journalctl -o short-iso[-precise]: timezone as +02:00 instead of +0200 by nabijaczleweli · Pull Request #29134 · systemd/systemd

nabijaczleweli · 2023-09-08T15:32:26Z

Before:

$ journalctl -u r710_fan_controller -o short-iso -n10
2023-09-08T17:20:30+0200 tarta r710_fan_controller.sh[20911]: Highest drive temperature: 43 Celcius.
2023-09-08T17:20:47+0200 tarta r710_fan_controller.sh[20911]: Changing to fan level 4.
2023-09-08T17:20:47+0200 tarta r710_fan_controller.sh[20911]: Highest CPU core temperature: 36 Celcius.
2023-09-08T17:20:47+0200 tarta r710_fan_controller.sh[20911]: Highest drive temperature: 44 Celcius.
2023-09-08T17:22:28+0200 tarta r710_fan_controller.sh[20911]: Changing to fan level 3.
2023-09-08T17:22:28+0200 tarta r710_fan_controller.sh[20911]: Highest CPU core temperature: 35 Celcius.
2023-09-08T17:22:28+0200 tarta r710_fan_controller.sh[20911]: Highest drive temperature: 43 Celcius.
2023-09-08T17:22:56+0200 tarta r710_fan_controller.sh[20911]: Changing to fan level 4.
2023-09-08T17:22:56+0200 tarta r710_fan_controller.sh[20911]: Highest CPU core temperature: 35 Celcius.
2023-09-08T17:22:56+0200 tarta r710_fan_controller.sh[20911]: Highest drive temperature: 44 Celcius.
$ journalctl -u r710_fan_controller -o short-iso-precise -n10
2023-09-08T17:20:30.655028+0200 tarta r710_fan_controller.sh[20911]: Highest drive temperature: 43 Celcius.
2023-09-08T17:20:47.560211+0200 tarta r710_fan_controller.sh[20911]: Changing to fan level 4.
2023-09-08T17:20:47.560211+0200 tarta r710_fan_controller.sh[20911]: Highest CPU core temperature: 36 Celcius.
2023-09-08T17:20:47.560211+0200 tarta r710_fan_controller.sh[20911]: Highest drive temperature: 44 Celcius.
2023-09-08T17:22:28.387008+0200 tarta r710_fan_controller.sh[20911]: Changing to fan level 3.
2023-09-08T17:22:28.387008+0200 tarta r710_fan_controller.sh[20911]: Highest CPU core temperature: 35 Celcius.
2023-09-08T17:22:28.387008+0200 tarta r710_fan_controller.sh[20911]: Highest drive temperature: 43 Celcius.
2023-09-08T17:22:56.388270+0200 tarta r710_fan_controller.sh[20911]: Changing to fan level 4.
2023-09-08T17:22:56.388270+0200 tarta r710_fan_controller.sh[20911]: Highest CPU core temperature: 35 Celcius.
2023-09-08T17:22:56.388270+0200 tarta r710_fan_controller.sh[20911]: Highest drive temperature: 44 Celcius.

After:

$ build/journalctl -u r710_fan_controller -o short-iso -n10
2023-09-08T17:20:30+02:00 tarta r710_fan_controller.sh[20911]: Highest drive temperature: 43 Celcius.
2023-09-08T17:20:47+02:00 tarta r710_fan_controller.sh[20911]: Changing to fan level 4.
2023-09-08T17:20:47+02:00 tarta r710_fan_controller.sh[20911]: Highest CPU core temperature: 36 Celcius.
2023-09-08T17:20:47+02:00 tarta r710_fan_controller.sh[20911]: Highest drive temperature: 44 Celcius.
2023-09-08T17:22:28+02:00 tarta r710_fan_controller.sh[20911]: Changing to fan level 3.
2023-09-08T17:22:28+02:00 tarta r710_fan_controller.sh[20911]: Highest CPU core temperature: 35 Celcius.
2023-09-08T17:22:28+02:00 tarta r710_fan_controller.sh[20911]: Highest drive temperature: 43 Celcius.
2023-09-08T17:22:56+02:00 tarta r710_fan_controller.sh[20911]: Changing to fan level 4.
2023-09-08T17:22:56+02:00 tarta r710_fan_controller.sh[20911]: Highest CPU core temperature: 35 Celcius.
2023-09-08T17:22:56+02:00 tarta r710_fan_controller.sh[20911]: Highest drive temperature: 44 Celcius.
$ build/journalctl -u r710_fan_controller -o short-iso-precise -n10
2023-09-08T17:20:30.655028+02:00 tarta r710_fan_controller.sh[20911]: Highest drive temperature: 43 Celcius.
2023-09-08T17:20:47.560211+02:00 tarta r710_fan_controller.sh[20911]: Changing to fan level 4.
2023-09-08T17:20:47.560211+02:00 tarta r710_fan_controller.sh[20911]: Highest CPU core temperature: 36 Celcius.
2023-09-08T17:20:47.560211+02:00 tarta r710_fan_controller.sh[20911]: Highest drive temperature: 44 Celcius.
2023-09-08T17:22:28.387008+02:00 tarta r710_fan_controller.sh[20911]: Changing to fan level 3.
2023-09-08T17:22:28.387008+02:00 tarta r710_fan_controller.sh[20911]: Highest CPU core temperature: 35 Celcius.
2023-09-08T17:22:28.387008+02:00 tarta r710_fan_controller.sh[20911]: Highest drive temperature: 43 Celcius.
2023-09-08T17:22:56.388270+02:00 tarta r710_fan_controller.sh[20911]: Changing to fan level 4.
2023-09-08T17:22:56.388270+02:00 tarta r710_fan_controller.sh[20911]: Highest CPU core temperature: 35 Celcius.
2023-09-08T17:22:56.388270+02:00 tarta r710_fan_controller.sh[20911]: Highest drive temperature: 44 Celcius.

Since now

$ build/systemd-analyze timestamp 2023-09-08T17:20:30+0200
Failed to parse "2023-09-08T17:20:30+0200": Invalid argument
$ build/systemd-analyze timestamp 2023-09-08T17:20:30+02:00
  Original form: 2023-09-08T17:20:30+02:00
Normalized form: Fri 2023-09-08 17:20:30 CEST
       (in UTC): Fri 2023-09-08 15:20:30 UTC
   UNIX seconds: @1694186430
       From now: 5min ago

Also, idk how --truncate-newline made it there, but it shouldn't have. Moved it out, but there's a better place in the list for it probably.

poettering · 2023-09-08T17:41:56Z

hmm is the old syntax incorrect?

nabijaczleweli · 2023-09-08T17:50:08Z

We don't parse it, and it's not part of the RFC3339 profile of ISO8601. I guess it is an ISO8601, sure. But it's extremely anti-user to have a format that, for no good reason, cannot be fed back to --since &c., and we parse a format that's a byte away. Especially since you said that RFC3339 is the only sane profile of ISO8601.

poettering · 2023-09-08T18:40:59Z

ok make sense.

bluca · 2023-09-26T10:06:36Z

@nabijaczleweli this just needs a rebase to fix merge conflicts (and re-test just in case) then looks ready

…o values list

nabijaczleweli · 2023-09-26T12:47:57Z

rebased

poettering · 2023-09-27T21:27:47Z

-                        xsprintf(usec, "%06"PRI_USEC, display_ts->realtime % USEC_PER_SEC);
-                        memcpy(buf + 20, usec, 6);
+                        int8_t h = tm.tm_gmtoff / 60 / 60;
+                        int8_t m = labs((tm.tm_gmtoff / 60) % 60);


hmm, where does the int8_t type come from? this is so weird, a signed byte? we have never used such a type, as when put in registers or in the stack is upgraded to int anyway.

moreover tm_gmtoff actually uses long as type, hence the artificially shorter width is just weird and smeels a lot like accidental loss of precision.

Quite frankly the compiler should warn assigning the result of a funciton like labs() that operates on long ints onto an int8_t...

