#qi-hardware IRC log for Saturday, 2010-10-09

qi-bot[commit] Alan Post: silence error when package has no patches directory. http://qi-hw.com/p/openwrt-xburst/d340b5b00:18
qi-bot[commit] Alan Post: check in package lojban-wordlists http://qi-hw.com/p/openwrt-packages/725355200:18
qi-bot[commit] Alan Post: jbofihe: The Lojban Parser http://qi-hw.com/p/openwrt-packages/8834e0200:18
qi-bot[commit] Alan Post: Add package description for makfa. http://qi-hw.com/p/openwrt-packages/8c2ff7800:18
qi-bot[commit] Alan Post: Move makfa from education to Languages/Lojban. http://qi-hw.com/p/openwrt-packages/9ef379300:18
qi-bot[commit] Alan Post: Record myself as the maintainer of makfa. http://qi-hw.com/p/openwrt-packages/a3c756200:18
qi-bot[commit] Alan Post: makfa package python cleanup http://qi-hw.com/p/openwrt-packages/0847b7300:18
aisaIt is adding two packages and cleaning up the original package I ported, now that I know more about OpenWrt.00:26
wolfspraulthat first patch into openwrt-xburst, let me see00:26
aisabut as is my habit, I split each change in the cleanup into its own patch.00:26
aisaI'll bet that one is upstream in trunk.00:26
aisaI didn't check,00:26
aisabut it is a very obvious bugfix.00:27
wolfspraulin general we try to avoid any diff between openwrt-xburst and upstream openwrt that is not absolutely necessary00:27
wolfspraulyes great, then maybe submit it as a patch upstream too00:27
aisaI could post it to OpenWrt devel.00:27
aisaI was just about to ask ;-)00:27
wolfspraulprobably a good idea, although I am also not 100% sure what upstreaming workflow the openwrt guys prefer00:27
aisaI'll ask as part of my bug report, those guys have been great to me so far.00:28
wolfsprauljust try out see what happens...00:28
wolfspraulok sounds good00:28
wolfspraulin openwrt-packages, in general we are much more free00:28
wolfspraulbecause even from the perspective of upstream, that can be seen like an additional 'feed' (=package repository)00:28
wolfspraulso it's a complement to upstream00:28
wolfspraulif one of the packages in openwrt-packages would show up in another feed, we could remove it in openwrt-packages00:29
wolfspraulbut since many of them are console tools or NanoNote specific, or at least assume a screen/keyboard etc. I think it might be that openwrt-packages becomes another stable long-term feed for OpenWrt00:29
wolfspraulbut in openwrt-xburst, we are trying to manage (i.e. keep small) the diff to upstream at any time00:29
aisaIndeed.  My previous change I'll be shortly reverting and placing somewhere else, so it won't show up as a diff vs upstream.00:30
aisabut I want to first test this permission problem I thought I saw.00:30
wolfspraulsure sure. don't feel kicked around. I am just explaining the current school of thought.00:30
wolfspraulall of this can change too.00:30
wolfspraulmirko made some good suggestions to me recently in a phone call00:30
aisaI've tried to "behave as if" and act as if you were 100% right,00:31
wolfspraulhe proposed to delete the 'tracking_trunk' branch in openwrt-xburst, because nobody is active there00:31
aisataking your advice that I should get to know how Qi works first.00:31
aisaSo far it has been a good experience.00:31
wolfspraulpeople interested in that could take the upstream trunk as a starting point instead, tracking_trunk is just a trap00:31
wolfspraulI agree00:31
aisaI'm not sure about tracking_trunk, yeah.00:31
wolfspraulthen he suggested the name 'tracking_backfire' is not ideal, he proposed that 'tracking_backfire' should do just that, i.e. track backfire00:31
aisait isn't like we'd even use it for the next stable branch.00:31
wolfsprauland there should be a second branch named 'backfire_qi' or so00:31
wolfspraulbut then we said we wait with another rename until more pain has built up :-)00:32
wolfspraulalso we need to keep in mind that Milkymist is coming00:32
wolfspraulwhich I hope over the next year or so we see support for in OpenWrt as well00:32
wolfspraulthen the name of the whole project would be wrong 'openwrt-xburst'00:32
wolfspraulbut instead of renaming like crazy now disturbing many people, let's wait until there is more need for it00:33
aisaDo you remember the mailing list message about LibreWrt?  And their desire to get FSF sanction as a free distribution?00:33
wolfspraultracking_trunk can be deleted I think, if there are no objections I think it will just be done soon00:33
aisaI've been wondering if as part of that, or even just for testing,00:33
aisawe might need an x86 branch at some point.00:33
wolfspraulah that's a long story00:33
wolfspraulthe main LibreWrt guy is on this channel, djbclark00:34
aisaI'm ok deleting tracking_trunk, personally.00:34
aisadjbclark: I'd like to brainstorm with you a bit, if you see this.  I'm User:Alanpost on the wiki.00:34
aisathat takes care of that ;-)00:34
wolfspraulif you read between the lines of his last mail or post somewhere (I think it was him), it read some people might not see the need or usefulness of FSF endorsement in the first place00:35
wolfspraulhe he00:35
wolfspraulwell, between the lines, some of those people are me00:35
wolfspraulthe FSF has great merits in free software, copyright, and the GPL00:35
wolfspraulbut in copyleft hardware they have proven to be useless00:36
wolfspraulin fact, counter-productive00:36
aisadang.  It certainly isn't their primary focus.00:36
wolfspraulthey will endorse a product with hidden proprietary software over the same product where the same (!) proprietary software is exposed00:36
wolfspraulsince we need to extend the freedoms into the hardware world, that's really bad to us00:37
wolfspraulbecause the FSF will rather have the things we are trying to free be locked away00:37
wolfspraulbut we need it the other way round00:37
wolfspraulwe need to expose any proprietary stuff, and write free replacements00:37
wolfspraulOpenWrt is a tool in that process00:37
wolfspraulthank you OpenWrt00:38
wolfspraulno thanks, I don't need you FSF00:38
aisaThe bugfix I made to OpenWrt isn't in trunk either.00:38
aisaw00t, I fixed a bug in OpenWrt :-)00:38
wolfspraulhaving said that, I hope djbclark can become an active distributor of copyleft hardware in the US :-) (www.freedomincluded.com is his site)00:38
wolfspraulso I hope they can finish LibreWrt and start shipping, or maybe now that they realized 'Wrt' is too much of an advertisement for the wrong thing, they rename it first. don't know. ask them.00:39
wolfspraulthey are advertising the sale of hidden proprietary software all the time00:39
wolfspraulbut since it's hidden they can claim it doesn't exist, even though they know better00:39
wolfspraulpretty amazing00:39
wolfspraula bit Chinese actually :-)00:39
wolfspraulaisa: bottom line: I focus on OpenWrt, OpenWrt is great. LibreWrt solves a strange problem from my perspective, although the result will still be free software, and run on the NanoNote, so fine.00:40
wolfspraulis anyone HURD?00:41
aisaHaving worked with OpenWrt now, I really understand why you like it.00:41
aisaIt is amazing.00:41
aisaI sent the patch to openwrt-devel.00:46
wolfspraulif you are already at the level that you can provide actual fixes to them that would be awesome00:46
wolfspraullet's see...00:46
wolfspraulaisa: hmm, looking at the patch, you are removing the redirection of stderr to /dev/null ?00:54
wolfsprauland that makes it ignore the missing directory?00:54
wolfspraul-       @if [ -d "$(2)" -a "$$$$(ls $(2) 2>/dev/null | wc -l)" -gt 0 ]; then \00:55
wolfspraul+       @if [ -d "$(2)" -a "$$$$(ls $(2) | wc -l)" -gt 0 ]; then \00:55
wolfspraul(for reference)00:55
aisait sends the error, which happens on stderr, to /dev/null.00:55
aisaso, I'm adding it.00:55
aisadid I get the patch backwards?00:55
wolfspraulah :-)00:55
wolfspraulthat explains it :-)00:55
wolfspraulredirecting stderr to /dev/null also isn't the nicest thing to do. why not add a separate check whether the directory exists?00:56
