#qi-hardware IRC log for Sunday, 2010-05-02

qi-commitsMirko Vogt: Merge branch 'xburst' of git@projects.qi-hardware.com:openwrt-xburst into xburst http://qi-hw.com/p/openwrt-xburst/5d5a4c005:59
qi-commitsMirko Vogt: use qi repository for packages http://qi-hw.com/p/openwrt-xburst/50c305905:59
qi-commitsMirko Vogt: enable kernel inotify support for lb60 target http://qi-hw.com/p/openwrt-xburst/c1f7c6605:59
qi-commitsMirko Vogt: build script should make sure etc/files exists http://qi-hw.com/p/openwrt-xburst/fb94b6007:02
qi-commitslars: uboot-xburst: Fix lcd display init sequence. http://qi-hw.com/p/openwrt-xburst/34dab4607:05
qi-commitsLars-Peter Clausen: Merge commit 'backfire/master' into xburst http://qi-hw.com/p/openwrt-xburst/78fcab607:25
qi-commitsnico: [backfire] merge r21282 (closes: #6918) http://qi-hw.com/p/openwrt-xburst/04e95f907:25
qi-commitsjow: [backfire] merge r21288 http://qi-hw.com/p/openwrt-xburst/e49cd5307:25
qi-commitsjow: [backfire] merge r21280 http://qi-hw.com/p/openwrt-xburst/4b7df9807:25
qi-commitsnbd: [backfire] backport ssb updates from r21269, this is required for future mac80211 updates http://qi-hw.com/p/openwrt-xburst/9631bda07:25
qi-commitsnbd: [backfire] merge vsc73x5-ucode download patch from r21274 http://qi-hw.com/p/openwrt-xburst/5970ed407:25
qi-commitsjow: [backfire] merge r21270 http://qi-hw.com/p/openwrt-xburst/a9553d907:25
qi-commitsjuhosg: backfire: ar71xx: create separate 'factory' and 'sysupgrade' images for the Planex boards (backport of r21175) http://qi-hw.com/p/openwrt-xburst/24ddbc007:25
qi-commitsjuhosg: backfire: tools/firmware-utils: add size check to the mkplanexfw tool (backport of 21174) http://qi-hw.com/p/openwrt-xburst/3bed6fa07:25
qi-commitsjuhosg: backfire: ar71xx: remove config partition from the mtd map on the M07:25
vegyraupetuxbrain_away: ping08:01
mirkoqi-commits, welcome back ;)08:08
larsci killed him :/08:08
mthmaybe it can be configured to only post the merge commit itself and not the commits that are being merged?08:10
qi-commitsLars-Peter Clausen: pass strings by reference where possible http://qi-hw.com/p/gmenu2x/40a26e108:40
qi-commitsLars-Peter Clausen: Cache current icon surface. By doing so we safe a huge number of hash map http://qi-hw.com/p/gmenu2x/a794a1c08:40
qi-commitsLars-Peter Clausen: Kill float and double http://qi-hw.com/p/gmenu2x/667fb0508:40
qi-commitsLars-Peter Clausen: Don't do this. It's pure evil. http://qi-hw.com/p/gmenu2x/8a87e8e08:40
qi-commitsLars-Peter Clausen: Fix pxml.cpp compile http://qi-hw.com/p/gmenu2x/a107dd808:40
wolfspraullarsc: yes you did :-)09:04
wolfspraulI checked the server and it looked like spamd sucked up all available memory09:05
wolfspraul2 gb or so09:05
wolfsprauluntil eventually other processes died from side-effects09:05
wolfspraullike the irc_cat09:05
wolfspraulmth: maybe it could be improved but right now it's not very high on my todo list. I think it works OK right now.09:06
wolfspraullarsc: which of the Hanvon e-books are you most interested in nowadays?09:06
wolfspraulthe reason I am asking is that I had quite some fun with the C3100 MTK phone for Harald, taking the PCB apart09:07
wolfspraulsee http://en.qi-hardware.com/wiki/CECT_C3100_Analysis09:07
wolfspraulcost me only about 100 USD for the service09:07
wolfspraulso I am thinking about buying some used Hanvon e-books on taobao, then do the same thing which maybe helps us follow a few traces and decode a few pins09:08
wolfspraulfor those C3100 pics they needed 5 boards, but only because I asked them to return each individual layer to me. I think they can do it with 2 boards.09:08
