Crash on unexpected datetime format

[ericaltendorf]

Crash on unexpected datetime format

[altendky]

Seems ok in Linux in 3.5-3.9, fwiw.

Linux 3.5-3.9

$ python3.5 -c 'import time; f = "%a %b %d %H:%M:%S %Y"; d = "Fri Feb 12 16:02:56 2021"; r = time.strptime(d, f); print(r)'
time.struct_time(tm_year=2021, tm_mon=2, tm_mday=12, tm_hour=16, tm_min=2, tm_sec=56, tm_wday=4, tm_yday=43, tm_isdst=-1)

$ python3.6 -c 'import time; f = "%a %b %d %H:%M:%S %Y"; d = "Fri Feb 12 16:02:56 2021"; r = time.strptime(d, f); print(r)'
time.struct_time(tm_year=2021, tm_mon=2, tm_mday=12, tm_hour=16, tm_min=2, tm_sec=56, tm_wday=4, tm_yday=43, tm_isdst=-1)

$ python3.7 -c 'import time; f = "%a %b %d %H:%M:%S %Y"; d = "Fri Feb 12 16:02:56 2021"; r = time.strptime(d, f); print(r)'
time.struct_time(tm_year=2021, tm_mon=2, tm_mday=12, tm_hour=16, tm_min=2, tm_sec=56, tm_wday=4, tm_yday=43, tm_isdst=-1)

$ python3.8 -c 'import time; f = "%a %b %d %H:%M:%S %Y"; d = "Fri Feb 12 16:02:56 2021"; r = time.strptime(d, f); print(r)'
time.struct_time(tm_year=2021, tm_mon=2, tm_mday=12, tm_hour=16, tm_min=2, tm_sec=56, tm_wday=4, tm_yday=43, tm_isdst=-1)

$ python3.9 -c 'import time; f = "%a %b %d %H:%M:%S %Y"; d = "Fri Feb 12 16:02:56 2021"; r = time.strptime(d, f); print(r)'
time.struct_time(tm_year=2021, tm_mon=2, tm_mday=12, tm_hour=16, tm_min=2, tm_sec=56, tm_wday=4, tm_yday=43, tm_isdst=-1)

[altendky]

Likewise in Windows 3.5-3.9. No other inspiration at the moment. Also confirmed with romp across macOS as well.

Windows 3.5-3.9

PS C:\Users\sda> py -3.5 -c "import time; f = '%a %b %d %H:%M:%S %Y'; d = 'Fri Feb 12 16:02:56 2021'; r = time.strptime(d, f); print(r)"
time.struct_time(tm_year=2021, tm_mon=2, tm_mday=12, tm_hour=16, tm_min=2, tm_sec=56, tm_wday=4, tm_yday=43, tm_isdst=-1)

PS C:\Users\sda> py -3.6 -c "import time; f = '%a %b %d %H:%M:%S %Y'; d = 'Fri Feb 12 16:02:56 2021'; r = time.strptime(d, f); print(r)"
time.struct_time(tm_year=2021, tm_mon=2, tm_mday=12, tm_hour=16, tm_min=2, tm_sec=56, tm_wday=4, tm_yday=43, tm_isdst=-1)

PS C:\Users\sda> py -3.7 -c "import time; f = '%a %b %d %H:%M:%S %Y'; d = 'Fri Feb 12 16:02:56 2021'; r = time.strptime(d, f); print(r)"
time.struct_time(tm_year=2021, tm_mon=2, tm_mday=12, tm_hour=16, tm_min=2, tm_sec=56, tm_wday=4, tm_yday=43, tm_isdst=-1)

PS C:\Users\sda> py -3.8 -c "import time; f = '%a %b %d %H:%M:%S %Y'; d = 'Fri Feb 12 16:02:56 2021'; r = time.strptime(d, f); print(r)"
time.struct_time(tm_year=2021, tm_mon=2, tm_mday=12, tm_hour=16, tm_min=2, tm_sec=56, tm_wday=4, tm_yday=43, tm_isdst=-1)

PS C:\Users\sda> py -3.9 -c "import time; f = '%a %b %d %H:%M:%S %Y'; d = 'Fri Feb 12 16:02:56 2021'; r = time.strptime(d, f); print(r)"
time.struct_time(tm_year=2021, tm_mon=2, tm_mday=12, tm_hour=16, tm_min=2, tm_sec=56, tm_wday=4, tm_yday=43, tm_isdst=-1)

[ericaltendorf]

stacktrace from reporter:

    jobs = Job.get_running_jobs(dir_cfg['log'])
  File "/chia/plotman/job.py", line 66, in get_running_jobs
    jobs.append(Job(proc, logroot))
  File "/chia/plotman/job.py", line 135, in init
    self.init_from_logfile()
  File "/chia/plotman/job.py", line 156, in init_from_logfile
    self.start_time = datetime.strptime(m.group(1), '%a %b  %d %H:%M:%S %Y')
  File "/usr/lib/python3.8/strptime.py", line 568, in strptime_datetime
    tt, fraction, gmtoff_fraction = strptime(datastring, format)
  File "/usr/lib/python3.8/strptime.py", line 349, in strptime
    raise ValueError("time data %r does not match format %r" %
ValueError: time data 'Fri Feb 12 16:02:56 2021' does not match format '%a %b  %d %H:%M:%S %Y'

platform 5.8.0-43-generic #49-Ubuntu SMP Fri Feb 5 03:01:28 UTC 2021 x86_64 x86_64 x86_64 GNU/Linux

One thing that looks suspicious is the two spaces between %b and %d in the format string?

[johnj212]

Hi, I've tested and to reproduce the issue set the locale time is set to : LC_TIME=de_CH.UTF-8 it crashes, but when I set this to LC_TIME=en_US.UTF-8 no problem.

[ericaltendorf]

Thanks! I suspected it might have to do with locale. I'm guessing in de_CH the expectation for Friday abbreviation is 'Fr' for Freitag?

FWIW, we're reading from a file produced by a different program in a different code base (the chia plotter) so we have no real guarantee of the format or locale with which it was produced.

The right approach here is probably to do something fuzzier and best-effort to make as much sense of the log datetime as we can.

[loysollivier]

A workaround is to use:
export LC_TIME=en_US.UTF-8

[BasilHorowt]

The linked fix seems to be on development only, this issue can be closed once in main?