Re: bug#54764: encode-time: make DST and TIMEZONE fields of the list argument optional ones

[Paul Eggert]

On 4/20/22 00:23, Eli Zaretskii wrote:
> > Date: Tue, 19 Apr 2022 15:22:29 -0700

> Thanks, the test-gettime-res test says "gettime_res returned 625000
> ns", which is a strange number: it doesn't fit any MS-Windows system
> time resolution figure I know about.  Do you happen to know what does
> this number represent, and why it is the result of gettime-res.c when
> it runs on MS-Windows?

It comes from current_timespec samples taken by gettime_res. Evidently 
something is going wrong, either in gettime_res or in current_timespec.

I stared at the code a bit and see one possible problem, which I fixed 
by installing the attached patch into Gnulib. I then generated a new 
test-gettime-res.tgz compressed tarball (also attached); could you give 
it a try?

This tarball is the result of running ./gnulib-tool as before, except I 
added an extra print statement executed if you pass an extra argument to 
that test program. So if you run the test and it's still outputting an 
outlandish value for the resolution, please run the command:

gltests/test-gettime-res x

and let's take a look at its (long) debugging output.

0001-gettime-res-more-robust-sampling.patch
Description: Text Data

test-gettime-res.tgz
Description: application/compressed-tar