board/native: set TZ to UTC by default by kfessel · Pull Request #20023 · RIOT-OS/RIOT

kfessel · 2023-10-26T14:30:48Z

Contribution description

board/native: set env TZ to UTC by default
also fix the failing minema test - that assumend to be run in Berlin/Paris/Hamburg/Rom/Wien

most (all) embedded libc do not care for TZ and time functions don't do timezone conversion,
making native use UTC for time.h functions does the same

you can still force localtime timezone

TZ=":/etc/localtime" <elf_file> or
TZ=":/etc/localtime" make term or test or
export TZ=":/etc/localtime" or

or another timezone

export TZ="LEET+1:33:7" (timezone name and offset from UTC optional DST
see man timezone for TZ specifier (not all of it is supported by new/picolib(c)

Testing procedure

run the minema test in Berlin and see that the time stamp is now UTC

Issues/PRs references

#20018 #20005

kfessel · 2023-10-26T14:33:45Z

setting the enviroment is actually the suggested portable solution by gnu libc

https://www.gnu.org/software/libc/manual/html_node/Broken_002ddown-Time.html#index-timegm

benpicco · 2023-10-26T14:46:59Z

I assume this will also affect the time returned by periph_rtc?

Then this will cause confusion because the time 'inside' native does no longer line up with the system time.
And one might want to correlate an external action (e.g. script) with a log timestamp on native when running a simulation.

benpicco

> rtc gettime
2023-10-26 14:53:09

$ date
Do 26. Okt 16:53:17 CEST 2023

I think this will lead to confusion.

benpicco · 2023-10-26T14:51:36Z

boards/native/board_init.c

 */
 void board_init(void)
 {
+    setenv("TZ", "UTC", 0);


Let's guard this behind a pseudo-module or config option that gets selected by Makefile.tests_common

how did you set the rtc ? (we should probably start it at 0 where most of our boards start) :)

I did not set it at all, it's synced with the system time by default.

I like sane defaults and the time functions are insane and even more so if timezone things are happening lets ditch then in a more sane direction by default.

if you like your localtime to much TZ=/etc/localtime <elf_file> or TZ=/etc/localtime make term or test
is an option
using mktime on native is a gamble otherwise maybe we should run a Linter that warns and errors about that.

IMHO native showing the same time as the host system is a sane default.
Why not make it a config option so we can have predictive behavior for tests that need it without forcing everyone to accept UTC as the one true time?

cause in networking our timestamps are usually UNIX time - which is kind of UTC.

since RIOT is an IoT OS we should probably be as compatible as we can.

https://www.sourceware.org/git/?p=newlib-cygwin.git;a=blob;f=newlib/libc/time/tzvars.c;h=d1f17649b7112ea520032f2bfc109191febb5a2c;hb=HEAD#l5

riot-ci · 2023-10-26T19:32:45Z

Murdock results

✔️ PASSED

fd11e3e board/native: add RTC at UTC documentation

Success	Failures	Total	Runtime
10295	0	10295	09m:06s

Artifacts

Documentation preview

maribu · 2023-10-30T16:40:16Z

Maybe we could have the periph_rtc start at RIOT_EPOCH on a cold boot similar to how it is on real hardware?

Not being slightly off but a lot off may look more intentional to a user and is more consistent to the behavior of real hardware.

Having the timezone of the host not leak into the native board is IMO still a good idea, not only but especially for tests.

kfessel · 2023-10-30T17:01:55Z

#20031

mguetschow · 2025-04-03T11:25:23Z

What's the status on this one? Mind looking into it before the upcoming release?

kfessel · 2025-04-03T11:49:51Z

rebased to resolve merge confict

benpicco · 2025-04-03T11:51:00Z

I blocked that because back then I was doing a lot of tests with native and timestamps in the logs matching the real system time really helped with debugging.