aisaI think that is the better plan,00:56
wolfspraulwell, let's see what feedback they give...00:56
aisaok, corrected patch sent.00:57
aisathank you for catching that!00:57
wolfspraulif you still redirect I wouldn't be surprised if they don't like it00:58
wolfspraulI wouldn't00:58
wolfspraulI'm always amazed how well this stderr and stdout stuff works00:58
wolfspraulsometimes I run a high-level command on a server, and it fails00:58
wolfspraulthen I think "oh my god, how long will it take until I figured out where in the system the error comes from"00:58
wolfspraulbut there it is... from deep inside the scripting and utilities magic... reported all the way through to me00:59
wolfspraul(sometimes, I admit, but when it works it's really nice)00:59
wolfspraulbut that only works if people have some discipline in not silencing stdout/stderr, but keeping them untouched so the entire stack can benefit00:59
wolfspraulif I assume I'm the topmost script/caller I admit I also often just silence it, to get my job done01:00
aisaIndeed.  If I understood better the syntax being used right there, I'd do if [ -f $(1) ]; ls $(1)01:00
aisaor something like it.01:00
aisaI think it is good to start with a solution that can be clearly improved on, it gives someone an opportunity to teach :-)01:01
aisaAlso, I copied that strategy from elsewhere in the code.01:01
wolfspraulhe, yeah01:01
wolfspraulI also don't know right away how to get the check inserted01:01
aisaoh, hey: question01:02
wolfspraulisn't the -d at the beginning checking whether the directory exists?01:02
aisaoh, hmmm.  Well, it isn't working :-)01:02
aisaWill you look at the header at the top of that url?01:02
aisaI have no idea what I should put for copyright and/or boilerplate.01:03
aisaat the top of my package Makefiles.01:03
wolfspraulif you are the author, copyright is yours01:03
wolfspraulas for license preference, mine is GPL v3 or higher01:03
aisayou're right, something deeper is going on there.01:03
aisaexcellent.  I'll clean them up next time I come around.01:04
aisaI think that -d line is broken,01:04
wolfspraulhe, OK I see your header01:04
wolfspraulthe license really is your choice01:04
aisaI can't figure it out yet.01:04
wolfspraulbut you currently have 'v2 only' (not even 'or higher')01:04
aisaI copied that from somewhere,01:05
wolfspraulmy choice would be 'v3 or higher'01:05
aisait is going to say that elsewhere too.01:05
aisayeah, no reason to use v2 any longer.01:05
wolfspraulalso don't forget the 'or higher', unless you specifically don't want that01:05
aisaright, that is very important.01:05
wolfspraulfor copyright, it's probably better if you keep yourself in there01:05
wolfspraulwe have no system of copyright transfers or handovers at qi-hardware.com and nobody is planning on setting one up01:06
wolfsprauland I don't like to hoarde copyrights unless there is real management/representation along with it01:06
wolfspraulwe are giving a stronger picture if we put the copyright on more shoulders, on the real authors01:07
wolfspraulthat's again just my opinion, but since you ask I answer01:07
qi-bot[commit] Alan Post: Fix license statement to conform to gplv3 and properly credit author. http://qi-hw.com/p/openwrt-packages/db7d69101:10
wolfspraulhe, perfect :-)01:11
wolfspraulI'm not sure what the latest guidelines are about author reachability, email addresses etc.01:12
wolfspraulin some (rare but important) cases the ability to reach many authors of whatever project quickly and effectively can be very helpful01:13
aisaah, indeed.01:13
aisaI though that having my name there would be tied to my commit access.01:13
aisaand of course, if my e-mail address changes, I can update in one place.01:13
wolfspraulsure it may be01:13
aisahowever, this is only true so long as the code is stored at this repository.01:13
aisaand being free software,01:14
aisait may well get copied around.01:14
wolfspraulbut like I said "quickly and effectively"01:14
wolfspraulI don't know what the latest state-of-the-art there is.01:14
wolfspraulbut just "Alan Post" is probably not ideal01:14
aisalet me think the night on this one.01:14
wolfspraulI always check with the Linux kernel, they tend to have the most verbose and thought through policies.01:14
aisaIt could be best to have an AUTHORS file,01:14
aisawith many ways of contacting.01:15
aisaor it could be better to add an e-mail to every header.01:15
aisaah, indeed.01:15
aisawonderful idea.01:15
wolfspraulabout why -d doesn't work..01:17
wolfsprauldo you know why there are 4 $$$$ ?01:17
wolfspraulit isn't immediately obvious to me01:17
aisaI don't know, other than two escapes.01:17
wolfspraulmaybe there need to be 4 $$$$ in the -d also ? :-)01:18
wolfsprauljust guessing around...01:18
aisaha!  right.01:18
wolfspraulwell not sure01:18
aisaHey, it is late here for me, and now that I've made these commits, I have accomplished my goal for the evening.01:18
aisaI'm going to go to sleep.01:18
wolfspraulafter -d is a test after -a (stands for 'and', right) is an expression01:18
aisaI believe that is all true.01:19
wolfspraulok 'night01:19
wpwraki think that should be ` ... ` instead of $(...)01:27
wpwrakor wait, no, i'm confusing this with $(shell )01:27
wpwrakstrange then01:28
wpwrakah no. makes sense. use [ X ] && [ Y ] instead of [ X -a Y ] :)01:32
wolfsprauloh the -a will not abort after the left part failed?01:35
wolfspraulI didn't know that...01:35
wolfspraulI would have thought every language does that01:35
wpwrakthe -a can't. you're invoking "test" with the arguments already expanded. the error happens in the expansion.01:36
wpwrakon the other hand, with &&, the shell sees that the left-hand side is false and doesn't expand the right-hand side01:37
wolfspraulso the fixed line is? @if [ -d "$(2)" ] && [ "$$$$(ls $(2) 2>/dev/null | wc -l)" -gt 0 ]; then \01:40
wolfspraul(too lazy to try this out)01:40
wolfspraulwpwrak: can I delete s.try on fidelio:/home/schhist ?01:40
wpwrakme too ;-)  looks good excelt that you don't need the 2>/dev/null01:41
wolfspraulah sure01:41
wolfspraulaisa: try this one @if [ -d "$(2)" ] && [ "$$$$(ls $(2) | wc -l)" -gt 0 ]; then \01:42
wpwrakor, much simpler [ ! -z `ls $(2)` ]01:43
wolfspraul(hope this won't wake you up, but I cannot imagine anyone configuring their irc client to connect nick mentioning to an alarm bell or such :-))01:43
wpwrak(for the 2nd expression. still need the first)01:43
wolfspraul@if [ -d "$(2)" ] && [ ! -z `ls $(2)` ]; then \01:44
wpwrakor, still better01:47
wpwrak@if [ -d "$(2)" ] && [ `ls $(2)` ]; then \01:47
wpwrakah no, that's too efficient :)01:49
wpwrak@if [ -d "$(2)" ] && [ "`ls $(2)`" ]; then \01:49
wolfspraulabout schhist, we leave the SCHHIST_ADDLIBS=device on for every project?01:49
wolfspraulcan I delete s.try?01:50
wpwrakthe ls $(2) is also potential trouble, though. if $(2) has a space, this will fail01:50
wpwrak(s.try) ah yes, please01:50
wpwrak(devices) not sure how common the problem is. i would hope it isn't, but i didn't examine the other projects.01:51
wolfspraulok but in a server/buildhost environment, is there any advantage of not setting it?01:51
wolfspraulbecause if not, I will just always set it01:51
wpwrakin case you have your own set of replacements with the same name, "devices" could get in the way01:52
wpwrakwe,, i'm not sure entirely about that. since it gets added at the end, maybe it's searched last, so it wouldn't matter01:53
wolfspraulI leave it in for all right now01:53
wpwrakthat's fine for now01:53
kyakkristianpaul: in fact, everything was already there.. had to port mpfr and gcc, now fighting with uClibc (had to copy some dev files manually)02:47
wolfspraulkyak: what are you porting?02:54
kyakit's gcc02:55
wolfspraulkyak: you mean into OpenWrt?02:58
wolfsprauloh wow02:59
kyakactually, this is working already.. but  not clean - had to copy some files manually, and links only statically03:00
