
wolfsprauldrc will fill it?19:15
wolfspraulbut if you run --drc from the cmdline, nothing gets saved19:15
wolfspraulok I think it's clear. I will add a --fill-all-zones option19:15
wpwrak(--drc) yes, don't know why not. it seems that you're not running the entire DRC process19:16
wolfspraulno19:16
wolfspraulI don't save on purpose.19:16
wolfspraulI wouldn't know why printing a drc report should modify the source. that looks very wrong to me.19:17
wpwrak(--drc) e.g., a pcbnew --plot takes about 0.5 s, --drc and --plot takes about 0.6 s. when i do a DRC interactively, it takes about 5 s. so unless you did some very clever optimization, something is missing :)19:17
wolfspraulwait, one by one19:17
wolfspraulI know it's fast, but I do think it checks everything. The GUI is only marginally slower on my notebook.19:17
wolfspraulcomplaining about 'too high speed' is not a good bug report :-)19:18
wolfspraulso unless we know what's wrong with --drc, I assume it's all fine19:18
wpwrakon my pc, the difference is quite large :)19:18
wolfspraulif you want me to save back the fill-zone side-effect when running --drc, I can add that19:18
wolfspraulI would think that's really ugly.19:18
wpwrakwhen i put --plot and --drc, which gets done first ? does it depend on the order in the command line ?19:19
wolfspraulalso it would cause issues with my plans to run some of those reports on the server, at least I would definitely not automatically commit some of those changes back.19:19
wpwrak(save back) you mean to in-core state ? or to the file ?19:19
wolfspraulah good point.19:19
wpwraki don't think you really have a choice about in-core state :)19:20
wolfspraulthat's not really thought through. it depends on the order I check the cmdline options in the sources.19:20
wolfspraulmaybe we should only support one 'major' command per pcbnew process?19:20
wpwrakwhen DRC does its filling, then it will mess with in-core state19:20
wolfspraulyes19:20
wolfspraulbut we can only accept 1 major command, then exit the process19:20
wpwrakyou could. or just put the one(s?) with side-effects first19:20
wpwrakor something :)19:21
wolfspraulno I think I will only support the first major command.19:21
wolfspraulso if you do --plot --drc, it will only plot19:21
wpwrakwarn if there are more ?19:21
wolfspraulif you do --drc --plot, it will only run the drc report19:21
wolfspraulyes I can warn, although the wx cmdline parsing class forces me into certain expected behavior19:22
wolfspraulonce we move away from that, the code gets ugly (but no problem)19:22
wolfspraulok, so that's settled too? only 1 major command per process...19:23
wolfspraul1) --fill-all-zones19:23
wolfspraul2) only process first major command, warn if there are more19:24
wpwraksounds fine19:24
wpwraka warning would be useful because it may not be obvious for the user what constitutes a "major command". and then you get silent failure.19:25
wpwraki also expected that wx may not be overly helpful here. that's why i suggested to just make the order deterministic but accept any set of commands. as long as all commands get processed, there's no error to report or warn about ;-)19:26
wolfspraulI rather operate more narrowly. one command per process is enough, then it's cleanup time :-)19:28
wolfspraulyes the wx --help thing is quite limited, they work from a structure and I need to fill my stuff in there.19:28
wolfspraulwarning is good19:28
wolfspraulI don't think it's a good idea to run multiple totally different files in one shot anyway.19:28
wolfsprauland it introduces potentially very hard to track down error cases. not worth it! Need to focus...19:29
wpwrakheh, whatever works :) for my Makefiles, one file at a time is nice enough19:30
wpwrakah nice. today, we had a "subjective" temperature > 40 C19:33
wolfspraulzrafa: you lost some files on the openmoko projects server?19:39
zrafawolfspraul: yes, the web page of the fatfinger shell. ANd the web page had a copy of the sources as well IIRC19:56
zrafakristianpaul: http://www.xatakamovil.com/motorola/game-gripper-sencilla-forma-de-convertir-un-teclado-qwerty-en-un-pad-para-videojuegos19:56
zrafakristianpaul: that would solve our game pad idea :)19:56
zrafabah.. wolfgang is not here19:58
andres-calderonzrafa, kristianpaul: ttp://www.youtube.com/watch?v=GVfydHEsXX819:58
zrafaandres-calderon: I would like something a lot easier :) http://www.engadget.com/2010/04/23/game-gripper-review/19:59
zrafaandres-calderon: check the idea.. it is just a game pad which press keys on qwerty :D19:59
andres-calderonnice, easy to implement through the microSD slot20:01
wpwrakadd an avr and connect an old analog joystick ? :)20:02
andres-calderonI like the openpandora keyboard:  http://www.openpandora.org/20:03
zrafaandres-calderon: I think that the gamepad link I pasted just press keys on qwerty20:04
zrafaandres-calderon: it is not connected to any IO20:04
andres-calderonandres-calderon: wow... so trivial :)20:09
zrafakristianpaul: kyak: did you see the doxbox for dingoo?.. it seems me that those guys are running dosbox (I tested their binary on nn and it works) like if the screen would have 640x480. It would be interesting to check how them are running an application which uses 640x480 on a 320x240. Of course, surely they are doing some scaling.. but it would be nice to know how, so we can run any 640x480 application on our screens. A new world would appear on our fro20:17
zrafaI have no much free time, that is why I just point the idea :)20:18
qi-bot[commit] Werner Almesberger: Patches to enhance KiCad, mainly for non-interactive use http://qi-hw.com/p/eda-tools/d2da52b21:05
qi-bot[commit] Werner Almesberger: atusb.brd: saving with the zones filled, to avoid surprises http://qi-hw.com/p/ben-wpan/5dde61722:30
qi-bot[commit] Werner Almesberger: atusb/Makefile: targets "front" and "back" to print layers for toner transfer http://qi-hw.com/p/ben-wpan/fe6070522:30
qi-bot[commit] Werner Almesberger: atusb/cam2/: streamlined CAM data processing, added generation of prerequisites http://qi-hw.com/p/ben-wpan/a0b064b22:30
qi-bot[commit] Werner Almesberger: atusb: added "clean" targets to main and CAM makefile http://qi-hw.com/p/ben-wpan/3385afc22:30
--- Tue Dec 28 201000:00

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