<wolfspraul>
wpwrak: I had one additional idea, since we have all the data already, we should also make one PDF that includes all schematics pages in one file, for each commit (I mostly see the need only for the 'latest' full schematics PDF, but since all the data is there we might as well create it for each commit)
<SiENcE>
@ bartbes, well i fixed 2 problems, now i have a nlove binary :)
<SiENcE>
bartbes : but nlove-0.0.2/modules/image/sdl/Image.cpp:36: error: 'IMG_INIT_JPGÂ Â <- is indeed an error
<wpwrak>
wolfspraul: look at the result page, top line, above the commits :)
<wolfspraul>
ok will do :-)
<wpwrak>
i dind't generate a full PDF set at each commit, though.
<wolfspraul>
wpwrak: the scripts are holding up nicely, not checking output yet though
<wolfspraul>
I have added blinkenlights, milkymist jtag/serial cable, SIE
<wolfspraul>
with the AVT2 board it runs into an error
<wolfspraul>
"cannot open function_block.sch"
<wolfspraul>
I think we may have committed some things in a step by step way that creates this broken dependency in a particular commit
<wpwrak>
eek. what's the project called ?
<wolfspraul>
it's in the board-qi-avt2 project, main schematic is main/Qi_AVT2.sch
<wpwrak>
let's see what happens there ...
<wolfspraul>
we may have simply committed broken stuff at some point in the history
<wolfspraul>
how about that -S sanitize option, can't that just always be on?
<wpwrak>
yes, -S can always be on. if there's no problem, then it does nothing
<wolfspraul>
not sure whether it would help here, but it did help me with some missing library dependencies
<wpwrak>
but i'm not sure if it catches missing sheets
<wpwrak>
yes, it does track libraries
<wolfspraul>
if -S is safe to always be on, can't you just remove the option and leave the code always running?
<wpwrak>
sometimes, you may want to see the errors
<wpwrak>
(Failed to open ...) i get it, too
<wpwrak>
and -S doesn't help, as expected
<wolfspraul>
btw, the eeschema segfault came from not having an X server, and even when I had one the missing DISPLAY variable will segfault eeschema
<wolfspraul>
a bit rough, it should handle missing DISPLAY or missing X server more gracefully
<wpwrak>
aah ! yes, it doesn't like that at all. how did you solve it ?
<wolfspraul>
I have now installed a permanently running VNC session on the buildhost
<wpwrak>
(rough) well, the only case in which it makes sense to run eeschema without display would be my --plot patch :)
<wpwrak>
perfect ;-)
<wolfspraul>
yes but a segfault is not nice
<wpwrak>
and the problem is indeed that function_block.sch wasn't committed. guess we need a sanitize-schem, too
<wpwrak>
should be fixed RSN ...
<wolfspraul>
take your time I need a long time anyway to hook everything up etc.
<wpwrak>
whoa. lots of libraries it can't find :-(
<wpwrak>
some of the revisions won't look pretty ...
<wolfspraul>
yeah I can imagine
<wolfspraul>
this was our first KiCad projet
<wolfspraul>
but for me it's also important that the schhist scripts are as robust as possible, and always generate something
<wolfspraul>
an ugly diff is better than no diff
<wolfspraul>
because people will look at it and then work with cleaner commits
<wpwrak>
yes, it's important that the scripts can handle some mistakes. missed commits are unfortunately all too common
<wpwrak>
the AVT2 revisions do actually loop pretty good. now let's see what happened with those problems ...
<wpwrak>
also some mystery diffs - in the FETs, the font style changed. my detection algorithm finds this, but it's invisible in the "for viewing" images. (you can figure it out with the before/after PDFs)
<wpwrak>
heh, and the line width for capacitors :) you need to zoom into the pdf to see *that* ;-)
<wpwrak>
(missing libs) the bug is in sanitize-profile. yet another $1 vs. $2
<wpwrak>
interestingly, eeschema didn't mind. probably because also Qi_AVT2.cache.lib is in git, so it never had to actually look for these items.
<qi-bot>
[commit] Werner Almesberger: New script sanitize-schem to remove subsheets for which no file can be found. http://qi-hw.com/p/eda-tools/75a4ea8
<qi-bot>
[commit] Werner Almesberger: sanitize-profile accidently ignored LibDir, causing many incorrect "corrections" http://qi-hw.com/p/eda-tools/d4405e4
<wpwrak>
much better now
<wolfspraul>
wpwrak: yes it runs through now, great! congrats!
<wolfspraul>
I need to do more plumbing, make it all robust and automatic. will probably continue tomorrow.
<wpwrak>
good. so you locate the projects automatically or do you just keep a list/Makefile of them ?
<wpwrak>
s/so/do/
<wolfspraul>
locate?