qi-bot[commit] kyak: gcc: initial port http://qi-hw.com/p/openwrt-packages/285555c03:39
qi-bot[commit] kyak: Merge branch 'master' of projects.qi-hardware.com:openwrt-packages http://qi-hw.com/p/openwrt-packages/568ac7b03:39
kyaki've commited it for now, but it still requires some polishing (it builds only with "-static")03:40
kyakalso the package size is ~11 Mb, i think some files can be thrown away03:41
kyakbut it's good to test03:43
wolfspraulnice, congratulations!03:51
kyakthanks, but as i said, it was already there :)03:53
wolfspraulsorry can't follow now03:54
wolfspraulit was there? where?03:54
kyakbinutils, gmp were already ported; mpfr is a part of toolchain, i only packaged it; uClibc is in toolchain; gcc was also ported for x86, i just had to modify it03:57
wolfspraulah OK, now I got it :-)03:58
wolfspraulsorry..., lack of overview03:58
kyakit is amazingly slow, btw.. during usual ./configure at takes ~1-2 seconds for each test line to appear04:31
kyakfor example: the 'patch' utility takes 0m 06.24s for configure, 0m 02.78s for make on my host.. and on Ben configure takes 3m 10.02s and make is 5m 49.95s04:52
kyaki guess i have to throw away the idea of building kernel and rootfs on Ben :)04:53
wolfspraulthat's a pretty radical idea anyway04:54
wolfspraulI think gcc is nice to have, in a number of cases04:54
wolfsprauland also patch, make, etc.04:54
wolfspraulbut to build the rootfs? not sure...04:54
wolfspraulbut hey, who knows. maybe performance can be improved? don't know.04:55
kyaknow, it's just for fun04:55
wolfspraulI'm wondering how this was when we had Pentium-90 machines04:56
kyakbesides, would be cool to state to someone that Ben is self-sufficient and can be used to build a firmware for itself04:56
wolfspraulI think I remember having had 16 and 33 mhz 386 computers.04:56
wolfspraulyes agree04:56
kyaki rememeber i built linux kernel on AMD K6-2 (450 Mhz)04:57
kyakit took couple of days.. i wonder if that COU is fast than that of Ben04:57
kyakwould be good to hae an understanding what is the analogue of our JZ4740 CPU in x86 world05:00
kyaklike, it is PEntium-I? or II?05:00
wolfspraulhmm, not sure05:08
wolfspraulthere are many differences, in instruction set, pipeline/caching architecture, bus speed and width05:08
wolfspraulbest is probably to just run the same load and then compare, like building patch :-)05:09
kyakyep.. now to find such old trash :)05:10
kyaki wonder if this can be emulated in virtual machine05:14
tuxbrainkyak: time ago there was a benchmark test with pentium openmoko nanonote and pentiumII search the archives05:34
tuxbrainkiak my mistake it was other procesors but a benchmark any way it was called "A few benchmarks for the Nanonote" from Delbert Franz at May 4th05:36
kyaktuxbrain: thanks for the hint05:47
kyakyes, nice post from Delbert Franz05:50
kyakas i see, Ben is 2 to 10 times slower than a PC of Pentium II class05:51
B_LizzardI just fixed me a nice gtk-1.2 theme08:06
B_LizzardNext up, implementing applications08:06
wpwrakkyak: maybe try to use uSD instead of NAND or vice versa. also, adding swap may help. it allows the kernel to get rid of idle pages, so it can do better caching.09:12
wpwrakwolfspraul: on those older systems, we also had leaner tools. e.g., gcc 1.something ;-)09:13
wpwrakwell, on the 386. on the pentium, it was already gcc 2 or such. i remember that going from gcc 1 to 2 was quite a bit of pain, because it was so slow.09:14
wolfspraulthe first version I used was 2.9509:15
wolfspraulbefore that only Microsoft compilers09:15
wolfspraulor Borland before that :-)09:15
wolfspraulwpwrak: did you get my reply about the patent policy?09:17
wolfspraulI think we should just upload the text to that page, then we can go from there09:17
wolfspraulit's a wiki :-)09:17
kristianpaulxiangfu: i'm confused, i tough there was a target system for SIE in xburst openwrt?..09:18
wolfspraulthe hunt for the nonexisting SIE target maintainer has started09:18
wolfspraulkristianpaul: maybe you can say some magic words in Spanish to Carlos, and he will switch from massive extremely hacky patches that are sent around as attachments in private mails to something a bit more collaboration friendly :-)09:19
wolfspraulwe have patches flying aroud that brutally just comment out stuff in the Ben sources, and replace it with SIE stuff. If we would 'apply' those, the Ben build would, small side-effect, completely break right away09:20
wolfspraulso big surprise - the patches cannot be applied :-)09:20
wolfspraulfirst step would be to stop patching (breaking) Ben sources, and instead have a few dedicated SIE source files where it's necessary09:20
wpwrakwolfspraul: (policy) i saw that i have mail. will read it over breakfast :)09:20
wolfsprauloh sure take your time09:20
wolfspraulyour text is the best that was written down on that subject since Qi was started09:21
wolfspraulwe don't seem to have the people (me included) that have enough motivation to try to tame that mess with a good text09:21
wolfspraullook at the patent-free page :-)09:21
wolfspraulthat's as much as we could squeeze out from our results-oriented minds...09:21
kristianpaulwolfspraul: so the image my SIE board have came up from a Carlos private mail?.. :/09:21
wpwrakwolfspraul: (magic words) kristianpaul said that he doesn't know carlos. so he may not date to use "boludo", which is i think the word you're looking for ;-) (well, i'm not sure they have it in colombia)09:22
wolfspraulthey know each other a bit, I mean we all hung out at the Campus Party together09:22
kristianpaulwpwrak: we dont but there are some similars tought09:22
wolfspraulkristianpaul: I don't know09:22
wolfspraulthe sources are all there09:22
wolfspraulthere is not that much missing - BUT, the big force of gravity that would clean this up and make it easily accessible to others does not exist09:23
wpwrakwolfspraul: (patent-free) okay, stiff competition :)09:23
wolfspraulthe 'SIE target maintainer'09:23
wolfsprauldoes not exist09:23
wolfspraulthis is also not helped by very few people having the board09:23
wolfspraul95% of which are in Colombia09:23
wolfspraulit's definitely not the commit rights into openwrt-xburst that stops it09:24
kristianpauli have a word for that is "mess"09:24
wolfspraulwpwrak: yeah that page is great, isn't it?09:24
wolfspraulI think Jon wrote it after some lengthy discussions with me.09:24
kristianpaulpatent-free great :)09:25
wolfspraulwe were both exhausted. in my case that translated to 'sleep', in his case I guess he managed to squeeze this out of his fingers before getting drunk or so.09:25
wpwrakwolfspraul: it's very to the point on the motivation. alas, a bit light on the practical implications09:25
wolfspraulyeah come on. time is limited. I'm no lawyer, cannot charge anybody for my words.09:25
wolfspraulanyway it's the best/most Qi has managed to write up on the subject so far09:26
wolfspraulmaybe you can leave it as an intro09:26
wolfspraulI still like it.09:26
wpwrakhmm, its scope is broader than the policy. the policy is only about avoiding trouble with other people's patents.09:28
wpwrak"patent-free" is mainly about the use of patents on the qi-hw side09:28
wolfsprauldoes this mean the policy will get even longer before it says what the current text already says?09:29
wolfspraul(just joking)09:29
wolfspraulyou asked for my feedback. my feedback is the text is great, add it to that page, and if you ask me - leave the current 2 lines on that page as an intro or somehow merged into your text somewhere.09:29
wpwrakit woud have to get longer to cover that stuff at the same level of detail as well :-(09:30
wolfspraullet's just upload it now, merge with the existing 2 lines in a smart way, and done09:30
wolfspraulwe can polish / extend this over time09:30
wolfspraulthe important thing is to clearly communicate to newcomers or interested people what we collectively believe in09:31
