Nanoseconds support request

[azmeuk]

Hello, I am a contributor of python-libfaketime (but no maintainer). This is a project that wraps libfaketime in python so it can be used in unit tests suits.

Testing for nanoseconds is not uncommon in unit testing, so python-libfaketime patches libfaketime to bring support for it:

diff --git a/src/libfaketime.c b/src/libfaketime.c
index e632395..09d9019 100644
--- a/src/libfaketime.c
+++ b/src/libfaketime.c
@@ -2384,10 +2384,16 @@ static void parse_ft_string(const char *user_faked_time)
       user_faked_time_tm.tm_isdst = -1;
       nstime_str = strptime(user_faked_time, user_faked_time_fmt, &user_faked_time_tm);

+      /* the actual format has a " %f" appended. Parse out the microseconds. */
+      char nanosecond_str[7];
+      memcpy(&nanosecond_str, user_faked_time + 20, 6);
+      nanosecond_str[6] = '\0';
+      int nanoseconds = atoi(nanosecond_str) * 1000;
+
       if (NULL != nstime_str)
       {
         user_faked_time_timespec.tv_sec = mktime(&user_faked_time_tm);
-        user_faked_time_timespec.tv_nsec = 0;
+        user_faked_time_timespec.tv_nsec = nanoseconds;

         if (nstime_str[0] == '.')
         {
-- 
2.45.0

And then python-libfaketime passes the time to mock to libfaketime with the expected format.

I thought this would be better if nanosecond resolution was implemented the right way here instead of a downstream patch.

What do you think?

[azmeuk]

Nevermind, I could actually do what I intended with FAKETIME_FMT = '%Y-%m-%d %T.%f'.