<wolfspraul>
I try to automate it as much as possible, but also I want to get it up and running asap
<wolfspraul>
in the end even hooking up a new project to the commit bot is manual right now
<wolfspraul>
hooking it up to the kicad schhist is also manual
<wolfspraul>
plus, like in the case of ben-wpan, there can be multiple schhist histories per project
<wolfspraul>
that's just fine right now, no need to automate this as the amount of work to automate it would not be recouped any time soon
<wolfspraul>
even if we expect 100 more KiCad project in the next year
<wolfspraul>
100 * 2 minutes = 200 minutes
<wolfspraul>
not worth automating further
<wolfspraul>
I'm writing a little script and cronjob on the buildhost, plus a little plumbing here and there
<wolfspraul>
will all be documented as usual
<wolfspraul>
once it's setup, adding another KiCad project to schhist will be < 2 minutes
<wolfspraul>
that's the plan
<wolfspraul>
makes sense?
<wpwrak>
yup, sounds good. when you add projects manually, you can also run a quick check to see if everything is well.
<wolfspraul>
sure
<wolfspraul>
I know the real costs of automation :-)
<B_Lizzard>
Is there source code available for libfitz etc?
<B_Lizzard>
Ah, it's part of mupdf
<B_Lizzard>
My bad
<SiENcE>
@ bartbes, after some work i can successfully run love2d files on dingoo. everytime i have to adjust the screensizes and disable mouse, sound.
<beegit>
hard drive not getting identified on comp?
<kristianpaul>
?
<SiENcE>
hey
<bartbes>
SiENcE: have you tried npong and urandom__'s snake?
<SiENcE>
i ask because some lua functions did not worked
<SiENcE>
but maybe this has todo with old demos
<SiENcE>
love.graphics.getWidth() did not worked
<SiENcE>
and i get an segfault when starting nlove alone
<SiENcE>
and i have to disable the sound module
<SiENcE>
bartbes, Pong works
<SiENcE>
Snake also works
<SiENcE>
:)
<Greek_o_nikos>
Hello.
<Greek_o_nikos>
Is it easy to transfer files from one's computer to the Ben Nanonote?
<Greek_o_nikos>
(what I'm first thinking about are simple text files)
<nebajoth>
yes
<nebajoth>
well
<nebajoth>
are you familiar with scp?
<Greek_o_nikos>
scp, could it be something I found in a ssh tutorial, I'm not sure. (to transfer files)
<Greek_o_nikos>
well, I can easily regain some basic knowledge about how to use scp, I just found one of the tutorials available.
<Greek_o_nikos>
I've used it just as it was stated in one tutorial I tried in the past and I successfully copied files from and to another networked computer.
<mth>
Greek_o_nikos: does your computer run Windows or Linux or something else?
<Greek_o_nikos>
mth: gNewSense on the Lemote Yeeloong.
<mth>
with GNOME or KDE?
<Greek_o_nikos>
GNOME.
<mth>
in either case, you can probably use sftp:// URLs in the file manager
<mth>
I know it works in KDE, but afaik GNOME has the same feature
<Greek_o_nikos>
mth: but first I have to setup an ftp server, I think if I want that method?
<nebajoth>
if you're using a windows desktop, grab winscp
<nebajoth>
ftp-clientish
<nebajoth>
uses scp
<nebajoth>
drag and drop files to your ben
<nebajoth>
if you are using gnome, make an SSH bookmark
<nebajoth>
open with nautilus
<nebajoth>
drag and drop files to your ben
<mth>
Greek_o_nikos: not needed: SFTP is part of the SSH protocol
<Greek_o_nikos>
mth: ok. I didn't know.
<nebajoth>
oh, I see
<nebajoth>
gnome
<mth>
at least with OpenSSH it's available by default, I'm not sure if the NanoNote uses OpenSSH
<nebajoth>
Places -> Connect to Server ->
<nebajoth>
enter the nanonote ip address
<nebajoth>
opens in nautilus
<nebajoth>
save as a bookmark while you're in thar
<Greek_o_nikos>
While service type: Public FTP?
<Greek_o_nikos>
There's also an option for SSH there.
<Greek_o_nikos>
(on service type)
<Greek_o_nikos>
I live in Greece. When I got the Lemote Yeeloong it was shipped from the Netherlands. I'll check now where I can get a Ben Nanonote device.
<nebajoth>
ssh
<Greek_o_nikos>
Is there any image of the ben nanonote which shows what the usb port looks like?
<Greek_o_nikos>
I found it by searching: mini-USB 2.0
<kristianpaul>
:D
<kristianpaul>
what for VNC server is fidelio?