wpwrakbut i think it's two completely different audiences anyway. one is for those who wonder if they/we should patent something. the other is something anyone placing public content on qi-hw should be aware of09:31
wolfspraulcould be even more motivations to read this09:31
wolfspraul"virtual warfare over an artificial problem"09:32
wolfspraulhe :-)09:32
wpwrakit's more like patent philosophy vs. appropriate use of the public qi-hw infrastructure09:33
kristianpaularggg carlos !!09:33
kristianpaul(chatting privatelly)09:34
wpwrakso i wouldn't mix the two. too distant. the patent-free text is for the legislative, the policy for the executive. don't make the cops read works contemplating the human condition. at least not on the job ;-)09:34
wpwrak(virtual warfare) that's a nice one :)09:35
wolfspraulwpwrak: well ok let's translate to reality. the policy needs to go into the wiki. if you want a separate page just create one.09:36
wolfspraullet's just get this text online then it can start its life.09:36
wpwrakalright. i can put a link to patent-free. something in the sense of "we don't like patents [link] but we have to play by the rules."09:37
wolfspraulbasically you say one text is about what we are trying to achieve (around patents). the other one is about how we are going to achieve it.09:37
wolfspraulwhat we are trying to achieve is to have the strongest, highest performing, most innovative, yet 100% patent-free hardware09:38
wolfsprauland how we are going to achieve it is documented in the patent policy09:38
wpwrakeven less. patent-free is a little about what we're trying to achieve and more about what part we plan to play in the patent system.09:38
wpwrakthe policy is about what rules apply for content on qi-hw.09:39
wolfsprauljust let us know what we need to take out, and for how long09:39
wolfspraulwe focus on innovations elsewhere09:39
wpwrakthe patent policy doesn't cover areas like whether we should apply for patent, how we should disseminate our prior art, how we should respond to patents covering out prior art, how else we can help fight patents, and so on.09:40
wolfspraulfair enough09:41
wpwrakon the other hand, the policy could be extended to cover other areas of IP. it's basically the guideslines for how a contributor to qi-hw should behave to avoid getting others (qi-hw, distributors) in trouble.09:41
wolfspraulwe would need community members that are passionate about these things and start acting on those issues09:41
wolfspraulafaik we don't have them right now09:42
wolfspraulwell I think any text needs to be reasonably specific otherwise we are wasting everybodys time09:42
wolfspraulso it should be about 'patents', and about 'mp3/h.264/mpeg4', just as you did09:42
wolfspraulyour text is great09:42
wolfspraulwhy don't you just upload it then we have a base...09:43
wpwrakone is clearly for activists. those who choose to simply ignore patents as much as possible don't need to read "patent-free" and any proactive recommendations that may get added to it.09:43
wpwrakthe other is for everyone. road rules. spitting on the street allowed in our village, but please use a toilet for the rest.09:44
wpwrak(my text) there are a few questions in there, in square brackets. any comments on these ?09:45
wolfspraulI need some valium or so to actually read every word.09:45
wpwrakparticularly, the one on "planet"09:46
wpwrakthe others are more editorial09:46
wolfspraulone sec09:46
wpwrakyeah, writing that stuff does get you angry, doesn't it :) i just HAD to write that last section to let off steam :)09:46
kristianpaulwolfspraul: i giveup with carlos, sorry :)09:47
wpwrakkristianpaul: let us know the words you used :)09:47
kristianpaulwpwrak: (pues bueno..)09:47
kristianpaulchao !09:48
kristianpaulthats it ;)09:48
wolfspraulkristianpaul: what does that mean?09:49
wpwraki think "shrug. bye then"09:49
wolfspraulthe planet...09:51
wolfspraulwell since the policy is for people to guide them on actions, ok, maybe it's similar to irc09:51
wpwrakwolfspraul: http://www.heise.de/newsticker/meldung/Microsoft-ruestet-sich-gegen-Smartphone-Patentklagen-1104882.html09:52
wolfspraulif you are hosting a blog that is syndicated in the planet, and you are posting instructions that infringe, the qi-hardware web admins need to censor your post from the archives09:52
wolfspraulif you do it frequently/all the time, the qi web admins may remove your blog from the planet09:53
wpwrakwolfspraul: (planet) the thing is, planet re-publishes content of others who may not even know that we're using their stuff.09:53
wpwrakwolfspraul: okay, so edit after the fact09:53
wpwrakthat works09:53
wolfspraulbtw, the planet html page itself has no archive, but entries are visible for 1-2 months or so09:53
wolfspraulbut then we have 2 mailing lists (rss2email), which also have archives09:53
wolfspraulbut anyway09:53
wolfspraulthat is all the problem of the qi admins09:54
wolfspraulbecause we want to say to manufacturers or distributors that we are keeping our technology in such a state that if those guys manufacture or sell the resulting devices, the chances of someone claiming infringement of their patents are very low09:54
wolfspraulwpwrak: of course, same on the list or irc09:55
wolfspraulwe still have the concept of freedom of speech09:55
wolfspraulthat's a 'passing' conversation09:55
wolfspraulthe difference is if someone archives that, and keeps it available for others to reuse09:55
wolfspraulthat someone cannot just claim 'freedom of speech' anything, in fact it wasn't even that someone who spoke09:56
wolfspraulso why keep the archives?09:56
wolfspraulTO INFRINGE!09:56
wolfspraulkristianpaul: well I tried to understand what you asked for, and why the conversation ended.09:57
wpwrakso you'd say also the mailing list should be treated as being as ephemeral as IRC ?09:57
wolfsprauland how can we now help you?09:57
wpwrakor a little less ?09:57
wolfspraulephemeral? wait I need to look this up...09:57
wpwrakvolatile, not lasting09:57
wpwrak"vergaenglich" :)09:57
wolfspraulfrom the perspective of a patent owner there is no difference between an irc log and a mailing list archive09:58
wolfspraulit's text, hosted on a server, server is run/paid for by abc09:58
kristianpaulwolfspraul: iwas a talk about, why you this thet should do it and not you..09:58
wpwrakgood that the irc logs are being archived. that can serve as proof for due diligence ;-)09:59
wolfspraulif someone posts a mail to the mailing list, with instructions/links, then the qi admins have to remove this mail, or parts of it, from the archives09:59
wolfspraulthat's our service to the copyleft hardware community, to keep the copyleft hardware manufacturable and sellable with small risk of successful patent infringement claims09:59
wolfspraulI wouldn't think that makes the mailing list 'ephemeral'09:59
wolfspraulthe issue is about WHAT you post09:59
wolfspraulI mean we would also remove spam09:59
wolfsprauland we also remove stuff that will make it possible for patent holders to launch a successful patent infringement claim10:00
wpwrakmailing lists get replicated more than irc goes. e.g., each subscriber implicitly gets a copy, gmane may have qi-hw, etc.10:00
wolfspraulno problem10:00
wolfspraulgmane is someone else anyway10:00
wolfspraulthe issue is only about what is hosted on qi-hardware.com servers10:00
wolfspraulthe copies that are sent to people don't matter either10:01
wolfspraulit is an undue burden that the qi admins would have to review every mail before it is being sent to the list10:01
wolfspraulkristianpaul: ok but what is your problem, and how will you solve it now?10:02
wpwrakthat's one thing. violation of usage policy. but can the usage policy itself allow such things then ?10:02
wolfspraulyou have to drill into every last corner10:02
wolfspraulI don't know.10:02
kristianpaulwolfspraul: was pity, (missing password) so i think i know what do to to reflash the SIE and make some changes i need to the FPGA10:03
