<lekernel>
got an idea about the status of the elphel theora encoder?
<lekernel>
kristianpaul: no, it's not moving. so no update except that i'll remove all fedora references from the milkymist material so people stop asking me this question.
<wolfspraul>
lekernel: I think it is only partially implemented
<wolfspraul>
and it has not been moved to newer camera models since a number of years
<wolfspraul>
and Andrey (the author) has a reputation of writing extremely hard to read and undocumented Verilog code
<lekernel>
I see ...
<wolfspraul>
there might be a jewel there though, I don't know
<wolfspraul>
a jewel covered in some mud maybe :-)
<wolfspraul>
the sources must be online somewhere...
<lekernel>
well I checked out some code yesterday from elphel which pretty much matches this description
<lekernel>
the weirdest point is that there's no test bench, but there sure has been one at some point. you can't write and debug a complete theora encoder by just downloading to fpga and not doing simulations...
<lekernel>
they just did not publish it apparently
<lekernel>
hi xiangfu
<xiangfu>
lekernel: hi
<lekernel>
what's up?
<wolfspraul>
lekernel: that may be resolvable if you just contac them, I suggest the #elphel channel on freenode
<xiangfu>
lekernel: fine. I am work on nanonote last week :)
<lekernel>
did that, no answer. same on their mailing list (actually they even moderate everything including new subscriptions, so my message didn't even go through yet)
<lekernel>
if you're subscribed to their list, maybe you can post my message so it goes a bit faster ... I don't know how reactive they are, but not a lot apparently
<wolfspraul>
what is your exact question?
<wolfspraul>
the #elphel channel should definitely work, they read the backlog too
<wolfspraul>
you can also ask olga there if andrey is somehow not seeing it, then she will
<lekernel>
let's start with getting the test benches for this code :-)
<wolfspraul>
do you have the url of the code you looked at?
<lekernel>
yeah... tbh I don't hold my breath regarding this code
<lekernel>
I'm afraid the solution will be "roll your own", as often
<wolfspraul>
lekernel: as for responsiveness, they are in Utah, that's an unusual time slot for me and you probably too unless you have a lot of people in the US you chat with
<wolfspraul>
so once you know the Utah thing, they are quite responsive, or used to be
<wolfspraul>
right now I'm pretty sure they are offline though, for example
<wolfspraul>
next chance, let me see. Utah Monday morning - still a long way out :-)
<wolfspraul>
kristianpaul: what milkymist-related things were you waiting for in Fedora 15?
<lekernel>
well, they were supposed to add the lm32-rtems toolchain for example, but it takes ages (even though most of the packaging work was already done by the rtems project)
<lekernel>
FEL is really dead I'm afraid
<lekernel>
and yeah, I did ping them, many times, just to hear a new excuse every time. now I just give up.
<lekernel>
if they wake up one day, good. but enough on my side.
<wolfspraul>
lekernel: Chitlesh told me he basically wanted to fold FEL into the regular Fedora releases
<wolfspraul>
if I understood that correctly he did not want the overhead and pressure of the FEL release cycles anymore
<wolfspraul>
just make it a group in Fedora, and then let Fedora do the rest. so yes, you could say 'FEL is dead', but from his perspective it's more about maintaining a group instead of a spin.
<lekernel>
no matter how they set it up, things do not get done
<wolfspraul>
also his main focus was the usability of these tools on the paid RedHat Linux, because that's what is used at ON
<lekernel>
even recompiling the flterm package with the new source code (with the gdb support) which should take what? 5 minutes?
<lekernel>
they didn't do
<wolfspraul>
what is the recommended devel platform now? what are you using?
<lekernel>
just run xiangfu's build script on any platform :-P
<lekernel>
anyway fedora would only solve part of the problem
<lekernel>
with the ~12 libraries (and counting) compiled for LM32 that Flickernoise needs, it's getting more and more difficult to get an "out of the box" development environment into a distro
<lekernel>
those sure can be packaged as well, but it would require a certain effort
<kristianpaul>
wolfspraul: (fedora 15) urjtag from upstream, and rtems toolchain already included :-), may be llhdl as a demo?..
<kristianpaul>
ah ok lekernel ;)
<lekernel>
huh? they added the rtems toolchain ?!
