[PATCH] Enable automake silent rules

Brian Inglis Brian.Inglis@SystematicSw.ab.ca
Wed Dec 8 06:26:30 GMT 2021

On 2021-12-07 13:03, C Howland wrote:
> On Tue, 7 Dec 2021 at 14:51, Mike Frysinger wrote:
>> On 07 Dec 2021 11:10, C Howland wrote:
>>> Should this really be made the new default?

>> yes

>>> I personally don't want it as a default because I want to see 
>>> command line arguments. Easy for me to get different, but only
>>> knowing how. That leads to the next question.

>> then use the standard `make V=1` or `./configure --disable-silent-rules`

Does "$ export V=1" provide the OP's preferred behaviour?

>>> Given that silent were made default, it is a big change in terms
>>> of what people see.  How would you know to change it? (New
>>> feature, most people are far from experts in automake.)
>>> Therefore, the directions most definitely would need something
>>> added that explains it. (Without knowing about it, I can see it
>>> taking someone a long time to chase down why they suddenly lost
>>> the make output.) Something along the lines of Jon's explanation
>>> in the patch email needs to end up in README and probably also 
>>> printed by a configure help.

>> you can make this argument about any change.  we shouldn't be paralyzed by
>> historical behavior and never make progress.  the automake manual provides
>> background info about why chatty make is not a good thing.

> I did not at all say that we should not do it.  I just pointed out there
> are reasons some people might not want it as the new default, while
> acknowledging that others might (e.g you and Corinna) and that it is a very
> visible one that needs additional consideration.

>> https://www.gnu.org/software/automake/manual/automake.html#Silencing-Make
>> the option is standard in the GNU ecosystem and adoption rate is only going
>> up.  it's already in the --help of configure.  no one has to know or be an
>> expert in automake to find it.

> OK, good. But this underscores the need to know something for a change
> for such a large impact.  (Why would one look under configure when make
> output has a very large change?)

>> adding the option to the newlib/README sounds reasonable since it already
>> has a bunch of these. want to send a patch ?

> The whole point of my response is that a README entry should be part of the
> patch to begin with--you should be supplying it. (I didn't know about the
> help in configure, what else would I miss?)

Try "$ less +/^Optional .../configure" to see Optional Features 
--enable-/--disable- and Optional Packages --with-/--without- where 
available and supported.

As with the best configure, build, and test suites, most maintainers 
would ideally like to see an options summary, build and test command 
summaries e.g CC c.c/CCLD/GEN/PASS/FAIL c to show progress, with the 
full commands and paths on any failures, and a results summary.

Take care. Thanks, Brian Inglis, Calgary, Alberta, Canada

This email may be disturbing to some readers as it contains
too much technical detail. Reader discretion is advised.
[Data in binary units and prefixes, physical quantities in SI.]

More information about the Newlib mailing list