Release 2016.10 - RC1 about release-specs HOT 14 CLOSED

I volunteer to have an eye ( or maybe two) on 02-Tests

from release-specs.

I've tested 04/Task 03 between:

• iotlab-m3 - iotlab-m3
• iotlab-m3 - samr21-xpro
• pba-d-01-kw2x - samr21-xpro
• remote-reva - samr21-xpro

Last combination seems unstable, had a relatively high loss (40%) and the remote got stuck after 612 pings. I assume it's a known issue but I forget which issue it was (@smlng, @aeneby)

from release-specs.

Last combination seems unstable, had a relatively high loss (40%) and the remote got stuck after 612 pings. I assume it's a known issue but I forget which issue it was (@smlng, @aeneby)

I have no clue about that, but does RIOT-OS/RIOT#5754 fix it for you (it still needs a backport for the release)?

from release-specs.

@PeterKietzmann or did you mean RIOT-OS/RIOT#5719?

from release-specs.

I had #5786 in mind which I thought to be solved. However, #5719 seems similar and still persists. @smlng might have more insights.

from release-specs.

@PeterKietzmann I'm asking since this would trigger a new RC if RIOT-OS/RIOT#6041 and it gets merged. So please repeat the tests with RIOT-OS/RIOT#6041.

from release-specs.

No it does not solve the issue. The problem lies somewhere in the driver and only occurs when packets get fragmented. Probably it is even on the Atmel side (#5486). IIRC Sebastian observed that fragments from the cc2538 are sent out faster than any other device can receive/reassemble.

from release-specs.

Build error on unittests:

/usr/local/src/RIOT/sys/net/gnrc/application_layer/coap/gcoap.c:62:5: error: missing initializer for field ‘target’ of ‘gnrc_netreg_entry_t’ [-Werror=missing-field-initializers]
.netreg_port = {NULL, 0, KERNEL_PID_UNDEF},
^
In file included from /usr/local/src/RIOT/sys/include/net/gnrc.h:289:0,
from /usr/local/src/RIOT/sys/include/net/gnrc/coap.h:138,
from /usr/local/src/RIOT/sys/net/gnrc/application_layer/coap/gcoap.c:22:
/usr/local/src/RIOT/sys/include/net/gnrc/netreg.h:182:7: note: ‘target’ declared here
} target; /**< Target for the registry entry */
^
cc1: all warnings being treated as errors
/usr/local/src/RIOT/Makefile.base:64: recipe for target '/usr/local/src/RIOT/tests/unittests/bin/native/gcoap/gcoap.o' failed

from release-specs.

I volunteer to have an eye ( or maybe two) on 02-Tests

Me too :)

from release-specs.

Great! @kYc0o @A-Paul could one of you open an issue to organize the effort?

from release-specs.

Hi @kYc0o, I have to leave for 2-3 hours and have already started Task #1. Maybe you can start Task #2 or #3, so we get not race conditions? :)

from release-specs.

Yes perfect, no problem, I had planned to start at 6pm so I'm ok with that.

from release-specs.

Okay, tests already show some issues with this RC. So please finish up the tests you are currently running and then stop testing for now. Let's merge and backport

• RIOT-OS/RIOT#6028
  • [merged]
  • [backported]
• RIOT-OS/RIOT#6045
  • [merged]
  • [backported]
• RIOT-OS/RIOT#6046
  • [merged]
  • [backported]
• RIOT-OS/RIOT#6047
  • [merged]
  • [backported]

and any PR that might fall out of

• #27 (comment),
• #27 (comment),
• #27 (comment) #28,
• #26 (comment)

and start a new RC from there

from release-specs.

Superseded by #29

from release-specs.