00:01
sandeepkr has quit [Ping timeout: 276 seconds]
00:58
awallin has quit [Ping timeout: 276 seconds]
01:11
awallin__ has joined #m-labs
02:07
fengling has joined #m-labs
02:21
<
whitequark >
bb-m-labs: force build artiq
02:21
<
bb-m-labs >
build forced [ETA 18m52s]
02:21
<
bb-m-labs >
I'll give a shout when the build finishes
02:21
<
whitequark >
bb-m-labs: help
02:21
<
bb-m-labs >
Get help on what? (try 'help <foo>', 'help <foo> <bar>, or 'commands' for a command list)
02:22
<
whitequark >
bb-m-labs: commands
02:22
<
bb-m-labs >
buildbot commands: commands, dance, destroy, force, hello, help, last, list, mute, notify, shutdown, source, status, stop, unmute, version, watch
02:22
<
whitequark >
bb-m-labs: stop build artiq broken
02:22
<
bb-m-labs >
build 644 interrupted
02:22
<
whitequark >
sb0: you unplugged the kc705 and did not plug it in, right?
02:22
<
whitequark >
well, now you've broken all builds, and manuals aren't uploaded too
02:48
<
GitHub9 >
artiq/master de73bd1 Sebastien Bourdeauducq: gui/moninj: sort by channel. Closes #413
02:55
sb0 has joined #m-labs
02:56
<
sb0 >
whitequark, if you want to unbreak the build, revert the mtu commit (that looks innocuous but lwip is another serious piece of bugware)
02:59
kristianpaul has quit [Quit: leaving]
02:59
kristianpaul has joined #m-labs
03:00
<
whitequark >
alright
03:24
kristianpaul has quit [Read error: Connection reset by peer]
03:24
kristianpaul has joined #m-labs
03:27
<
sb0 >
whitequark, for supporting inheritance, will it need those slow virtualized methods you mentioned earlier?
03:32
<
whitequark >
depends
03:33
<
whitequark >
do you want to put a subclass where a superclass is expected? then yes
03:33
<
whitequark >
do you just want to do class C: def x(); class A(C); class B(C); ?
03:33
<
whitequark >
then no
03:33
<
whitequark >
but without virtual methods, I will make a copy of the method for every class, so that bloats code.
03:37
<
whitequark >
(LLVM can remove duplicate functions for trivial cases, but not all)
03:37
kristianpaul has quit [Quit: leaving]
03:37
kristianpaul has joined #m-labs
03:41
<
sb0 >
if that's easy go for it. it will solve the #414 case
03:47
<
sb0 >
again, python fucked up with inspect.getsource(). an optimist would think linecache is connected to the parser and getsource() returns the source corresponding to the object in memory no matter what
03:50
<
sb0 >
at least the 3.5 behavior is more consistent, even if still stupid: it returns whatever is on disk looks like a block of code near the position the parser read initially
03:51
<
sb0 >
you can make it return total garbage by inserting code before the object you are getsourcing
03:53
fengling has quit [Ping timeout: 240 seconds]
04:04
<
whitequark >
sb0: yes. python fucked up getsource. you cannot get the source for a lambda reliably at all
04:05
<
whitequark >
typical broken python API for which no thought went into it design
04:28
<
whitequark >
sb0: that will delay 1.0...
04:30
<
sb0 >
how large of a change is that?
04:31
<
sb0 >
are you talking about inheritance, or finding a solution to this python bug?
04:32
<
GitHub22 >
artiq/release-1 06268d1 Sebastien Bourdeauducq: gui/moninj: sort by channel. Closes #413
04:37
<
whitequark >
inheritance
05:06
sb0 has quit [Quit: Leaving]
06:19
mumptai has joined #m-labs
06:21
<
GitHub37 >
artiq/master ff721ff Robert Jordens: Revert "lwip: set MTU to 9000 to support jumbo frames"...
06:23
<
GitHub143 >
artiq/release-1 dcf082e Robert Jordens: Revert "lwip: set MTU to 9000 to support jumbo frames"...