wolfspraulso the total damage would only be 2*used Hanvon device + 100-200 USD + a few days09:08
wolfspraulbut I don't know which one you guys focus on nowadays - the old N516, or newer N526 or something newer still?09:09
larscwolfspraul: well we mostly know how things are connected, so i'm not sure if it would be worth the effort09:22
larsconlything still unkown is the protocol of the ts controller09:24
larsc...but i'm getting a logic anlayser soon09:24
mirkolarsc, the merge screwed up some stuff09:30
wolfspraullarsc: not sure this got through, chinese govt is erratically fighting openvpn traffic now...09:31
wolfspraulwhich of the hanvon models are you or openinkpot most interested in nowadays?09:31
mirkolarsc, busybox patches do not apply anymore correctly09:31
mirkoi wonder how this can happen after a merge09:31
wolfspraulstill the 516, or newer ones?09:31
larscwell, the main focus is still 516.09:36
larscbut mostly all the hardware connections are known09:36
larscso i don't know if it would be worth the effort to take the pcb apart09:37
wolfspraulsure I got that, no worries09:38
wolfspraulfor the future it's still good for me to know I can take pics to follow any trace in the PCB for 200 USD09:41
wolfspraulcould speed up a lot of things09:41
wolfspraulplus I also need to go back to Hanvon one day09:41
qi-commitsLars-Peter Clausen: reset to backfire. http://qi-hw.com/p/openwrt-xburst/99c6c7309:42
vegyraupetuxbrain_away: ping09:53
zoltanh7211Hi guys10:55
zoltanh7211I have several questions about nanonote10:56
zoltanh7211Who could help me10:56
gnufshow can we say if we can without knowing the question?10:59
zoltanh7211Oh, sorry my net provider has long lag11:00
zoltanh7211My question is is it possible to expand this PCB to an table device?11:01
zoltanh7211Bigger touchscreen, and bigger memory?11:01
zoltanh7211The target what we would like to see is an multimedia capable table deviice maybe with wifi module, there is an copyleft solution?11:05
jxselfzoltanh7211: If you're asking if the NanoNote could be endowed with a larger screen & more memory, sure. Go for it.11:11
zoltanh7211jxself: Our group would like to redesign the nanonote, keep as much hw as possible and run as os an rebulded Fedora - precisely -  with 512 Mbyte RAM, and bigger (9 col) screen with touchpad.11:26
qi-commitsLars-Peter Clausen: Use a simple array instead of a hashmap for the colors. http://qi-hw.com/p/gmenu2x/7fd1ab013:54
qi-commitsLars-Peter Clausen: iconbutton: Cache icon surface instead of doing serveral hashtable lookups per http://qi-hw.com/p/gmenu2x/28d3fc913:54
qi-commitsLars-Peter Clausen: Fix topbar icon rects http://qi-hw.com/p/gmenu2x/475fdb714:18
qi-commitsLars-Peter Clausen: remove redundant assignments http://qi-hw.com/p/gmenu2x/d4490ba14:18
zearlarsc, i see you're fixing gmenu :)14:33
zearsay, is the path length for file browser in gmenu fixed yet?14:33
zearif not, let me tell you what lines of code need fixing14:33
larsci don't think that it's fixed14:36
zearit's in a few places in the code14:36
zearyou'll have to look for the rest of them, but14:36
zeardirdialog.cpp, line 18414:36
zearyou should replace 11 with the length of the path in " "14:37
zearin our case it's 5 or 4 (i don't know if it counts from 0 or not)14:37
zeari forgot to change it from the dingoo version, where it pointed to /usr/local14:38
zearif the length of the path is wrong, you won't be able to go up the dirs in browser14:38
zearalso, this might be a good opportunity to change the path to "/" and see how well gmenu2x works with the nand filesystem14:39
zearlarsc, you copy?14:48