wolfspraulI don't think we need to, or even would have the authority to, restrict someones rights to speak freely.10:03
wpwrak /n the_sadist (-:C10:03
wolfspraulthey will just do it.10:03
wolfspraulthe policy doesn't need to pickup this fight10:03
wolfspraulit can jsut say "not tolerated and will be removed from archives"10:03
wolfspraulthe policy neither needs to encourage people to use the list to distribute infringement stuff, nor does it need to install any sort of gag order, which couldn't be enforced anyway (we already say we remove such stuff from the archives)10:05
wpwrakhmm. i can see how to put the cya angle. but it would then fall short of giving people specific guidelines for what they should and shouldn't do.10:05
wolfspraulok let's talk about friendliness then10:05
wolfspraulif you want to be friendly to the project, do not use project resources to spread information about how to get mp3 to play, mpeg4, h.26410:06
wolfspraulit doesn't help us10:06
wpwrakalright. that's good.10:06
wolfspraulbecause we cannot include such functionality anyway, and in fact we need to constantly cleanup behind out chatty community members :-)10:06
wolfspraulbehind our10:07
wolfspraulkristianpaul: do you think you are able to create the full image you have on your SIE from source?10:08
wolfspraulor is it entirely undocumented/messy?10:08
kristianpaulwell not so messy atl all10:08
kristianpaulwolfspraul: yes sure, the source actually is same Ben image :)10:09
kristianpaulreplace root_fs and thats it10:09
kristianpauland uboot too, but i dont need reflash uboot10:09
wolfspraulthe kernel is not the same10:09
kristianpauli think is, or carlos said that10:09
wolfspraulok I'm a software guy, when I say 'same' I mean md5sum(ben_uImage) == md5sum(sie_uImage)10:10
wolfspraulI have massive doubts that that is the case.10:10
wolfspraulmassive as in 99.99999%10:10
wolfspraulthe only .000001% only to save Carlos face :-)10:11
wolfspraulbut seriously, no, there must be changes10:11
wolfspraulso if you reflash, be careful10:11
wolfspraulyou may end up with a system that won't work anymore, and then getting it back to work may be hard10:11
kristianpaulwell any way if carlos desided to move to something more colabortive will be SIE as a fork from Ben as in sofware and hardware (wich is right now)10:11
wolfspraulI think the SIE target maintainer (in OpenWrt) needs to come from outside of Bogota10:12
wpwrakwouldn't having all the SIE stuff nicely out in the open be a prerequisite for selling it to others than carlos' group ? :)10:12
kristianpaulwe are chatting again about create a project in indefero10:12
wolfspraulwpwrak: it is all there, just messy (imo)10:12
wolfspraulit's not about lack of motivation to do so, it's lack of skills10:12
kristianpaulidependent from uboot,linux,root fs  frm the current one10:13
kristianpaulwolfspraul: (skills) agred10:13
wpwrakah, that's harder. but it also seems to be in part poor practice - you mentioned patches sent by private mail and such.10:13
kristianpaulcarlos students are so vage mostly so nobody  besides carlos to support sofware side.10:13
kristianpaulthast why i ends in privatelly mails to qi people ;)10:13
wolfspraulwpwrak: yes that too.10:13
wolfspraulthe pieces are all there10:14
wolfspraulbut not in one place, not clean10:14
wolfspraulalso there are too few boards now10:14
wolfspraulmy hope was that David would get 5 or so, and me another 5 or so10:14
kristianpaulwolfspraul: sure not in Bogota for sure10:14
kristianpaulsee what i dont live there ;)10:14
wolfspraulthere are plenty of people that are interested and able to cleanup the software (which is not that much work since the Ben SW is in great shape, I think)10:15
wpwrakstudent work is generally volatile :) PhD students have a long-term interest but all the rest is hard to integrate into such projects. occasionally, someone has good enough skills to get started quickly and it thorough enough to make something stick, but it's not very common10:15
wolfspraulbut now I have 0 boards, and David has 110:15
kristianpaulall the hacks are in the curret nn-usb-fpga10:15
wolfspraulbecause the V2 yields were a little lower than hoped, and Carlos students had priority (and paid)10:15
wolfspraulso I cannot help with the maintainer thing either10:15
wolfspraulwith 0 boards :-)10:15
wolfsprauland David has 110:15
wolfspraulthat won't change until a V3 run which may be months out10:16
kristianpaulbut more features for SIE will come, so it shows definelly that SIE need a new project :)10:16
kristianpaulsoftware porject*10:16
wpwrakadam has given up on fixing the remaining boards ?10:16
kristianpaullarsc: xiangfu openwrt tarjets can habdle different uboots and linux images isnt?10:17
wolfspraulto put my words into perspective a bit, I may be too hash10:17
wolfspraulCarlos has done a great job, all software exists, and, it is based on the Ben software which imo is in excellent shape10:18
wolfspraulmy estimate is it would take a really skilled sw guy maybe 2 weeks to clean it all up in OpenWrt, especially if that person had some prior knowledge of OpenWrt10:18
wolfspraulbut for that this person would need the board, and all patches/hacks/scripts/whatnot10:19
wolfspraulthen 2 weeks, then all is perfect10:19
wolfspraulbut this step will, again imo, not happen in Bogota10:19
wolfsprauland me and David have too few boards10:19
wolfspraulso we are a little stuck on this10:19
wolfsprauleventually we find a way out :-)10:19
wolfspraulwpwrak: yes Adam has stopped, because I asked him to. it's about cost. right after the run 19 boards had problems.10:20
larscbest would be to get the SIE platform patch into upstream linux10:20
wolfspraulhe fixed 9 of them, it got slower and slower.10:20
kristianpaulthe issues i see, is openwrt able to build system that can handle different kernel confifuration even for uboot, and no afect current ben target?10:20
wolfspraulkristianpaul: worst case you can have a uboot-sie package10:20
kristianpaullarsc: not posible yet, well or think so,  as more features are coming10:21
wolfspraulmaybe even as a diff on top of the uboot-xburst? :-) don't know how flexible OpenWrt is10:21
wolfspraulwpwrak: so if Adam would continue fixing those 10 boards, those days would eat into Milkymist One RC210:21
kristianpaullarsc: i think i wroing in the last, coming features dont care isnt?10:21
wolfsprauland we have 27C3 coming up10:21
wolfspraulI cannot take the risk. Adam already spent 5 days or so fixing the first 9 boards, and was quite exhausted at the end. The last 2 days only 1 or 2 came out of the fixing.10:22
larsckristianpaul: well those could be added in later patches10:22
wolfspraulthe remaining 10 would need serious root cause searching, to prepare for the next run. So I'm sitting on them for that purpose.10:22
larsccode is not frozen once upstream10:22
wpwrakwolfspraul: (adam) okay, so he's already saturated with more important stuff. makes sense then.10:24
wolfspraulwpwrak: yes. it's just risky with 27c3 coming.10:26
wolfspraulI cannot let down Sebastien on this.10:26
wolfsprauland there will always be surprises.10:26
wolfspraulit's already tight10:26
wolfspraulI need to move everything out of the way for that now.10:26
wolfspraulonce the RC2 boards are made maybe Adam can go back a few days and try to fix a few more SIE10:26
wolfspraulmaybe late November even? don't know10:27
wolfspraulI need to get RC2 done first10:27
kristianpaulwolfspraul: when RC2 should be done?10:27
wolfspraulcalculate backwards from 27C310:28
kristianpaulwolfspraul: and the boards will be sell in 27C3?10:28
wolfspraul27C3 starts Dec 2710:28
kristianpaulor pre-oder?10:28
wolfspraulBearstech offered their booth10:28
kristianpauli'm just curios10:28
wolfspraulyou don't want to try shipping stuff to europe on Dec 21 or so10:28
wolfspraulso they need to leave Asia no later than first week of December10:28
wolfsprauland they need to be tested10:28
wolfspraul(going backwards)10:28