06:41
sb0 has joined #m-labs
06:41
<
sb0 >
i had done incremental changes to avoid this problem but NO
06:42
<
sb0 >
i suppose that when building misoc, and it triggers an artiq build, conda still uses the old version of misoc somehow...
06:44
<
GitHub158 >
misoc/master fe550f3 Sebastien Bourdeauducq: Revert "liteeth: increase MTU to support jumbo frames"...
06:44
<
sb0 >
so maybe it's liteeth that has the problem and lwip is fine ...
07:03
<
sb0 >
bb-m-labs, force build artiq
07:03
<
bb-m-labs >
build forced [ETA 18m52s]
07:03
<
bb-m-labs >
I'll give a shout when the build finishes
07:13
<
sb0 >
rjo, the proper way to fix xilinx issues in migen is to describe the desired function (e.g. a non-power-of-2 block RAM) and then have a transform in migen.build that modifies them
07:14
<
sb0 >
ad-hoc hacks in cores are not portable
07:16
<
sb0 >
what the fuck
07:17
<
GitHub142 >
misoc/master 6e86189 Sebastien Bourdeauducq: Revert "liteeth: use non power-of-two RAM depth"...
07:24
<
rjo >
sb0: completely agree. was i inferring something else?
07:27
<
rjo >
sb0: i checked it again. ise for s3 tells you it inferred a non-power-of-two bram but then proceeds to allocate the next higher power-of-two size in terms of brams.
07:52
<
sb0 >
bb-m-labs, force build artiq
07:52
<
bb-m-labs >
build forced [ETA 18m52s]
07:52
<
bb-m-labs >
I'll give a shout when the build finishes
07:53
<
sb0 >
this makes no sense, unless the network broke for another reason (see murphy's law) ...
08:02
<
sb0 >
yup, i still cannot ping after downgrading
08:02
<
sb0 >
unbelievable ...
08:08
<
sb0 >
bb-m-labs, force build artiq
08:08
<
bb-m-labs >
build forced [ETA 18m52s]
08:08
<
bb-m-labs >
I'll give a shout when the build finishes
08:10
<
sb0 >
ok, ping works again now (though I have no idea why) with git master flashed manually, let's see if the buildbot passes again
08:11
<
_florent_ >
sb0: I'm going to look at liteeth, I saw the revert from you/rjo, what was the problem? (just trying to understand)
08:13
<
sb0 >
no ping. didn't look further
08:17
<
_florent_ >
ok thanks, I'll look at that then
08:57
fengling has joined #m-labs
09:43
mumptai has quit [Remote host closed the connection]
10:07
kuldeep_ has quit [Ping timeout: 244 seconds]
10:17
<
sb0 >
bb-m-labs, force build artiq
10:17
<
bb-m-labs >
build forced [ETA 18m52s]
10:17
<
bb-m-labs >
I'll give a shout when the build finishes
10:18
kuldeep_ has joined #m-labs
10:31
<
sb0 >
whitequark, the buildbot flashes wrong versions of the gateware
10:33
<
whitequark >
conda madness again?..
10:34
<
sb0 >
it just flashed 2.0.dev+260.gff721ff
10:34
<
whitequark >
sb0: I don't get it
10:34
<
whitequark >
it just built that
10:34
<
sb0 >
"Built Apr 30 2016 14:24:07"
10:35
<
whitequark >
ok and? what is the problem?
10:35
<
sb0 >
it's nowhere near 14:24, the board doesn't ping, and when i install the conda package it just uploaded manually and reflash, it does ping
10:36
<
sb0 >
I suppose that misoc modifications without artiq modifications cause issues
10:36
<
whitequark >
Date b'2016/04/30\x00'
10:36
<
whitequark >
Time b'14:33:15\x00'
10:37
<
whitequark >
this seems to check out
10:37
<
whitequark >
you are right.