I'm not doing that now, so if you really think running native at UTC time has some kind of advantage that is greater than the confusion it will introduce for native users, go ahead.

kfessel · 2025-04-04T11:15:14Z

Why not add an
export TZ := UTC
to the two affected test apps?

because it effects the application that use mktime and co and is not isolated to tests
the test are just showing the issue

benpicco · 2025-04-04T11:53:12Z

boards/common/native/doc.md

+    - execute `export TZ=":/etc/localtime"` prior `make term` in the very shell you intend to run make term in
+    - add `export TZ=":/etc/localtime"` to your `~/.profile`
+    - run `TZ=":/etc/localtime" make term`
+    - set `TZ` in riot application picolibc and newlib are able to respect that setting if `+/-hh:mm:ss`


Can't we just add this to pyterm so it's less user hostile?

i will not: i think doing things wrong (make embedded things that assume no timezone behave wrong) should not be the default.

i think mktime behaving different on native and on embedded is much more user hostile

if one wants mktime behave like it is in a timezone there are easy ways in application or environment
and i put them in the documentation

mguetschow · 2025-04-07T14:30:04Z

boards/common/native/doc.md

+- RTC: host-time but at 'UTC' to behave consistent with embedded systems if TZ is set that timezone will be used.
+    - set `TZ` environment variable e.g.:
+    - execute `export TZ=":/etc/localtime"` prior `make term` in the very shell you intend to run make term in
+    - add `export TZ=":/etc/localtime"` to your `~/.profile`
+    - run `TZ=":/etc/localtime" make term`
+    - to set set `TZ` in riot application as picolibc and newlib are able to respect that setting if `NAME+/-hh:mm:ss`
+      (-time is added, +time is subtracted from UTC (+ pointing WEST England towards America) see `man timezone`
+      eg.: `IST-5:30` or `TZ="NZST-12NZDT-13,M10.1.0,M3.3.0"` or `TZ="CET-1CEST-2,M3.5.0,M10.5.0"` or `TZ="ACST-9:30ACDT-10:30,M10.1.0,M4.1.0"`


Thanks for the detailed documentation! It's a lot though now, and I'd say it warrants a separate subsection.

boards/common/native/doc.md

mguetschow

Thanks for the documentation, some formulation suggestions below.

boards/common/native/doc.md

mguetschow

One last round, then I'd be satisfied :)

boards/common/native/doc.md

mguetschow

LGTM now, thanks!

kfessel requested review from MichelRottleuthner, aabadie, fjmolinas, leandrolanzieri and smlng as code owners October 26, 2023 14:30

github-actions bot added Platform: native Platform: This PR/issue effects the native platform Area: tests Area: tests and testing framework Area: boards Area: Board ports labels Oct 26, 2023

kfessel requested a review from maribu October 26, 2023 14:31

kfessel requested a review from benpicco October 26, 2023 14:41

maribu approved these changes Oct 26, 2023

View reviewed changes

benpicco previously requested changes Oct 26, 2023

View reviewed changes

kfessel added the CI: ready for build If set, CI server will compile all applications for all available boards for the labeled PR label Oct 26, 2023

kfessel mentioned this pull request Oct 26, 2023

tests/pkg/minmea: fixing RMC timestamp #20018

Merged

kfessel requested a review from bergzand October 27, 2023 10:16

kfessel force-pushed the p-nativ-TZ-UTC branch from 2423137 to 842a74e Compare October 30, 2023 16:21

github-actions bot removed the Area: tests Area: tests and testing framework label Oct 30, 2023

kfessel mentioned this pull request Oct 30, 2023

tests/test-rtc: test retrieval of unix-time #20031

Merged

board/native: set TZ to UTC by default

0574878

kfessel force-pushed the p-nativ-TZ-UTC branch from 842a74e to fc4e189 Compare April 3, 2025 11:49

github-actions bot removed the Platform: native Platform: This PR/issue effects the native platform label Apr 3, 2025