wolfsprauland we need the jtag/serial cables10:29
wolfsprauland power adapters10:29
wolfsprauland boxes10:29
wpwrakyou have 2-3 weeks to finish the design and order the production run :)10:29
wolfspraulwe need to be careful about customs (first shipment of anything is always risky)10:29
wolfspraulthe SMT date needs to be scheduled10:29
wpwrakpcb have to be made10:30
wolfspraulthe PCBs need to be ordered, made, received, tested10:30
wpwrakparts need to be ordered10:30
wolfspraulbefore that the layout house needs to go back work their way through our change list10:30
wolfspraulit's October 9 now10:30
wolfsprauldoes Adam have time to fix a few more SIE right now?10:30
wpwrakmaybe if he can work negative hours on them :)10:31
wolfspraulAdam may also get sick, or, big surprise, need a few days off somewhere10:31
wolfspraulsometimes vendors make mistakes10:31
wolfspraulship the wrong thing10:31
wolfspraulor apply the wrong production process, which leads to unusable pcbs10:31
wolfspraulor or or10:31
wpwraklet's hope there be no bad shipments. your schedule doesn't supprt that error or margin.10:32
wpwrakerr, margin of error :)10:32
wolfspraulwpwrak: I think it's still OK10:32
wolfspraulit's the second run10:32
wolfspraulnumber of changes is manageable10:32
wolfspraulmany parts still in stock10:33
wpwrakokay if you order next week. are you that close ?10:33
wolfspraulbut I agree this is tight and I explain it all to explain why I really cannot risk loosing days right now over SIE fixes10:33
wpwrakalso, how about the case ?10:33
wolfspraulno case, gave up on that for 27C310:33
wolfspraulcannot make it10:33
wpwraksure sure. "adam is busy elsewhere" explains enough ;-)10:33
wolfspraulwell I like to explain the whole picture10:34
wolfspraulmaybe someone else wants to manufacture 10 or 20 SIE?10:34
wpwrak(case) maybe find some off-the shelf case that's not too horrible10:34
wolfspraulI could probably supply such a run with a lot of helpful material, components, etc.10:34
wpwrak(sie) golden delicious ? :)10:35
wolfspraulI will try to sneak in a few days of SIE fixing (out of those 10), when I feel those days are not on the Milkymist critical path10:35
wolfspraulthe moment I have such days I will try to 'make' another 2-3 fully functional SIE from the 1010:35
wolfspraulbut right now I feel the Milkymist critical path everywhere :-)10:36
wolfspraullogging out, n810:36
qi-bot[commit] Alan Post: separate conditions rather than ignoring error,  thx @werner http://qi-hw.com/p/openwrt-xburst/b48d96a12:42
qi-bot[commit] Alan Post: Add my e-mail address to the copyright header. http://qi-hw.com/p/openwrt-packages/4877e6912:49
tuxbrainmmm, I can ask about the cost of doing 20 Sie boards here..., maybe I need help on the right files to send to the manufacturer, what components would you sendme wolfgang and what I have to source locally12:50
wpwrakaisa: btw, i think the  ls $(2)  is still inconsistent with the  -d "$(2)" . if $(2) needs quoting at all, then it would need quoting also in the ls. but then, i don't know if it actually needs this or not12:53
kristianpaultuxbrain: http://downloads.qi-hardware.com/hardware/sakc/gerber/v2/GERBERs_SIE_V2_20100811.tar.gz < manfacture PCB first12:54
aisaoh!  you're right.12:54
wpwraktuxbrain: while you're here, any comments on http://en.qi-hardware.com/wiki/Patent_Policy_Public_Face before i announce it on the list ?12:54
aisaI think that just means \"$(2)\"12:54
wpwrak(as request for comments for now. everybody can play policy maker here :)12:54
aisatesting now, but I may have to leave before I can finish.12:55
wpwrakaisa: not sure what degree of escaping you need, with make adding one layer of indirection. you'll have to try and (maybe) err :)12:56
kristianpaultuxbrain: but as you said just wolfgang know what he will send you :)12:56
aisaright, #1 didn't work.12:56
tuxbrainok I will ask when he reconnects12:57
aisatrying: '$(2)'12:57
tuxbrainwpwrak: reading page12:57
aisathat seems to have worked.12:57
aisaI suppose I should really test with "patches with space"12:58
wpwrakyes :)12:59
aisabecause thinking about it, '$(2)' shouldn't work.13:00
aisaman, I thought I knew sh and makefile better than I do...13:00
aisaactually, it should.  $(2) is expanded by make.13:01
aisaand that can be single-quoted in sh.13:01
aisanot that it works...13:01
tuxbrainwpwrak: I think also worth to mention about pictures in wiki not totally free, any picture with non-comercial clause, must be avoided too13:04
wpwrakthe $$$$ hints at there being yet another level. at least i don't normally see that sort of thing.13:04
wpwraktuxbrain: so that would be the extension into copyrighted material.13:05
tuxbrainyeah but as you put the emepetres examples I think this example also need to be mentioned13:05
wpwraktuxbrain: not sure if we should cram everything into one document (that would then be on IP, not just patents), or treat this separately13:05
tuxbrainunderetand your poing13:06
wpwraktuxbrain: i would lean towards combining things too, but i'm not sure if i understand the matter well enough to not overlook some subtle differences :)13:06
tuxbrainwow that nap and your text is a bad combination for my finguers13:06
aisawpwrak: I've got to go for today, but I'm testing now and will figure it out.13:06
aisathank you for reminding me.13:06
wpwraklet's do it like this: for now, i'll just cover patents. i'll mention that general IP (copyright, trademarks, etc.) is similar and that it may or may not make sense to include it, too.13:07
wpwrakaisa: always happy to heckle ;-)13:07
wpwraktuxbrain: i'm also a little afraid that, if i make it too long, nobody will read it ;-)13:08
tuxbrainalso why instead of a link to http://en.qi-hardware.com/wiki/Patent-free just c&p this as intro13:09
tuxbraina more agressive summary should be we piss of on pattent stuff, we want this fucking crap as far away from us as posible.13:11
wpwraktwo reasons: a) patent-free has a different focus. it describes what we do with OUR original ideas, while the patent policy describes what we do with idea others have removed from the public domain.13:11
wpwrakb) the patent-free page may get extended, which would make the thing unwieldy.13:12
tuxbrainok I think my sentence include both cases :)13:12
tuxbrainok again understand your point13:13
wpwrakafter all, the goal is to have guidelines people will actually follow. not a scholarly work on theory and practice of intellectual property legalese, to be read by lawyers only.13:13
wpwrak(aggressive summary) yeah. i'm thinking of finding a way to put "we prefer lawyers to be starved and too weak to fight, not bursting with confidence and eager to make the world quiver before their powers" somewhere.13:15
wpwrak(piss of on) it's just "piss on" :)13:15
wpwrakalas, that's not an option. well, if YOU decide not to worry about patents anymore, i'm sure you can be accommodated ;-)13:16
tuxbrainI hope one day we can have a lawyer in the community that can take care of this stuff, until now your guide is the best think I read about , so go for it13:16
wpwrakit may just be a somewhat hazardous decision13:16
tuxbrainI don't say I dont take care, I want them far away from me so it implies I take care of they doesn't being close :P13:17
tuxbraintime to time I still had nighmares of nanos stopped at customs, brrrrr, terrifing13:18
wpwrakah, "piss on" usually means to disregard. see definition 2 on http://www.urbandictionary.com/define.php?term=piss%20on13:20
wpwrakyeah, if patent holders are the nazis, customs are their gestapo13:20
wpwraktuxbrain: to sing along with, "we don't want no confiscation. we don't want no patent trolls. hey, sisvel, leave our bens alone !"13:26
wpwrakwell, s/sisvel/lawyers/ and we could use it fairly universally. naturally, violating pink floyd's copyright :)13:27
rafawpwrak: http://en.qi-hardware.com/wiki/Patent_Policy_Public_Face  :( .. it is going deeper and deeper ..13:27