larscis there any reason to set it not to /15:25
zearlarsc, it was designed to run on an fat filesystem sd card15:50
zeari'm not sure if it can handle symlink loops15:50
zearand i bet there are some in /dev/ or /proc/15:50
zearit might scan the filesystem for ever15:50
zearadditionally, it will show the capacity of the nand instead of sd card on the gui bar at the bottom15:51
zeareventually, we can set the automatic game/app scanner to look only in a selected dir, let's say /card15:54
zearoh and larsc, why do you use the pandora's theme? it's designed to run in 800x480 resolution, hence the huge icons and wrongly positioned wallpaper15:57
tuxbraingood night from spain nanonoters15:58
zeartuxbrain, you mean good night, or good evening? :)15:58
tuxbrainwell it's 22:00 here15:59
zearyeah, but "good night" is when you leave15:59
tuxbrainwhatever :)15:59
zeari think "good evening" is the proper greeting until the morning15:59
tuxbrainand going to sleep early is proper for health and here we are :)16:00
tuxbrainI see a lot of commits this two days :) , any release date scheduled?16:02
zeartuxbrain, larsc is improving/fixing gmenu2x, that's probably gonna go for the next release16:03
tuxbrainalso I want to confirm that with the lastest uboot patch from xianfu, tp4,tp5 used as UART0 has no interference with keyboard and work like a charm :)16:04
larschm, looks like upstream gmenu2x is dead16:34
tuxbrainlars:time for a fork?16:40
larscit is already16:40
larscbut i was looking forwared to sending our patches upstream16:41
larscbut nothing was commited in their svn during the last year16:41
tuxbrainI think we will use gmenu2x in different way than is pretended, we have a full distro than some in next future will full managed by opkg16:42
tuxbraingmenu2x was intended to manage individual ports16:42
tuxbrainand mess with overclocking16:43
tuxbraintwo characteristics we really don't need16:43
larscthe overclocking feature was one of the causes for screen flicker16:45
larschm and their bugtracker is full with spam16:52
tuxbrainlasrc: If I have time I will like to work with the "easy" part of pimping gmenu2x, (theme/icons/shortcuts...)17:24
qi-commitsMirko Vogt: append to logfile, do not overwrite... http://qi-hw.com/p/openwrt-xburst/b3e52c117:40
qi-commitsMirko Vogt: evaluate error messages (this way e.g. newlines can be passed) http://qi-hw.com/p/openwrt-xburst/c8b4f9e17:40
qi-commitsMirko Vogt: as usbboot often exists with exit code 0 (success), even if an major error occured, we have to handle things differently... http://qi-hw.com/p/openwrt-xburst/76a779f17:40
qi-commitsMirko Vogt: wget should append to logfile, do not overwrite it http://qi-hw.com/p/openwrt-xburst/1c0e1e417:40
qi-commitsMirko Vogt: reformatting error messages, adding hints about latest versions http://qi-hw.com/p/openwrt-xburst/601659517:40
qi-commitsMirko Vogt: function <log> should be defined before being called http://qi-hw.com/p/openwrt-xburst/f2f946317:40
mirko__a bit hacky, however it works17:43
tuxbrainmirko__:what are you talking about?17:44
mirko__usbboot does exit with SUCCESS even when flashing failed17:45
tuxbrainok :) big league coding :P17:46
qi-commitsMirko Vogt: move /etc/opkg/opkg.conf to /etc/opkg.conf as otherwise we're not overriding the default configuration http://qi-hw.com/p/openwrt-xburst/f9b833c17:58
qi-commitsMirko Vogt: forgot to delete the obsolete opkg config directory http://qi-hw.com/p/openwrt-xburst/cd2cd8d18:00
tuxbrain_awayc u tomorrow guys, happy hacking, coding, learning, and some sleeping :P18:25
--- Mon May 3 201000:00

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