mguetschow mentioned this pull request Apr 3, 2025

.github/labeler: recognize boards/common/native as native #21351

Merged

board/native: add information about the use of UTC

ffcfc56

kfessel force-pushed the p-nativ-TZ-UTC branch from fc4e189 to ba80533 Compare April 4, 2025 11:27

kfessel requested a review from jia200x as a code owner April 4, 2025 11:27

github-actions bot added Platform: native Platform: This PR/issue effects the native platform Area: doc Area: Documentation labels Apr 4, 2025

kfessel force-pushed the p-nativ-TZ-UTC branch from ba80533 to 71b6a72 Compare April 4, 2025 11:50

benpicco reviewed Apr 4, 2025

View reviewed changes

kfessel force-pushed the p-nativ-TZ-UTC branch 4 times, most recently from ae55ecc to 933a498 Compare April 4, 2025 13:18

mguetschow reviewed Apr 7, 2025

View reviewed changes

kfessel force-pushed the p-nativ-TZ-UTC branch 3 times, most recently from 55466d4 to f73e6cc Compare April 8, 2025 08:01

mguetschow reviewed Apr 9, 2025

View reviewed changes

boards/common/native/doc.md Outdated Show resolved Hide resolved

mguetschow reviewed Apr 9, 2025

View reviewed changes

boards/common/native/doc.md Outdated Show resolved Hide resolved

boards/common/native/doc.md Outdated Show resolved Hide resolved

boards/common/native/doc.md Outdated Show resolved Hide resolved

boards/common/native/doc.md Outdated Show resolved Hide resolved

mguetschow requested changes Apr 10, 2025

View reviewed changes

boards/common/native/doc.md Outdated Show resolved Hide resolved

boards/common/native/doc.md Outdated Show resolved Hide resolved

boards/common/native/doc.md Outdated Show resolved Hide resolved

mguetschow reviewed Apr 10, 2025

View reviewed changes

boards/common/native/doc.md Outdated Show resolved Hide resolved

kfessel force-pushed the p-nativ-TZ-UTC branch 2 times, most recently from c1ffdbb to 8f8c136 Compare April 11, 2025 08:38

board/native: add RTC at UTC documentation

fd11e3e

kfessel force-pushed the p-nativ-TZ-UTC branch from 8f8c136 to fd11e3e Compare April 11, 2025 08:56

mguetschow approved these changes Apr 11, 2025

View reviewed changes

kfessel added this pull request to the merge queue Apr 14, 2025

github-merge-queue bot removed this pull request from the merge queue due to failed status checks Apr 14, 2025

mguetschow added this pull request to the merge queue Apr 14, 2025

Merged via the queue into RIOT-OS:master with commit 103ab62 Apr 14, 2025
26 checks passed

Conversation

kfessel commented Oct 26, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Contribution description

Testing procedure

Issues/PRs references

Uh oh!

kfessel commented Oct 26, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

benpicco commented Oct 26, 2023

Uh oh!

benpicco left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

riot-ci commented Oct 26, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Murdock results

Artifacts

Uh oh!

maribu commented Oct 30, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

kfessel commented Oct 30, 2023

Uh oh!

mguetschow commented Apr 3, 2025

Uh oh!

kfessel commented Apr 3, 2025

Uh oh!

benpicco commented Apr 3, 2025

Uh oh!

kfessel commented Apr 4, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

kfessel Apr 4, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

mguetschow left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

mguetschow left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

mguetschow left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

kfessel commented Oct 26, 2023 •

edited

Loading

kfessel commented Oct 26, 2023 •

edited

Loading

benpicco left a comment •

edited

Loading

riot-ci commented Oct 26, 2023 •

edited

Loading

maribu commented Oct 30, 2023 •

edited

Loading

kfessel commented Apr 4, 2025 •

edited

Loading

kfessel Apr 4, 2025 •

edited

Loading