From stdint.h, like all the others. If I had a type that went [0, 60) I would've used it, but this uses int8_t because it's derived from voreutils vore-time which carries a generic version of date(1) %z, but there it's

std::int8_t hms[3]; hms[2] = std::abs(tm_gmtoff_for(tm, ts) % 60); hms[1] = std::abs((tm_gmtoff_for(tm, ts) / 60) % 60); hms[0] = tm_gmtoff_for(tm, ts) / 60 / 60;

poettering · 2023-09-27T21:33:10Z

+                        /* No usec in strftime, need to append */
+                        if (mode == OUTPUT_SHORT_ISO_PRECISE) {
+                                snprintf(buf + tail, ELEMENTSOF(buf) - tail, ".%06"PRI_USEC, display_ts->realtime % USEC_PER_SEC);
+                                tail += 7;


hmm, this is a bit icky, if the buffer was too short, then you just jump into the void here...

should at least have an assert(ELEMENTS(buf) - tail >= 7) first

yes and in the original if buf was smaller than 28 you also jumped into the void so

poettering · 2023-09-27T21:35:19Z

-                        memcpy(buf + 20, usec, 6);
+                        int8_t h = tm.tm_gmtoff / 60 / 60;
+                        int8_t m = labs((tm.tm_gmtoff / 60) % 60);
+                        snprintf(buf + tail, ELEMENTSOF(buf) - tail, "%+03"PRId8":%02"PRId8, h, m);


so, this is going to be one character longer than before. This must be reflected in the definition of FORMAT_TIMESTAMP_MAX

FORMAT_TIMESTAMP_MAX is

66 /* We assume a maximum timezone length of 6. TZNAME_MAX is not defined on Linux, but glibc internally initializes this 67 * to 6. Let's rely on that. */ 68 #define FORMAT_TIMESTAMP_MAX (3U+1U+10U+1U+8U+1U+6U+1U+6U+1U) 69 #define FORMAT_TIMESTAMP_RELATIVE_MAX 256U 70 #define FORMAT_TIMESPAN_MAX 64U

which very helpfully both references some glibc-private macro and doesn't say which format it's describing, but that works out to 38 and "2023-09-08T17:20:47.560211+0200" is 31, so no it doesn't

poettering · 2023-09-27T21:35:55Z

this really should have a test

nabijaczleweli · 2023-09-27T23:03:38Z

show_journal_entry() doesn't have tests, so unclear to me how (i'm not writing generic show_journal_entry() tests)

nabijaczleweli · 2023-09-27T23:05:12Z

#29356

When testing openshift/release#66397 - the pivot for the nodeimage to cs10 from cs9 seemed to work pretty well except one test which failed in the e2e with: ``` [sig-node] Managed cluster should verify that nodes have no unexpected reboots [Late] [Suite:openshift/conformance/parallel] ``` This test failed because the timestamp has changed to be more compliant with RFC339 (there was a PR in systemd about this a while back: systemd/systemd#29134). This PR fixes that test to also support this layout.

Related to: systemd/systemd#26639 Related to: systemd/systemd#29134 Signed-off-by: Volker Theile <[email protected]>

When testing openshift/release#66397 - the pivot for the nodeimage to cs10 from cs9 seemed to work pretty well except one test which failed in the e2e with: ``` [sig-node] Managed cluster should verify that nodes have no unexpected reboots [Late] [Suite:openshift/conformance/parallel] ``` This test failed because the timestamp has changed to be more compliant with RFC339 (there was a PR in systemd about this a while back: systemd/systemd#29134). This PR fixes that test to also support this layout.

github-actions Bot added documentation util-lib please-review PR is ready for (re-)review by a maintainer labels Sep 8, 2023

poettering reviewed Sep 8, 2023

View reviewed changes

Comment thread src/shared/logs-show.c Outdated

poettering reviewed Sep 8, 2023

View reviewed changes

Comment thread src/shared/logs-show.c Outdated

nabijaczleweli force-pushed the short-iso-timestamp branch 2 times, most recently from b91d3ed to 9705b89 Compare September 8, 2023 18:57

bluca added needs-rebase and removed please-review PR is ready for (re-)review by a maintainer labels Sep 18, 2023