<lekernel>
wow ...
<kristianpaul>
no no, i'm asking if?..
<kristianpaul>
i'm telling wolfspraul what i want for fedora next release related to mm1
<kristianpaul>
but seems scripts are more universal those days :-)
<lekernel>
well, tell the fedora people, not wolfspraul :-)
<lekernel>
but tbh if you want this to move, you probably have to do all the packaging yourself then keep bothering people until it is merged
<lekernel>
which I won't do
<kristianpaul>
me either ;)
<lekernel>
there are already more than enough people on my followup list, which is quite tiring for me and distracts me from coding. so I just keep the most important.
<lekernel>
scripts aren't more universal, they just require less emailing around
<kristianpaul>
got it !!, i must go, 1 of may event coming.. :-)
<lekernel>
haha
<kristianpaul>
heh ;)
<lekernel>
it seems there is a huge riot in Berlin every 1st may night. will be my first today ...
<wolfspraul>
yes Berlin culture
<wolfspraul>
maybe it's similar to some riots in suburbs of French cities? from the news I thought it sounded the same...
<wolfspraul>
you should be (slightly) careful, especially since it's your first year
<wolfspraul>
but I think it's quite localized and ritualized :-)
<wolfspraul>
maybe with your long hair if you are in the wrong place at the wrong time, some police may think you are one of the bad guys :-)
<lekernel>
yeah, usually it doesn't happen in my area
<wolfspraul>
unlikely though, no worries...
<lekernel>
it's more in the east (kreuzberg ...)
<lekernel>
from what I've seen the French riots looked softer :p
<mumptai>
aww, just don't dress like the involved parties, and look for somthing to take cover behind ;)
<GraveDigger>
hi folks
<GraveDigger>
is Sebastien B. around here?
<lekernel>
yes
<GraveDigger>
ohai
<GraveDigger>
regarding the rs232 uart, i didnt wanna start a huge discussion or such, it was just a simple question
<GraveDigger>
like
<GraveDigger>
was it supposed to be this way?
<lekernel>
no. but i'm not even sure the latest version does it. i rarely update the opencores stuff
<lekernel>
mwalle did some changes to the uart lately
<GraveDigger>
i guess the update source is available at milkymist.org?
<lekernel>
sure, and without registration (unlike opencores)
<CIA-48>
flickernoise: Sebastien Bourdeauducq master * r22d1a10 / (src/input.c src/shortcuts.c): Middle pushbutton to power down - http://bit.ly/lc4gAq
<CIA-48>
milkymist: Sebastien Bourdeauducq master * r2373578 / boards/milkymist-one/standby/standby.v : Standby: debounce keys, only power up on rising edge - http://bit.ly/lYb3yV
<wpwrak>
(german police mistaking people) well, a few days ago, during a demonstration, one tried to arrest his boss and treated him to a bit of pepper spray for good measure
<wpwrak>
(boss) not his direct superior, of course. the guy pretty much on top of the local food chain.
<azonenberg>
lol
<CIA-48>
flickernoise: Sebastien Bourdeauducq master * rf49776c / (src/sysconfig.c src/sysconfig.h src/sysettings.c): New system settings dialog layout - http://bit.ly/ljkuIH
<CIA-48>
flickernoise: Sebastien Bourdeauducq master * r33a127d / src/sysconfig.c : Bugfix: Use DHCP address when DHCP is enabled in default config - http://bit.ly/my9GoZ
<CIA-48>
flickernoise: Sebastien Bourdeauducq master * re722120 / src/cp.c : New control panel layout - http://bit.ly/jHRqEw
<kristianpaul>
"safe" at home again
<lekernel>
any good alternative to libcurl? its main problem is it uses GNU/Autocrap and GNU/Libtrash which, as often, miserably fail
<lekernel>
the problem du jour is that due to some obscure autocrap-libtool interaction, the gcc flags end up specified twice which doesn't work with the -specs flag used by RTEMS
<azonenberg>
lekernel: I'm intimately familiar with the failings of libcurl
<azonenberg>
one of my projects is in seirous trouble due to a segfault deep in the library that nobody can diagnose
<azonenberg>
But i havent had issues with its build system (yet)
<azonenberg>
Of course, that doesnt stop me from using cmake whenever possible :P