sb0 changed the topic of #m-labs to: ARTIQ, Migen, MiSoC, Mixxeo & other M-Labs projects :: fka #milkymist :: Logs http://irclog.whitequark.org/m-labs
ylamarre has joined #m-labs
fengling has quit [Quit: WeeChat 1.1.1]
ylamarre has quit [Quit: ylamarre]
attie has quit [Ping timeout: 248 seconds]
attie has joined #m-labs
FabM has joined #m-labs
bentley` has quit [Ping timeout: 240 seconds]
bentley` has joined #m-labs
sb0 has joined #m-labs
<GitHub194> [misoc] enjoy-digital pushed 1 new commit to master: http://git.io/vqiHA
<GitHub194> misoc/master e6da1d1 Florent Kermarrec: wishbone2lasmi: fix "READ_DATA" state
bentley` has quit [Ping timeout: 256 seconds]
bentley` has joined #m-labs
mumptai has quit [Ping timeout: 252 seconds]
sb0_ has joined #m-labs
sb0 has quit [Ping timeout: 276 seconds]
<GitHub160> [artiq] sbourdeauducq pushed 2 new commits to master: http://git.io/vqPxh
<GitHub160> artiq/master 96a5d73 Sebastien Bourdeauducq: worker: split build stage from prepare
<GitHub160> artiq/master 9d4073e Sebastien Bourdeauducq: master: remove simple parameter history
travis-ci has joined #m-labs
<travis-ci> m-labs/artiq#299 (master - 9d4073e : Sebastien Bourdeauducq): The build passed.
travis-ci has left #m-labs [#m-labs]
ylamarre has joined #m-labs
key2 has joined #m-labs
kristianpaul has quit [Quit: leaving]
balrog has quit [Remote host closed the connection]
kristianpaul has joined #m-labs
mumptai has joined #m-labs
balrog has joined #m-labs
ylamarre has quit [Quit: ylamarre]
<rjo> is there anything wrong with having a AsyncFIFO with depth=1 or 2 as a cdc (apart from the possibility of an overflow)?
<ysionneau> rjo: I'm having a look at renaming llvmlite to llvmlite_or1k, do you know any setup.py field that does that?
<ysionneau> I've searched a bit and didn't find how to do that
<ysionneau> looks like I need to rename the llvmlite directory ... no?
<rjo> name
<rjo> ah. you want the target directory in the python package namespace. not just the package name.
<rjo> one would first have to make sure that llvmlite only does relative imports from itself.
<rjo> and then, in setup.py, change the name= field and then package_dir={"llvmlite_or1k": "llvmlite"}
<rjo> that should be it.
<ysionneau> I already tried to change the name to "llvmlite_or1k" but it didn't change anything
<ysionneau> package is still installed in "llvmlite"
<ysionneau> but I didn't try package_dir
<ysionneau> let's try that
<ysionneau> package_dir didn't help :/
<ysionneau> (ack about checking llvmlite only does relative import)
<whitequark> sb0_: this turbopump uses two bearings
<whitequark> upper is maglev
<whitequark> bottom is ceramic oiled
<whitequark> astoundingly, I even found the oil, in reasonable qty, and for reasonable price: http://www.ajvs.com/cambridge-mills-cmp-500-turbo-pump-oil-full-synthetic-hydrocarbon-oil-25-liters-7237
<whitequark> it's presumably its stock oil
<rjo> ysionneau: package_dir is the right thing. you might have replace find_packages() by the explicit list as well.
<ysionneau> llvmlite setup.py does not use find_packages() (at least I don't see it)
<rjo> good. then add _or1k to all lines at https://github.com/numba/llvmlite/blob/master/setup.py#L82
<ysionneau> it is explicitly listing packages = ['llvmlite', 'llvmlite.binding', ...]
<rjo> and then do the package_dir remapping accordingly
<ysionneau> oh ok
<ysionneau> I should really learn to use setup.py correctly.
<rjo> otoh this entire problem might not be relevant at all.
<rjo> i have come to like pyvenvs (conda envs are just an instance) very much.
<rjo> and then llvmlit-host vs llvmlite-or1k is not a problem anymore.
<ysionneau> ah :)
<ysionneau> conda envs were not enough to allow you to use both numba and llvmlite-or1k ?
<ysionneau> I mean using both but in different conda envs
<ysionneau> ah, I guess no because both would install in the same lib directory
<rjo> so if you feel this is difficult, then skip it. the only case where it would matter is if you want numba and artiq in the same python session/interpreter/toolkit. while a possible use case it is a bit unlikely.
<rjo> having env_a with llvmlite and env_b with llvmlite-or1k only needs different conda package names to distinguis them in the conda root. the module directories can conflict.
<rjo> (unless somebody needs numba and artiq in the same env).
<ysionneau> all right, seems to copy in the right directories now
<ysionneau> but setup.py install fails at some point
<ysionneau> ah got it
<ysionneau> rjo: seems to have worked, and I can even import llvmlite_or1k and llvmlite_or1k.llvmpy which means at least those don't trigger import issue
<ysionneau> it's installing in miniconda3/lib/python3.4/site-packages/llvmlite_or1k/
<GitHub122> [artiq] jordens pushed 1 new commit to master: http://git.io/vqyXx
<GitHub122> artiq/master 80eea4c Robert Jordens: test: relax test_time_keeps_running...
<rjo> ysionneau: nice.
* ysionneau can even run the transforms.py test after a bit of renaming
<ysionneau> will commit that tomorrow after a bit more tests if it still works
<ysionneau> (never commit before going to sleep, always re-check in the morning :p)
<ysionneau> thanks for the help on setup.py btw
<rjo> np.
mumptai has quit [Quit: Verlassend]
<whitequark> first plasma