02-tests/task 01 Release 2016.07 - RC1

[BytesGalore]

Task # 01

Execute all tests on native and check their output:

• [x] bitarithm_timings
• [x] bloom_bytes
• [x] coap
• [x] conn_ip
• [x] cpp11_condition_variable
• [x] cpp11_mutex
• [x] cpp11_thread
• [x] driver_adt7310
• [x] driver_at30tse75x
• [x] driver_at86rf2xx
• [x] driver_bh1750
• [x] driver_bmp180
• [x] driver_dht -> compiles without warning runs and obviously fails
• [x] driver_enc28j60
• [x] driver_encx24j600
• [x] driver_hdc1000
• [x] driver_hih6130
• [x] driver_ina220
• [x] driver_isl29020
• [x] driver_isl29125
• [x] driver_kw2xrf
• [x] driver_l3g4200d
• [x] driver_lis3dh
• [x] driver_lis3mdl
• [x] driver_lps331ap
• [x] driver_lsm303dlhc
• [x] driver_mag3110
• [x] driver_mma8652
• [x] driver_mpl3115a2
• [x] driver_mpu9150
• [x] driver_mq3
• [x] driver_nrf24l01p_lowlevel
• [x] driver_nrfmin
• [x] driver_nvram_spi
• [x] driver_pcd8544
• [x] driver_pir -> compiles without warning runs and obviously fails on GPIO access
• [x] driver_servo
• [x] driver_si70xx
• [x] driver_srf02
• [x] driver_srf08
• [x] driver_tcs37727
• [x] driver_tmp006
• [x] driver_xbee
• [x] emb6 -> automatically sets samr21-xpro when BOARD is not set, build with BOARD=native make for the expected build fail
• [x] fault_handler
• [x] float -> only shows/spams - on the screen
• [x] fmt_print
• [x] gnrc_ipv6_ext
• [x] gnrc_sixlowpan
• [x] irq
• [x] leds
• [x] libfixmath
• [x] lwip -> automatically sets iotlab-m3 when BOARD is not set, build with BOARD=native make
• [x] malloc -> only malloc tested yet since I don't want to wait for filling my whole RAM with 1024 byte chunks
• [x] minimal
• [x] msg_send_receive
• [x] msg_try_receive
• [x] mutex_order
• [x] mutex_unlock_and_sleep
• [x] netdev2_test
• [x] netstats_l2
• [x] nhdp
• [x] periph_adc
• [x] periph_cpuid
• [x] periph_dac
• [x] periph_gpio -> compiles without warning and runs
• [x] periph_hwrng
• [x] periph_i2c
• [x] periph_pwm
• [x] periph_rtc -> compiles without warning and runs: f: rtc_set_alarm: not implemented
• [x] periph_rtt
• [x] periph_spi
• [x] periph_timer
• [x] periph_uart -> compiles without warning and runs: UART used for STDIO (the shell): UART_DEV(-1)
• [x] pipe
• [x] pkg_cmsis-dsp
• [x] pkg_jsmn
• [x] pkg_micro-ecc
• [x] pkg_u8g2
• [x] posix_semaphore
• [x] posix_sleep
• [x] pthread
• [x] pthread_barrier
• [x] pthread_cleanup
• [x] pthread_condition_variable
• [x] pthread_cooperation
• [x] pthread_rwlock
• [x] pthread_tls
• [x] saul -> compiles without warning and runs: No SAUL devices present
• [x] sched_testing
• [x] shell
• [x] sizeof_tcb
• [x] slip (compiled and started, function not tested yet)
• [x] struct_tm_utility
• [x] thread_basic
• [x] thread_cooperation
• [x] thread_exit
• [x] thread_flags
• [x] thread_flood
• [x] thread_msg
• [x] thread_msg_avail
• [x] thread_msg_block_wo_queue
• [x] thread_msg_block_w_queue
• [x] thread_msg_seq
• [x] warn_conflict
• [x] xtimer_drift
• [x] xtimer_hang
• [x] xtimer_longterm
• [x] xtimer_msg
• [x] xtimer_msg_receive_timeout
• [x] xtimer_now64_continuity
• [x] xtimer_remove
• [x] xtimer_reset
• [x] xtimer_usleep_until
• [x] zep (compiled and started, function not tested yet)

list made manually performing:

ls -1a | sed  -e 's/^/- [ ] /'

in RIOT/tests

[BytesGalore]

feel free to edit the list :)

[kYc0o]

Great! Thanks @BytesGalore @miri64 any hints on emb6 and lwip errors?

[miri64]

@BytesGalore @kYc0o can not reproduce, sadly :(

[BytesGalore]

@kYc0o @miri64 emb6 sets the samr21-xpro board as default [1]. using:

BOARD=native make

results in the expected behaviour, i.e. unsatisfied requirements and build fail [1] https://github.com/RIOT-OS/RIOT/blob/master/tests/emb6/Makefile#L5

[miri64]

The reason this application sets it as a default is because there is no native support for it.

[miri64]

(emb6 only has IEEE 802.15.4 support so porting it for native wouldn't make much sense anyway, at least until https://github.com/RIOT-OS/RIOT/pull/5582 is merged)

[miri64]

((also also: https://github.com/RIOT-OS/RIOT/issues/4672 needs to be solved, so auto-init also works for non GNRC-stacks))

[BytesGalore]

@miri64 yup I expected it to not compile and run but I expected a specific behaviour, i.e. failing with an unsatisfied requirements warning.

Similar applies for lwip the iotlab-m3 board is set when BOARD is not provided [1]. using:

BOARD=native make

results in a successful build and allows to start the compiled test-application (though untested yet)

[1] https://github.com/RIOT-OS/RIOT/blob/master/tests/lwip/Makefile#L3

[kYc0o]

So what's the conclusion on this? Can we say that the test ran successfully?

[BytesGalore]

@kYc0o don't know, I guess lwip is not supposed to run on native but it successfully builds and the test-application can be started. Beside this uncertainty I would say the test ran successfully.

btw. Since we have no release branch yet, I performed the tests on: https://github.com/RIOT-OS/RIOT/commit/b5cb68bd6673abe02166f8af7782aa2745ca08cd

[kYc0o]

Ok great. Yes that was the last commit before I started the tests too.

[miri64]

lwIP is supposed to run on native! So if it runs what is the problem? :/

[BytesGalore]

@miri64 if its supposed to run on native I propose that we change the default set BOARD accordingly.

So if it runs what is the problem?

When I start it on native ifconfig states:

ifconfig
ET_00:  inet6 fe80::d8c3:d5ff:fe46:7582

but sending some packet fails:

> udp send 2001::1 8080 1
udp send 2001::1 8080 1
could not send
> ip send 2001::1 inet6 10
ip send 2001::1 inet6 10
could not send

probably I just failed to use it properly.

[BytesGalore]

ok, I failed :blush:

> udp send fe80::1 8080 1
udp send fe80::1 8080 1
Success: send 1 byte to [fe80::1]:8080)
> ip send fe80::1 inet6 1
ip send fe80::1 inet6 1
Success: send 1 byte to fe80::1 (next header: 0)

[miri64]

:+1:

[BytesGalore]

I propose https://github.com/RIOT-OS/RIOT/pull/5698 to change the default BOARD to native for lwip