Nicolas Boichat

(btw the empty `TZ` case is fixed in chrono, but we need to wait for a new release: https://github.com/chronotope/chrono/pull/1691)

I... also lost all context... and if it's not clear from my comments, I was also a bit confused a few weeks ago when I tried to look at this...

I think we also do not handle negative exponents correctly: `1.0e-1` is not parsed correctly.

This misses the `fuzz/Cargo.lock` update, but also see https://github.com/uutils/parse_datetime/issues/151

> The `iana-time-zone` crate offers this functionality. We could mention it in our documentation. Not completely though. On POSIX systems we should also look at `TZ` variable, and it seems...

Yeah, it's a red herring, that error is _meant_ to happen. You can get the exact some error output by adding `exit 1` towards the end of `test/misc/tee.sh`, so something...

Got verbose logs here: https://github.com/drinkcat/coreutils/actions/runs/14574570622/job/40877858265 , https://productionresultssa10.blob.core.windows.net/actions-results/82c4e74b-7898-41a8-bff9-0e4502de1c29/workflow-job-run-2edfd9fd-13f0-51db-b86a-dc1022807e62/logs/job/job-logs.txt?rsct=text%2Fplain&se=2025-04-21T15%3A31%3A11Z&sig=vZcj8t4%2BtyQU4%2FckeoXa1a7itxw0ss%2FLX8sgHKUzsJQ%3D&ske=2025-04-21T23%3A22%3A59Z&skoid=ca7593d4-ee42-46cd-af88-8b886a2f84eb&sks=b&skt=2025-04-21T11%3A22%3A59Z&sktid=398a6654-997b-47e9-b12b-9515b896b4de&skv=2025-01-05&sp=r&spr=https&sr=b&st=2025-04-21T15%3A21%3A06Z&sv=2025-01-05 In a gist for easier parsing... https://gist.github.com/drinkcat/b406f61bed0422985015ff58fff0cc01 This seems to be the error: ``` 2025-04-21T14:20:20.3132187Z + read_fifo 2025-04-21T14:20:20.3132294Z + timeout 10 dd count=1...

Reproducer `tee-test.sh`: ``` # set -x rm -f fifo mkfifo fifo # Previous test, not 100% necessary but makes errors more frequent (timeout 10 dd count=1 if=fifo of=/dev/null status=none; echo...

`strace` is interesting. GNU ``` poll([{fd=0, events=POLLIN|POLLRDBAND}, {fd=1, events=POLLRDBAND}], 2, -1) = 1 ([{fd=0, revents=POLLIN}]) read(0, "y\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\n"..., 8192) = 8192 write(1, "y\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\ny\n"..., 8192) = 8192 poll([{fd=0, events=POLLIN|POLLRDBAND}, {fd=1, events=POLLRDBAND}], 2,...

Goodness, even a rebase was challenging, ended up squashing 2 commits together to make the rebase _slightly_ easier, preserved author info in commit message. #8329. I'll try to apply the...