Re: do not install *.in files

[G. Branden Robinson]

At 2022-03-31T18:43:15+0200, Ingo Schwarze wrote:
> Hi Branden,
> 
> G. Branden Robinson wrote on Thu, Mar 31, 2022 at 11:33:52PM +1100:
> > At 2022-03-30T13:25:24+0200, Ingo Schwarze wrote:
> 
> >>  6. DOCFILES_NOINST needs to be added to EXTRA_DIST unconditionally.
> 
> > I believe Automake already has an idiom for this, the "noinst_" prefix.
> > https://www.gnu.org/software/automake/manual/html_node/Uniform.html
> 
> I wasn't aware of "noinst_", so i looked at the page you pointed to,
> and found that it says:
> 
>   The special prefix ‘noinst_’ indicates that the objects in question
>   should be built but not installed at all.
> 
> None of the *.in files we are talking about need to be *built*,
> all four of them are already checked into the git repo.

Right.  I managed to overlook that.

> Do my above findings imply that you are OK with my original patch?
> (appended again for convenience)

Yes, please go ahead.

> P.S.
> I dislike how these *.am files are typically first defining variables
> with ad hoc, unsystematic names (for example SBOXES_NORMAL_FILES,
> SBOXES_PROCESSEDEXAMPLEFILES), then copying the contents of
> these into systematically named variables (in these two examples,
> into dist_sboxestmac_DATA and nodist_sboxesprocessedexample_DATA,
> respectively).  That adds yet another, IMHO gratuitious layer of
> additional complexity with an additional risk of terminological
> confusion.  It might be better to define systematically named variables
> right away, use them throughout, and delete the unsystematic names,
> except in those cases where a name is unsystematic on purpose in order
> to not trigger any automake magic (for example SBOXES_EXAMPLEFILES).

I agree--I find they make the build process harder to understand.

> But cleaning up the redundant naming should probably not be mixed
> into this bugfix patch.

Also agreed.

Regards,
Branden

signature.asc
Description: PGP signature