10:37
<
whitequark >
it re-uploaded the anaconda package, but it was cached locally.
10:37
<
whitequark >
I can absolutely clean all conda caches before every build, but then it would be incredibly slow, like Travis was
10:38
<
sb0 >
what about deleting migen, misoc and all artiq packages from conda caches
10:38
<
whitequark >
that's horribly hacky
10:38
<
whitequark >
I guess that will do.
10:39
sandeepkr has joined #m-labs
11:09
sandeepkr_ has joined #m-labs
11:10
sandeepkr__ has joined #m-labs
11:10
sandeepkr has quit [Read error: Connection reset by peer]
11:14
sandeepkr_ has quit [Ping timeout: 260 seconds]
11:18
sandeepkr_ has joined #m-labs
11:22
sandeepkr__ has quit [Ping timeout: 240 seconds]
11:22
sandeepkr__ has joined #m-labs
11:26
sandeepkr_ has quit [Ping timeout: 260 seconds]
11:29
sb0 has quit [Quit: Leaving]
11:56
cr1901_modern has joined #m-labs
11:57
fengling has quit [Ping timeout: 240 seconds]
12:06
cr1901_modern1 has quit [*.net *.split]
12:10
fengling has joined #m-labs
12:14
fengling has quit [Ping timeout: 240 seconds]
12:48
fengling has joined #m-labs
15:28
folkert has quit [Ping timeout: 276 seconds]
15:33
sb0 has joined #m-labs
15:53
sandeepkr_ has joined #m-labs
15:53
sandeepkr__ has quit [Read error: No route to host]
15:54
sandeepkr has joined #m-labs
15:58
sandeepkr_ has quit [Ping timeout: 260 seconds]
16:39
sb0 has quit [Quit: Leaving]
18:00
sandeepkr_ has joined #m-labs
18:04
kuldeep_ has quit [Ping timeout: 276 seconds]
18:04
sandeepkr has quit [Ping timeout: 260 seconds]
18:19
bb-m-labs has quit [Quit: buildmaster reconfigured: bot disconnecting]
18:19
<
GitHub4 >
buildbot-config/master fcddd57 whitequark: Partially clean conda caches.
18:20
<
GitHub21 >
buildbot-config/master c45c698 whitequark: Partially clean conda caches.
18:20
bb-m-labs has joined #m-labs
18:20
<
whitequark >
bb-m-labs: force build artiq
18:20
<
bb-m-labs >
build #653 forced
18:20
<
bb-m-labs >
I'll give a shout when the build finishes
18:21
bb-m-labs has quit [Client Quit]
18:21
bb-m-labs has joined #m-labs
18:21
<
whitequark >
bb-m-labs: force build artiq
18:21
<
GitHub13 >
buildbot-config/master db8ac6c whitequark: Partially clean conda caches.
18:21
<
bb-m-labs >
build #654 forced
18:21
<
bb-m-labs >
I'll give a shout when the build finishes
18:22
kuldeep_ has joined #m-labs
18:26
sandeepkr has joined #m-labs
18:30
sandeepkr_ has quit [Ping timeout: 240 seconds]
18:30
kuldeep_ has quit [Ping timeout: 260 seconds]
18:45
kuldeep_ has joined #m-labs
19:03
bb-m-labs has quit [Quit: buildmaster reconfigured: bot disconnecting]
19:03
bb-m-labs has joined #m-labs
19:03
<
GitHub131 >
buildbot-config/master 9b5b62d whitequark: Partially clean conda caches.
19:06
sandeepkr_ has joined #m-labs
19:06
<
whitequark >
bb-m-labs: force build artiq
19:06
<
bb-m-labs >
build #655 forced
19:06
<
bb-m-labs >
I'll give a shout when the build finishes
19:07
kuldeep_ has quit [Ping timeout: 260 seconds]
19:07
sandeepkr has quit [Ping timeout: 252 seconds]
19:23
<
whitequark >
sb0: something else fails now
19:25
kuldeep_ has joined #m-labs