rafawpwrak: I really do not like the whole thing. I need a while to read and read to understand well.. and still.. it is going deeper and deeper because just in few places around the world could have problems with that, and then the rest of us needs to follow some guidelines, ways, or whatever.13:29
wpwrakrafa: alas, the places where you have trouble are those where they project can get money. if the main market was south america, you'd see the same thing happen here. IP laws are fairly global :-(13:31
rafawpwrak: from time to time that all sucks for me and I feel that it is really a annoying thing which is not directly related with the rest of users/people. it is just one thing useful for qi and resellers.. which does both worlds hard to live together. I mean, I think that patents sucks of course, and if mp4/mp3.. whatever does13:31
wpwrakrafa: without qi-hw and resellers, there are no bens and no users. so it pretty much does affect them ...13:32
rafaproblems to nice projects like qi and family then I really would try to help.. but it is going so deep to me understand why I should follow some things to collaborate13:32
rafawpwrak: without qi and resellers there are no bens: I understand that.. but still, it is not my fault and that will not affect users at all surely :(13:33
wpwrakrafa: well, what can i say. these threats are not imagined. this sort of confiscation at customs has happened to openmoko.13:33
wpwrakrafa: (well, to one of their resellers - because openmoko had codecs on the device)13:34
wpwrakrafa: you could also view this as an opportunity - each time you avoid using a patent, you shrink the market for them :)13:36
wpwrakrafa: particularly if avoiding the patent means that you coe up with an alternative that achieves the same result. e.g., think of GIF. that patent was completely devalued by the avoidance of GIF files and the development of alternative and better file formats.13:37
rafawpwrak: :( yeah.. but I just do easy tasks just for fun.. and nobody is going to convert all his/her mp3 files to ogg.. still if here in this channel there 1 to 5 guys who did that.. the rest will not do that, just to give a bad example13:38
rafathere 1 to 5 = there are maybe 1 to 513:38
Action: tuxbrain has all his mp3 converted to ogg13:38
rafa1 guy so far :)13:38
rafawell, maybe more, just 1 who already confirmed it :)13:39
tuxbrainas easy as one order of a shell13:39
rafatuxbrain: yes of course, becuase all of the mp3 users have come surely from internet they will need to put that order into crontab or something similar13:40
rafabecause they will get more mp3 from time to time13:40
tuxbrainwpwrak: it's not copyright infringiment is a tribute :) I think I will record this one day heheeheh or at least my nightmares now will have a new backgroud music  "we don't want no confiscations... papara papapa papara we"13:45
tuxbrainc u later, guys, rafa, I see you a little low today, common!!! it is mostly a game, some of us has bet money others bet time, we are just finding strategies to win, all toguether, we are a team, users, developers, resellers and manufacturers don't give up in the boring part of the game, we will take care of it , just play the part you like also you are not tied to play if you don't find it fun, but be sure I will miss you :), btw send me the ogv pl13:49
tuxbrainayer before you leave :P13:49
kyakwpwrak: thanks for the hints; of course i have swap enabled on uSD, and i also doubt that it's drive performance that is a bottle neck13:50
kyakthough you are right, the speed of NAND could've been better13:51
wpwraktuxbrain_away: rafa had to go to the desert and he'll miss the nice barbecue we'll have at my place tomorrow. no surprise he's unhappy :)13:57
wpwraktuxbrain_away: (tribute) yeah ! ;-)13:57
tuxbrain_awayuff seems that finally we have found a good config of get rid of spam at drupal.... spammers (here are a lot of bad and ugly words that my daughter must not hear)15:23
wpwrakwolfgang: http://www.informationisbeautiful.net/2010/whos-suing-whom-in-the-telecoms-trade/15:30
wpwrakkyak: i'm not quite sure whether NAND or uSD is faster. by the bus width and bus type, NAND should win (8 vs. 4 bit, local bus vs. external bus). but then i've heard that uSD was faster. not sure how much overhead UBIFS adds.15:32
wpwrakrafa: (mp3/ogg recoding) how about a script that runs on the pc that uploads mp3s to the ben and recodes on the fly ?15:34
kyakaisa: there are a lot of packages not requiring patches, and during make only a notification is generated16:00
kyaki don't know why you decided so necessarily to get rid of it16:00
aisakyak: the principle I was using was to remove noise to make revelvant lines easier to see.16:04
aisathis turned out to be a two or three bugs.16:04
aisaone was an error/warning for a condition that is normal.16:05
aisatwo was the code being executed when it was not intended.16:05
aisathree is still not fixed, which is that the macro expansion is not protected from being expanded again by the shell.16:05
kyaknot following you here..16:06
kyakbugs where?16:06
aisaI'm not sure what you were asking about either.16:06
aisawhat does this: 'aisa: there are a lot of packages not requiring patches, and'16:06
aisarefer to?16:06
kyakare you Alan Post?16:07
aisaI am.16:07
kyakthen i was referring to your "no warning when there is no patches directory" patch16:08
aisaah, the "warning" in this case is an error message from 'ls'16:08
aisanot a warning that was intentional.16:08
aisain fact, the code tried to not have the error be there.16:08
aisabut it was written in such a way that the feature did not work.16:09
kyakcan you specify under which conditions you've seen such error message?16:09
aisaMy choice of wording re: warning was to state that the code still worked.16:09
aisamake package/lojban-wordlists/clean package/lojban-wordlists/compile V=9916:09
aisaNamely, when a package has no ./patches directory.16:10
kyakthen again16:10
kyakas i said above16:10
kyakthere are a lot packages without patches directory16:10
kyakand i never seen an error message16:10
aisaI can't speculate, you can look at the code and tell me ;-)16:11
aisaI'm not sure what you're driving at.  What are you intending for this conversation?16:11
aisaCan you tell me what you want?16:11
kyakjust trying to understand16:12
kyakwhy you removed that warning16:12
aisaHave you looked at the patch?16:13
wpwrakkyak: the code is quite clearly inconsistent in itself. whether it's actually necessary to worry about the directory being absent is a different question :)16:13
aisathe original code did not work.16:13
aisait wanted to test for the existence of the directory,16:13
kyakalways worked for me16:13
aisaand then if it existed, run ls ./patches | wc -l16:13
aisano matter the first condition,16:14
aisathe ls was always run.16:14
aisaso the patch corrects this,16:14
aisasuch that ls in only run if the first condition is true.16:14
aisaI promise this change works for you too.16:14
aisaand it works for me.16:14
wpwrakaisa: perhaps if there's a package that exhibits this, kyak can see for himself16:14
aisaI mentioned one earlier.16:14
aisathe command:16:14
kyakif there is patches directory, with the patches inside, it applies them. If there is no such directory, there is a warning.. That's how it worked16:15
aisamake package/lojban-wordlists/clean package/lojban-wordlists/compile V=9916:15
aisawill demonstrate the error,16:15
aisaif you revert or don't update tracking_backfire16:15
aisaif you want the warning,16:15
aisawe should add it on purpose.16:15
aisathe warning was a side effect of bad code.16:15
kyakbut it didn't throw an error or stop your build process?16:16
kyakbecause you said there was some error rather then warning16:17
aisaok: ls had an error.16:17
wpwrakkyak: if you look at the code, you'll see that it explicitly tests for the directory. this would only serve a purpose if the objective was to avoid the error/warning. otherwise, just run ls and it'll produce empty output as well16:17
