01:50 <sb0> maybe the solution to this annoying simulation bug is just ditch Verilog for simulation

01:51 <sb0> figuring out verilog tricks that get that mess to work sounds as complicated as just interpreting the logic directly

01:51 <sb0> plus, there will be no more VPI module to install (unless people want to co-simulate verilog code)

03:25 <sb0> rjo, http://www.thorlabs.com/thorProduct.cfm?partNumber=TCH002 << is this thing more than a USB hub from the PC point of view?

03:43 <rjo> sb0: just an expensive one.

03:43 <sb0> so no driver needed then

03:45 <rjo> no. who mentioned that tch002?

03:45 <GitHub151> [artiq] sbourdeauducq pushed 1 new commit to master: http://git.io/fhOvlA

03:45 <GitHub151> artiq/master 8d305e3 Sebastien Bourdeauducq: doc: some precisions about controllers

03:45 <sb0> it's in the original project spec

03:46 <sb0> "We require Controllers to be made [...] for the following additional hardware [...] Thorlabs TDC001 servo motor controllers, TPZ001 piezo controllers, ***and TCH002 hubs***"

03:47 <rjo> ack. trivially done then ;)

03:49 <rjo> sb0: and the others are just a more or less asynchronous single-shot devices like those labbricks.

03:49 <sb0> there's also the NI PXI-6733

03:50 <sb0> are you still using those, or switching to PDQ?

03:51 <rjo> sb0: some experiments use them.

03:52 <rjo> sb0: and it would be nice if the API for the PXI-6733 and PDQ2 would be very similar. But I did not have that in mind when I wrote the PDQ2 driver/api.

03:53 <rjo> sb0: i personally would push the 6733 to the bottom of the list.

03:54 <sb0> ok. speaking of pdq, I still have some questions: https://ssl.serverraum.org/lists-archive/artiq/2014-October/000150.html

03:55 <rjo> ah. forgot. sorry.

03:56 <rjo> anything else i promised and did not deliver (apart from getting more people deeply involved ;)?

03:57 <sb0> well, it seems that workshop was mostly successful :)

03:58 <sb0> and no, not that I remember. thanks :)

04:28 <rjo> sb0: i replied to your mail. let me know if i missed the point.

04:33 <rjo> sb0: ah. the talk slides and your comments on them. i'll try to do that tomorrow...

05:28 <sb0> rjo, about those slides, I dropped JSON because serializing numpy arrays was a mess

05:29 <sb0> and writing a numpy-aware serializer in python is easy, https://github.com/m-labs/artiq/blob/master/artiq/management/pyon.py

05:33 <zyp> especially when you don't need to deal with untrusted input

05:37 <rjo> sb0: i agree. this is why i tend to gravitate towards yaml which does support that.

05:38 <rjo> sb0: but pyon is also just fine.

05:39 <rjo> sb0: yaml may also be a bit too flexible for the embedded side of things.

05:54 <zyp> please don't believe that simply removing builtins will make it safe, ref: nparray.__globals__['__builtins__'].open('foo.txt', 'w').write('foo')

05:58 <sb0> zyp, this is not meant to be used over the open internet, but between trusted machines on a LAN

06:00 <zyp> fair enough, it's just a reminder :)

06:03 <zyp> the dangerous part is that it's seemingly secure, so somebody might decide to use it for untrusted input

06:09 <sb0> that's pretty much the case for most software ;)

06:10 <sb0> but yes, I see you point

06:10 <sb0> *your

06:18 <rjo> a funny sidestory to that: we have a so-called "research network" now (not "research on networks" but a "network for research equipment") because people definitely want to connect their 1990 oscilloscope with Windows NT 3.whatever to the network. virtually no rules enforced on that network. a honeypot. the story goes that some of the more powerful oscilloscopes were hijacked and used as warez/etc distribution bots. i am just waiting for some virus to run amok i

06:19 <zyp> haha

07:06 <sb0> ysionneau, can you save your simulator-breaking code somewhere?

07:06 <sb0> might want to use it as a test case for the bugfix

07:48 <GitHub105> [artiq] sbourdeauducq pushed 1 new commit to master: http://git.io/mrt3bQ

07:48 <GitHub105> artiq/master 4cae553 Sebastien Bourdeauducq: pc_rpc: server identification support

08:07 <sj_mackenzie> sb0: github.com/upstandinghackers/hammer

09:03 <ysionneau> 08:06 < sb0> ysionneau, can you save your simulator-breaking code somewhere? < sure!

09:07 <sb0> ysionneau, did the artiq tutorial go smoothly on your board/pc?

09:08 <ysionneau> yes, I don't remember any issue, except that I needed to re-run a few times experiments, but maybe I had another terminal with the serial port open (flterm), not sure

09:09 <ysionneau> sb0 how can I convert a 'Proxy' (from simulation context) into an integer?

09:09 <sb0> and was everything clear?

09:09 <sb0> ysionneau, the selfp proxy?

09:09 <ysionneau> some selfp.a signal

09:10 <ysionneau> so that I can do selfp.b[selfp.a]

09:10 <sb0> if selfp.a is a signal, then the getattr of selfp for "a" should return an integer

09:10 <ysionneau> ok maybe I misunderstood the error then

09:11 <ysionneau> ok I worked around the issue

09:11 <ysionneau> I could just use self.a instead of selfp.a

09:11 <ysionneau> since a was a constant value

09:12 <ysionneau> 10:09 < sb0> and was everything clear? < I think if it was not clear I would have added some text to my patch, so except the python3.3 issue and the sudo -E PATH=$PATH stuff I think it was OK

09:12 <sb0> ah yes, selfp doesn't work for anything else than a signal

09:12 <ysionneau> but I can try again

09:13 <sb0> in fact I should probably remove selfp, even though it gives some syntactic sugar in simple use cases, it causes obscure errors in others

09:13 <ysionneau> ah ok, I was doing selfp.dfi.phases[selfp.sdrphy.phy_settings.rdphase]

09:15 <ysionneau> 10:09 < sb0> and was everything clear? < I can try again all the install procedure from scratch, I have setup a 'fresh' VM snapshot for that kind of stuff, so that I can easily try 'from scratch' stuff

09:15 <ysionneau> I just have the toolchain already compiled

09:18 <sb0> ysionneau, your video: https://www.youtube.com/watch?v=AfEPbw4nREo

09:18 <sb0> from orconf

09:23 <ysionneau> ah thanks!

09:23 <ysionneau> maybe it will be hard to see since it's all console demo

09:24 <ysionneau> ok the focus is wrong on the slides :(

09:24 <sb0> ysionneau, I was asking more about using it, not installing it

09:24 <sb0> creating the pulses etc.

09:29 <ysionneau> I just went through the installing part so far, and then tried running a few pre-written experiments and had the issues I told you about, then I switched on the spiflash

09:29 <ysionneau> -on+to