[m-rev.] for review: include program name and date/time in deep prof data files' names

[Zoltan Somogyi]

On Sat, 5 Oct 2024 15:06:02 +1000, Julien Fischer <jfischer at opturion.com> wrote:

> On Fri, 4 Oct 2024 at 21:37, Zoltan Somogyi <zoltan.somogyi at runbox.com> wrote:
> >
> > better names than Deep.data for profiling data.
> 
> I assume the rest of that was accidently not included.

Correct: an initial "Use" was missing.

> > +            snprintf(date_name, MR_FILENAME_BUF_LEN, "_%d_%d_%d_%d_%d_%d",
> > +                tm->tm_year + 1900, tm->tm_mon + 1, tm->tm_mday,
> > +                tm->tm_hour, tm->tm_min, tm->tm_sec);
> > +            strncat(data_filename, date_name, MR_FILENAME_BUF_LEN);
> > +            strncat(procrep_filename, date_name, MR_FILENAME_BUF_LEN);
> 
> I would be inclined to borrow the 'T' separator used in the ISO 8061 formats
> and use that to separate the date and time components.

I am pretty sure that we cannot assume universal knowledge
of that ISO standard, since I, for example, do not know it.

> The timestamp could
> either be:
> 
>    YYYMMDD'T'HHMMSS (e.g. 20241005T145900)
> 
> which is compact,

Compactness is good, but only if it does not compromise readability.
In this case, it does.

> or you could use '-' to separate the date and time components.
> 
>    YYYY-MM-DD'T'HH-MM-SS (e.g. 2024-10-05T14-59-00)
> 
> which is a bit more readable.

I chose names like this:

mercury_compile_on_2024-10-05_at_08-43-34.data

I followed all your other suggestions. The updated log
and diff are attached. I committed the diff.

> The rest looks fine.

Thank you.

Zoltan.