01:55 <sb0> SOP can be deduced from EOP, but the opposite is not exactly true

01:56 <sb0> you'd have to wait for the next packet (which may possibly never come) to know the end of the current one

01:57 <sb0> not a problem for video streams, but generally it is

02:33 <sb0> _florent_, is the cache size always in 32-bit words?

02:33 <sb0> https://github.com/m-labs/misoc/blob/master/misoc/interconnect/wishbone.py#L459

02:33 <sb0> or is this dependent on the word widths of the two interfaces?

03:26 <GitHub139> artiq/master f33ed88 whitequark: language.core: implement round(value, width) (fixes #203).

03:26 <GitHub139> [artiq] whitequark pushed 1 new commit to master: http://git.io/vE36X

03:27 <GitHub72> [artiq] whitequark force-pushed master from f33ed88 to a250b5d: http://git.io/vYgPK

03:27 <GitHub72> artiq/master a250b5d whitequark: language.core: implement round(value, width) (fixes #203).

03:45 <GitHub128> [artiq] sbourdeauducq pushed 1 new commit to master: http://git.io/vE3Xo

03:45 <GitHub128> artiq/master fc299ca Sebastien Bourdeauducq: language/environment: disable processors by default

03:59 <GitHub114> [artiq] sbourdeauducq pushed 1 new commit to master: http://git.io/vE3Mq

03:59 <GitHub114> artiq/master 23355d8 Sebastien Bourdeauducq: coredevice: restore RTIOCollisionError

04:03 <GitHub22> [artiq] sbourdeauducq pushed 1 new commit to master: http://git.io/vE3Dk

04:03 <GitHub22> artiq/master f652292 Sebastien Bourdeauducq: coredevice/exceptions: PEP8

04:11 <GitHub188> [artiq] sbourdeauducq pushed 1 new commit to master: http://git.io/vE3yr

04:11 <GitHub188> artiq/master b4b0dcc Sebastien Bourdeauducq: test/coredevice/rtio: remove obsolete functions

06:59 <GitHub94> [migen] sbourdeauducq pushed 2 new commits to master: http://git.io/vEs8w

06:59 <GitHub94> migen/master eb8e8d4 Robert Jordens: sim: lower specials, closes #34

06:59 <GitHub94> migen/master a5fb9ba Sebastien Bourdeauducq: sim: make unlowered specials an error

07:06 <GitHub79> [migen] sbourdeauducq pushed 2 new commits to master: http://git.io/vEsB2

07:06 <GitHub79> migen/master 2a81389 Sebastien Bourdeauducq: fhdl/simplify/MemoryToArray: remove spurious memory ports from specials

07:06 <GitHub79> migen/master 98ce81b Sebastien Bourdeauducq: sim: run MemoryToArray before lowering specials

09:13 <GitHub10> [artiq] sbourdeauducq pushed 1 new commit to master: http://git.io/vEsb0

09:13 <GitHub10> artiq/master e4233d7 Sebastien Bourdeauducq: examples/speed_benchmark: fix scheduler arguments

15:04 <cr1901_modern> rjo: What are the constraints, if any, on the maximum/minimum values a scan can take? I want to implement axis scaling correctly, and the way I do it is different if the maximum/minimum are "can be stored in a typical datatype" versus "unbounded to the amount of memory your computer has"

15:08 <cr1901_modern> Actually, I'm not certain it matters; the "correct" way to implement an infinitely long line in Qt is to set it's length to a very large value (contrast to scaling a line to the size of the view each time it's scaled).

15:09 <cr1901_modern> s/is to/seems to be/

15:50 <cr1901_modern> rjo: Also, I've placed all the potential edge cases as TODOs in my src file, so I'll bring them up as I tackle each.

16:11 <sb0> cr1901_modern, what's complicated with rescaling the line everytime the zoom changes? or excluding the line from the zoomed bit, ie put it on a different layer? how does pyqtgraph does it?

16:11 <sb0> and why use qt's zooming features at all, if they're inappropriate?

16:38 <cr1901_modern> sb0: There's nothing complicated about it; I misread. The situations where zooming isn't appropriate doesn't apply in my case, so scratch that.

20:57 <rjo> cr1901_modern: the quantities usually have units. the values themselves can be anywhere from +-1e-12 (ps) to +-1e15 (THz) but floating point will be sufficient to represent the quantify. no need for arbitrary precision.