aisathe makefile neither cares, checks for, nor detects that ls had an error.16:17
aisaso the process continues.16:17
aisakyak: I think this would all be much clearer if you looked at the code and we spoke about it.16:18
kyakok, i got your point16:18
aisarather than abstractly talking about what it may or may not be doing.16:18
kyakaisa: i see the code16:18
kyakit's ok :)16:19
aisaWhat about the code don't you understand?16:19
wpwrakaisa: that's a nasty question for this kind of rather subtle code ;-)16:19
aisaheh.  Indeed, I'm still working on one of the bugs, because there is a piece I don't understand.16:20
aisaWhat I'm really driving at, however,16:20
aisais if we all understand the code,16:20
aisaI need to know what the code isn't doing that we would like it to.16:20
wpwrakaisa: i think your interpretation (that leads to the fix of avoiding the error) is plausible. another possibility would be that someone wanted the error (as a warning) but didn't trust the ls ... wc -l to yield zero in this case. occam's razor favours the former interpretation :)16:23
aisaha!  Indeed, I'd much rather see a warning generated by make.  Though I'd rather see no warning at all because the condition is normal.16:25
wpwrakaisa: yes. don't warn if it's a quiet, normal day. no need for a Homeland Security Advisory System that's always at yellow :)16:26
wpwraklike someone wrote, "yellow is the new green" :)16:27
aisakyak: I'm sorry I'm so grumpy, I am working on too many things at once today :-(16:44
kyakaisa: no problem, i just misunderstood you16:50
aisaNext time I'm confusing I'll endeavor to not be impatient about it, too.16:51
kyakand next time i'm drunk i'll not show up on IRC (i always say that) :)16:52
aisawhat is your name?  Either your wiki username or commit name?16:54
kyakit's kyak16:55
aisaoh looksie, it only took me needing to look at the list.16:56
kyakyeah, the same name in the list16:57
aisais it pronounced ki-yahk, kuyook, or exactly like it is spelled?16:57
kyakehm, it used to be a random and unique combination of letters, no intended for reading out load :) but i guess it's the same as kayak, the boat16:58
qi-bot[commit] Juan64Bits: Connecting some FPGA's free GPIO http://qi-hw.com/p/xue/874a01d17:11
qi-bot[commit] carlos: Removing battery circuit http://qi-hw.com/p/nn-usb-fpga/9da569a18:52
wolfspraulaisa: that patch looks like your first patch, no?20:22
wolfsprauldid you see Werner's final version here in irc?20:22
wolfspraulaisa: ah sorry, wrong! it is the new version, just not with the last ideas werner had around -z or so20:23
wolfspraulone sec let me find it in the logs20:23
aisaah, correct.  I haven't incorporated that yet.20:23
aisaI got stuck on the idea of getting the quoting right,20:23
aisaand I have been doing non-computer work most of the day.20:23
wolfspraulaisa: @if [ -d "$(2)" ] && [ "`ls $(2)`" ]; then \20:24
aisa(I'm walking away again for a bit, I'll be around variably.20:24
aisaI think that is the current version, missing the wc -l.20:24
wolfspraulI think that was Werner's last idea20:24
wolfspraulit's all fine now, we don't need to waste time over every last quote :-)20:25
wpwrakmy last idea is a use i also only spotted today. shell programming is full of surprises ;-)20:36
wpwrakwolfspraul: btw, two quick PR ideas: 1) next 1000 units, with the new stuff that's getting bundled. it's low-profile, but someone may pick it up.23:05
wpwrakwolfspraul: 2) when jlime is ready for inclusion, plus dual-boot is there as well, then you can start shipping bens with both options. (well, after passing a zen moment of reflashing :) that'll also be a newsworthy moment.23:06
wpwrak(that is, ben shipping with jlime dual-boot. not necessarily your reflash session ;-)23:07
wolfspraulyes I agree but this is all a lot of work23:10
wolfspraulI think I will focus on the Milkymist launch at 27C3 next23:10
wpwrakworthwhile work, i'd hope :)23:10
wpwrakwhy not try to make something out of the 2nd 1kunits ? see how it goes23:11
wpwrakyou may find things that help with the mm launch as well23:11
wolfspraulI've started to write down random thoughts on the milkymist release23:11
wolfspraulhaven't even talked with Sebastien yet, nothing can go in there that he doesn't agree with23:12
wolfspraulit takes time to craft a strong message23:12
wolfspraulI have started to contact Lattice about a quote, I think that would be nice23:12
wolfspraulbut I need to ask 5-10 people for a quote, it always depends on whom you talk to, whether they 'get it', see mutual benefit, etc. etc.23:12
wpwrakwell, you still have a bit of time for that. so that;s probably safe23:13
wolfspraul2 months23:13
wolfspraulnot much23:13
wolfspraulgetting a quote is hard work :-)23:13
wolfspraulthere is a PR department on the other side23:13
wolfspraulI remember one time it took 6 months, of constant back and forth, to get a quote from Sony-Ericsson for a press release23:13
wpwrakthey'll also want to find out who you are. they wouldn't want to support the local taliban-nazi-pedophile club, for example ;-)23:13
wolfspraulso we see23:14
wolfspraulbut I appreciate your comment23:14
wolfspraulyes, if we had a better/more efficient news machine, a press release about the first 1k being sold, and the second 1k being manufactured is not bad23:14
wolfspraulthere is bigger crap that gets broadcast23:14
wolfsprauland another one about Jlime/OpenWrt dual-booting is also not bad23:15
wpwraknot sure if you even want to define a "target market"23:15
wolfspraulwell it's full of question marks23:15
wolfspraulI haven't even sent my 1k manufacturing report email yet23:15
wolfspraulyet alone a fully crafted press release23:15
wolfspraullet alone23:15
wpwraki'd say, if you think you can do one about the 2nd k without sacrificing anything more important, then do it23:15
wpwraksee, so it's still news ;-)23:16
wpwraksometimes, procrastination works in your favour :)23:16
wolfspraulandres-calderon: hey there, good to see you here! :-)23:16
wpwrakanyway, i'll be afk for a bit23:16
wolfspraulwpwrak: check this http://en.qi-hardware.com/wiki/Xue/layout23:17
wolfspraulthey are working...23:17
andres-calderonHola wpwrak23:25
wolfspraulandres-calderon: he just went out a bit :-)23:40
andres-calderon  I think I will use this type of connector : http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=S9044-ND23:48
andres-calderon50 pins, pitch 2mm23:49
andres-calderonThey seem appropriate?23:49
wolfspraulI heard 100mil (2.54mm) is more common than 2mm23:50
wolfspraulthis connector is for what? expansion header?23:50
andres-calderonyes, GPIOs23:50
wolfsprauloh I see, 2 rows of 2523:51
andres-calderon2.54mm x 25 pos = 63.5mm23:51
wolfspraulmaybe we should then include a cable that has that connector on one side, and individual wires coming out on the other side23:51
wolfspraulsure, I agree23:51
wolfspraulsmaller = better, imo23:52
wolfspraulthere are 1.25mm pitch connectors too, I think23:52
wolfspraulI don't know much about the pros or cons for the intended use case later, though23:52
wolfspraulmy feeling is whether we have a 2mm or other connector, we should include a cable so that people can easily solder individual wires to whatever target they have23:53
andres-calderonWe also use two 2.54mm connectors (instead of one of 2mm)23:53
wolfsprauland in that case (if we include a cable anyway), I would think it doesn't matter so much whether the connector on the board is easy to get, or not so easy to get23:53
wolfspraulas long as at least going forward the connector we chose is not becoming obsolete23:53
wolfspraulthat would be bad23:53
wolfspraultoo bad that wpwrak is out right now...23:54
wolfspraulandres-calderon: if you are out later, I am sure wpwrak will have some feedback on this, you can check the irclogs at http://en.qi-hardware.com/irclogs23:54
--- Sun Oct 10 201000:00

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