YHNdnzj added this to the v255 milestone Sep 20, 2023

journalctl.1: move --truncate-newline to the options instead of the -…

bf83c67

…o values list

nabijaczleweli force-pushed the short-iso-timestamp branch from 9705b89 to ce49e60 Compare September 26, 2023 12:47

github-actions Bot added please-review PR is ready for (re-)review by a maintainer and removed needs-rebase labels Sep 26, 2023

bluca approved these changes Sep 26, 2023

View reviewed changes

bluca added journal good-to-merge/waiting-for-ci 👍 PR is good to merge, but CI hasn't passed at time of review. Please merge if you see CI has passed and removed please-review PR is ready for (re-)review by a maintainer labels Sep 26, 2023

YHNdnzj added the ci-fails/needs-rework 🔥 Please rework this, the CI noticed an issue with the PR label Sep 26, 2023

journalctl -o short-iso[-precise]: timezone as +02:00 instead of +0200

0693e6b

nabijaczleweli force-pushed the short-iso-timestamp branch from ce49e60 to 0693e6b Compare September 26, 2023 13:24

github-actions Bot removed the ci-fails/needs-rework 🔥 Please rework this, the CI noticed an issue with the PR label Sep 26, 2023

bluca merged commit 93bd6e3 into systemd:main Sep 27, 2023

github-actions Bot removed the good-to-merge/waiting-for-ci 👍 PR is good to merge, but CI hasn't passed at time of review. Please merge if you see CI has passed label Sep 27, 2023

poettering reviewed Sep 27, 2023

View reviewed changes

poettering mentioned this pull request Sep 27, 2023

[tracker] some merges need post-merge love #29355

Closed

poettering added the merged-but-needs-fixing label Sep 27, 2023

tompar28 mentioned this pull request Nov 16, 2023

RFE: journalctl -o short-rfc3339 / -o short-rfc3339-precise #26639

Closed

Prashanth684 mentioned this pull request Jun 26, 2025

OKD-267: Support RFC339 based timestamp layout for reboot tests openshift/origin#29942

Merged

votdev added a commit to openmediavault/openmediavault that referenced this pull request Nov 7, 2025

Adapt regex to new short-iso timestamp layout

824f0e0

Related to: systemd/systemd#26639 Related to: systemd/systemd#29134 Signed-off-by: Volker Theile <[email protected]>

rgwood mentioned this pull request Jan 9, 2026

Make date formatting more reliable rgwood/systemctl-tui#65

Merged

psviderski mentioned this pull request Apr 23, 2026

logging: client side psviderski/uncloud#283

Merged

Uh oh!

Conversation

nabijaczleweli commented Sep 8, 2023

Uh oh!

poettering commented Sep 8, 2023

Uh oh!

nabijaczleweli commented Sep 8, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

poettering commented Sep 8, 2023

Uh oh!

Uh oh!

Uh oh!

bluca commented Sep 26, 2023

Uh oh!

nabijaczleweli commented Sep 26, 2023

Uh oh!

poettering Sep 27, 2023

Choose a reason for hiding this comment

Uh oh!

nabijaczleweli Sep 27, 2023

Choose a reason for hiding this comment

Uh oh!

poettering Sep 27, 2023

Choose a reason for hiding this comment

Uh oh!

nabijaczleweli Sep 27, 2023

Choose a reason for hiding this comment

Uh oh!

poettering Sep 27, 2023

Choose a reason for hiding this comment

Uh oh!

nabijaczleweli Sep 27, 2023

Choose a reason for hiding this comment

Uh oh!

poettering commented Sep 27, 2023

Uh oh!

nabijaczleweli commented Sep 27, 2023

Uh oh!

nabijaczleweli commented Sep 27, 2023

Uh oh!

Reviewers

Assignees

Labels

Milestone

Development

Uh oh!

4 participants

nabijaczleweli commented Sep 8, 2023 •

edited

Loading