<mindrunner>
is there any way to figure out by myself? or build them by myself?
<sb0>
not sure where the sources are, and rjo is on holidays...
<mindrunner>
:)
<sb0>
a taobao vendor just recommended me to put "oil from the gas station" into a rotary vane vacuum pump. i wonder if that was lost in translation, if he's mocking me, or if he's being serious.
<GitHub149>
[migen] sbourdeauducq pushed 1 new commit to master: http://git.io/vfLox
<GitHub149>
migen/master 28dde1e Guy Hutchison: fhdl/verilog: add flag to produce ASIC-friendly output
travis-ci has joined #m-labs
<travis-ci>
m-labs/migen#11 (master - 28dde1e : Guy Hutchison): The build passed.
<mindrunner>
sb0, you remember that I got this uart problem where some strange 0-byte was read every now and then on my ppro?
<sb0>
yes
<mindrunner>
i got the same problem with the pipistrello
<sb0>
I don't have it on the kc705
<sb0>
(which doesn't use one of those nasty ftdi-chips for the uart)
<mindrunner>
hmm, so you think ftdi might be the problem?
<sb0>
I don't know, that's just one data point
<mindrunner>
pipistrello has some uart fifo on second channel, I thought i could use direct communication instead through the ftdi chip then
<sb0>
ftdi-chips are a crappy design (no CPU/microcode, all hardwired USB logic like huge FSMs) coupled with crappy drivers, so of course I'd tend to suspect them
<mindrunner>
:))
<sb0>
but they might not be the only part of the issue. lots of people use ftdi-chips without apparent problems...
<mindrunner>
yeah i know.
<mindrunner>
thats so weird.. :(
<GitHub4>
[artiq] sbourdeauducq pushed 5 new commits to master: http://git.io/vfLDe
<whitequark>
JRuby implements Ruby with C extensions (they look just like the Python ones) by taking a machine-code-optimizing Ruby interpreter and adding a machine-code-optimizing C interpreter to it
<whitequark>
and it gets some inner loops an order of magnitude faster than what an actual C compiler gives you
aeris_ has quit [Ping timeout: 245 seconds]
aeris_ has joined #m-labs
aeris_ has quit [Ping timeout: 248 seconds]
antgreen` has quit [Ping timeout: 256 seconds]
<sb0>
is there a simple way to tell gcc to align the 9th byte of a char array to a 32-bit boundary?
<whitequark>
no. I don't think there is any other way besides requesting a larger array and pointing inside it.
<whitequark>
(I'm assuming you mean a stack/static allocated one.)
<whitequark>
the symbols in ELF only have the n associated to them, where they're aligned to 2**n
mumptai has joined #m-labs
<GitHub64>
[artiq] sbourdeauducq pushed 2 new commits to master: http://git.io/vfYYn
<GitHub64>
artiq/master 17f61b8 Sebastien Bourdeauducq: pxi6733: minor fixes
<GitHub64>
artiq/master 9f52277 Sebastien Bourdeauducq: coredevice/core: set default for ref_period
<GitHub153>
[artiq] sbourdeauducq pushed 1 new commit to master: http://git.io/vfYWU
<GitHub153>
artiq/master 18106cc Sebastien Bourdeauducq: comm: refactor to support lwip event model
<ysionneau>
use a READ_INT() macro which will read 4 bytes and mask/shift to create the int?
<ysionneau>
use a union { int integer; char bytes[4]; }; ??
<cr1901_modern>
The code's C99?
<cr1901_modern>
If so, I THINK a union can be used liked that, but I'm not sure
<ysionneau>
code is C99 but there is no flag used to tell it's C99
<cr1901_modern>
My gut feeling is to try READ_INT() macro
<ysionneau>
alright, union made the warning go away
<ysionneau>
but I still feel like I'm using a hack instead of coding it clean
<cr1901_modern>
Do you intend to merge your LM32 port of Edge back into Net?
<ysionneau>
if it works, and if it's clean enough for NetBSD standards: yes
<ysionneau>
the NetBSD guys I met at EuroBSDcon were quite interested about it and wanted to make me a NetBSD developer right away so that I could push it
<ysionneau>
but it's way too early for that
<cr1901_modern>
Lol, good stuff. :P
<cr1901_modern>
I asked mainly because- yes, I'm interested on hacking on it, but I'd like to focus on hacking Net for now. I am capable of writing a very basic driver for net using the autoconf framework
<ysionneau>
I had to do it also, to have a basic clock/uart driver
<ysionneau>
but it was months ago and even at that time the autoconf framework was still blury in my mind :p
<ysionneau>
but anyway, I have two repositories , one at edgebsd, another one on github which is based on NetBSD
<ysionneau>
but the two are very similar anyway
<cr1901_modern>
Ahhh, okay I remember seeing the latter actually