Quickly closing/reopening applets (e.g. quickly clicking the checkbox
on an entire folder of applets) would previously lead to an occasional
KeyError on the self.dock_to_item access in on_dock_closed, as close()
would be invoked more than once.
The geometry/checked state handling can potentially be cleaned up
further, but at least this avoid the crash.
The leading empty line seems to be required by Sphinx 1.8.3.
The arguments must be strict JSON when prerendering for a target that is
not "html". Browser JSON parsing may have been more lenient.
Signed-off-by: Stephan Maka <stephan@spaceboyz.net>
* handle disconnects like core device address changes and do a
disconnect/connect iteration
* after connection failure wait 10 seconds and try again
* this addresses the slight regression from release-2
to release-3 where the moninj protocol was made stateful
(#838 and #1125)
* it would be much better to fix smoltcp/runtime to no loose the
connection under pressure (#1125)
* the crashes reported in #838 look more like a race condition
* master disconnects still require dashboard restarts
Signed-off-by: Robert Jördens <rj@quartiq.de>
This reverts 425cd7851, which broke the use of casts to define
integer width.
Instead of it, two steps are taken:
* First, literals are monomorphized, leading to predictable result.
* Second, casts are monomorphized, in a top-bottom way. I.e.
consider the expression `int64(round(x))`. If round() was visited
first, the intermediate precision would be 32-bit, which is
clearly undesirable. Therefore, contextual rules should take
priority over non-contextual ones.
Fixes#1252.
artiq_influxdb doesn't work with aiohttp 0.17 (anymore), as the
ClientSession API changed. I have not looked up precisely which
is the first version that works, but 3.x has been out for almost
a year and is available on the Anaconda/conda-forge channels.
* fix broken and old URLs to anaconda/miniconda
* append conda-forge, do not prepend it (consistent with conda-forge
instructions and does not blindly prefer packages in conda-forge over
packages in defaults)
* shorten m-labs repo
async is now a full (non-contextual) keyword.
There are two more instances:
- artiq/frontend/artiq_client.py
- artiq/devices/thorlabs_tcube/driver.py
It is not immediately clear how to fix those, so they are left for
later work.
This is difficult to test without hardware mocks or some
form of phase readback, but the symptom was that e.g.
`self.dds.set(…, ref_time=now_mu() - 1)` would fail
periodically, that is, whenever bit 32 of the timestamp
would be set (which would be turned into the sign bit).
This is a fairly sinister issue, and is probably a compiler
bug of some sort (either accepts-invalid or wrong type inference).
* tune_sync_delay: the opposite IO_UPDATE to SYNC_CLK alignment may not be perfectly
mis-aligned
* set_mu_speed: seems to be slower on the buildbot
Signed-off-by: Robert Jördens <rj@quartiq.de>
Previously the TSC was truncated to an even coarse RTIO periods before doing
the setting SPI xfer. Afterwards the the IO update pulse would introduce
at least one but less than two RTIO cycles. Ultimately the RTIO TSC was
truncated again to even. If the SPI xfer takes an odd number of RTIO
periods, then a subsequent xfer would collide.
close#1229
Signed-off-by: Robert Jördens <rj@quartiq.de>
This allows the migen namer to derive names for the ADC return clock
domain in the case of multiple SUServos
close#1201
Signed-off-by: Robert Jördens <rj@quartiq.de>
