[m-rev.] for review: derive getopt.m and getopt_io.m from same template

[Julien Fischer]

On Sun, 20 Sep 2020, Zoltan Somogyi wrote:

>
> 2020-09-19 13:53 GMT+10:00 "Julien Fischer" <jfischer at opturion.com>:
>> After we moved to git I kept the test_mercury script around for
>> reference purposes; it can probably be deleted now as the stuff managing
>> the test server is in now in the mercury-testing repository.
>
> Just to be sure: you mean github.com/Mercury-Language/testing?

Yes.

>> It should be a small adjustment, so just give me a heads up when you
>> commit.
>
> I don't want to screw up any cron jobs, so I am attaching the diff now,
> and leave it up to you whether you either
>
> - add the build_srcdist2 script, and the tar2 target in the diff, next to the
>  existing script and target, or
> - replace both the script and the target.

Ok, I will try it out this evening.

> On my laptop, the new script and target work, but of course it has to work
> on the machine that generates the rotds as well.
>
> The new script build the tar file uncompressed, so that it can be compressed
> twice by gzip and xz. I have no idea whether this is faster or slower than
> invoking tar twice, but never writing an uncompressed tar file to the filesystem,
> in part because this depends on whether the file is on a magnetic disk vs on an SSD.

The majority of the time is spent by xz when compressing; the rest
doesn't matter.

Julien.