<wolfspraul>
kristianpaul: I installed it as part of the schhist setup
<wolfspraul>
that's live from now on, any KiCad project on projects.qi-hardware.com can be hooked into it, and every commit will automatically generate an update to the visual schematics history!
<kristianpaul>
great
<kristianpaul>
oh even avt2 !
<wolfspraul>
of course
<wolfspraul>
from now on we will hook every KiCad project into this
<wolfspraul>
thanks a lot for your excellent work on this, I only did the plumbing!
<wolfspraul>
I will try to make sure the scripts keep working, next time someone does a KiCad commit we should watch out
<aisa>
I've just booted my NanoNote into USB mode for the first time, to update u-boot.
<aisa>
Things seem to be fine now, but how do I shut the machine down?
<aisa>
it seems I had to unplug the device and take my battery out.
<aisa>
I'd love a better suggestion when anyone passes by and reads this.
<wolfspraul>
aisa: you can try the hardware reset pin, the little hole on the bottom of the device
<aisa>
Indeed, I will next time.
<wolfspraul>
I rarely (never) use it, but I think it's the equivalent of unplugging USB cable and removing the battery
<aisa>
Is that how one exits USB BOOT mode?
<wolfspraul>
exit in which sense?
<aisa>
turn the machine off, or reboot it.
<wolfspraul>
USB boot mode is a state the CPU is in, it waits for commands over USB
<aisa>
what do I do when I'm done issuing commands over USB and want my screen back?  :-)
<wolfspraul>
so in that sense 'usb boot mode' is already exited when you run usbboot, especially when usbboot loads and executes binaries in the NanoNote RAM
<wolfspraul>
I unplug the USB cable, replug it
<wolfspraul>
you can also try the hardware reset button
<aisa>
awesome, thank you wolfspraul!
<aisa>
I'm trying to update my NanoNote using MicroSD.
<aisa>
I believe I've copied the uImage, root.ubi, and untarred the rootfs on an ext2 MicroSD card.
<aisa>
I boot by using Power+M, and I get the following error:
<aisa>
** File not found /boot/uImage
<aisa>
Wrong Image Format for bootm command
<aisa>
ERROR: can't get kernel image!
<aisa>
I think I should believe the first message,
<wolfspraul>
how big is your microSD card?
<aisa>
and that "Wrong Image Format" is essentially saying "zero sized file.
<aisa>
16gb, I believe.
<wolfspraul>
there is a bug in u-boot's ext2 handling
<wolfspraul>
yes, won't work
<wolfspraul>
filesystem bug in u-boot's ext2
<aisa>
dang.  how small should I be to avoid this bug?
<wolfspraul>
xiangfu is on it, but we have many bugs and unfortunately it also has been hard for us so far to get patches back upstream...
<aisa>
would reformatting and copying uImage early in the process work,
<wolfspraul>
hmm
<aisa>
or do I need to get a smaller microSD card?
<wolfspraul>
I think cards up to 4 gb will always work
<aisa>
good, ok.  Time to go purchase a new card :-)
<wolfspraul>
you can also try to make a small partition on your 16gb card
<aisa>
that is a good idea.
<wolfspraul>
try a small partition first
<wolfspraul>
at the beginning
<wolfspraul>
2 gb
<aisa>
Worth trying, it will be mucch faster
<wolfspraul>
this is a bug in the ext2 filesystem codes of u-boot
<wolfspraul>
we just haven't gotten aroudn to tracking it down and fixing it yet
<wolfspraul>
they even have 32 gb microSD cards now, but I was too lazy/stingy to buy one and test :-)
<aisa>
Indeed.  I've been sitting on flashing my NanoNote for way too long.
<aisa>
I needed like several hours to sit and work on it,
<aisa>
and finally made time tonight.
<aisa>
I'm really looking forward to getting this process done,
<aisa>
because I want to be faster at it so I can go find some of these problems.
<wolfspraul>
well thanks for your patience
<wolfspraul>
oh you can help us with that kind of low level bugs?
<aisa>
It's all fun.  :-)
<wolfspraul>
that would be awesome!
<aisa>
Indeed, I'm a long way from being an expert,
<aisa>
but I generally understand how the kernel runs,
<wolfspraul>
u-boot is a bit troublesome, we have been unsuccessful in getting some patches upstream, and quickly fixing some bugs such as this one
<aisa>
have a really strong notion of how libc works,
<wolfspraul>
well try the 2gb partition idea first, let's see whether it helps you
<aisa>
and know almost nothing about MIPS.
<wolfspraul>
it may also have to do with block sizes or so and then maybe the bug will always show on a large card, even if you have a small partition on it
<aisa>
got it.  I'll find out.  Is there an open ticket I should append information to?
<aisa>
I'm updating the wiki.
<aisa>
I'm about to stop working on it tonight, but will come back to it.