Hi,
On 2023-11-08 16:19:51 +0100, Peter Eisentraut wrote:
> On 08.11.23 13:55, Christoph Berg wrote:
> > Re: Peter Eisentraut
> > > > If the problem is broken doc patches, then maybe a solution is to
> > > > include the `xmllint --noout --valid` target in whatever the check-world
> > > > equivalent is for meson. Looking at doc/src/sgml/meson.build, we don't
> > > > seem to do that anywhere. Doing the no-output lint run is very fast
> > > > (375ms real time in my machine, whereas "make html" takes 27s).
> > >
> > > This would be a start, but it wouldn't cover everything. Lately, we require
> > > id attributes on certain elements, which is checked on the XSLT level.
> >
> > I'd think there should be a catchy "make check-world"-equivalent that
> > does run all reasonable check that we can tell people to run by
> > default. Then if that takes too long, we could still offer
> > alternatives that exclude some areas. If it's the other way round,
> > some areas will never be checked widely.
>
> I think we could build doc/src/sgml/postgres-full.xml by default. That
> takes less than 0.5 seconds here and it's an intermediate target for html
> and man.
That does require the docbook dtd to be installed, afaict. I think we would
need a configure test for that to be present if we want to build it by
default, otherwise we'll cause errors on plenty systems that don't get them
today. The docbook dts aren't a huge dependency, but still. Some OSs might
not have a particularly install source for them, e.g. windows.
I don't think that'd detect the missing ids?
Greetings,
Andres Freund