On Tue, Jul 28, 2020 at 16:14:03 +0200, Pavel Hrdina wrote:
On Tue, Jul 28, 2020 at 03:59:58PM +0200, Peter Krempa wrote:
> On Tue, Jul 28, 2020 at 15:57:16 +0200, Pavel Hrdina wrote:
> > On Tue, Jul 28, 2020 at 03:18:23PM +0200, Peter Krempa wrote:
> > > On Thu, Jul 16, 2020 at 11:59:17 +0200, Pavel Hrdina wrote:
> > > > Signed-off-by: Pavel Hrdina <phrdina(a)redhat.com>
> > > > ---
> > > > docs/Makefile.am | 26 ---------------------
> > > > scripts/meson-html-gen.py | 49
+++++++++++++++++++++++++++++++++++++++
> > > > scripts/meson.build | 1 +
> > > > 3 files changed, 50 insertions(+), 26 deletions(-)
> > > > create mode 100755 scripts/meson-html-gen.py
>
> [...]
>
> > > > +)
> > > > +
> > > > +html = subprocess.run(
> > > > + [args.xmllint, '--nonet', '--format',
'-'],
> > > > + input=html_tmp.stdout,
> > > > + stdout=subprocess.PIPE,
> > > > + stderr=subprocess.PIPE,
> > >
> > > We can then do this as separate stage.
> >
> > What do you mean by separate stage? Not following here.
>
> I was arguing that the reformatting done by xmllint can be done as a
> separate build target thus eliminating the need to have this script
> completely.
Right, I guess now it can be done like that since there is not the part
that modified the string directly in python. However, I would like to
keep it like this.
Well, and I don't like the extra python wrapper which obscures what is
happening. Arguably we could write everything ourselves and just invoke
a massive script from make/meson/whatever.
We will not have temporary files in the build directory. With two
targets the temporary file should not be removed otherwise ninja would
rebuild it every single time and that would cause to rebuild the
resulting HTML as well.
Umm there's plenty of temporary stuff in the build directory. I don't
think we care.