-----BEGIN PGP SIGNED MESSAGE-----
thanks for your comprehensive reply.
About the "mistaken" part (I should have explained myself better here): I
got the impression that a YML-like formatting of the spec files is one of
the next steps in RPM evolution. Since I pretty much like RPM, I decided to
go with that and tried to apply YML formatting rules to my spec files. "Am I
mistaken" just meant "did I get the wrong impression about the direction of
RPM spec file markup development?" :-)
On Sunday 10 May 2009, Jeff Johnson <firstname.lastname@example.org> wrote:
> On May 10, 2009, at 3:19 PM, Eric MSP Veith wrote:
> > -----BEGIN PGP SIGNED MESSAGE-----
> > Hash: SHA1
> > Hello,
> > since a while rpm supports spec files that do not only look like
> > YAML files,
> > but indeed are. At least as far as I know.
> > %prep
> > %setup -q
> > Becomes:
> > %prep
> > %setup -q
> > ... with proper indentation.
> FYi: Try
> rpm -qp --yaml foo*.src.rpm
> rpm -qp --yaml --specfile foo.spec
> for as close as rpm has come to YAMLspec!
> Yes, the indentation could be automated.
> > I've been trying to adopt this to my own projects. (I actually just
> > began
> > re-formatting my old specfiles.) What I found out is that it really
> > works in
> > the script parts, a thing that broke the spec file parser before.
> > However,
> > the description part still doesn't seem to get it, or I am making
> > mistakes.
> > When I write something like:
> > %description
> > This is some indented text for this package.
> > It shouldn't look like this when using "rpm -qi PKG".
> > The description is indented, too, when I run "rpm -qi PKG". While I
> > understand that there's no good way to have YML-like indentation *and*
> > preserve text formatting that might have been done in the
> > description part,
> > I long for a real YML-ish look of my spec files. :-)
> > Is there any way to do it right?
> With markup like XML/YAML, "right" is pretty much an aesthetic
> rather than engineering POV.
> But what is rate-limiting for RPM + YAML is the bikeshed discussions re
> what is "aesthetic" and what is "ugly".
> For example, I'm actively (for 9+ months) shopping
> markup that captures the same information as what
> is permitted in %trigger* sections in spec files.
> The trigger sections include multiple dependencies, as well
> as the usual -p and -n scriptlet flags, so example YAML markup
> for %triggers would generalize to all spec file sections rather
> quickly and easily.
> So if you are inclined, take a stab at example YAML markup
> for %trigger* sections, and send along please. I'll take a stab at
> a proof-of-concept parser for the markup using /etc/rpm/sysinfo/
> If/when I'm able to fire triggers from external configuration, then
> its time to start on the larger job of alternatives to *.spec file
> Note that the same YAML parser and markup would be used for both
> externally configured triggers and spec file parsing (in case the
> connection is not obvious).
> > Or am I perhaps totally mistaken regarding the YML thing?
> Not sure what "mistaken" means, since there has been almost
> no discussion about RPM + YAML.
> But IMHO, YAML is data focussed and human readable markup,
> very much not true for XML (which is syntax focussed) or JSON
> (which is hardly human readable, its just fancy CSV in most usage
> cases I've seen).
> Note that syck was re-added (it was in rpm-4.4.9 but too fragile
> and flimsy to carry forward to @rpm5.org) last month. syck adds
> a YAML parser to rpm, so far all the --yaml spewage is just
> 73 de Jeff
> RPM Package Manager http://rpm5.org
> User Communication List email@example.com
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
-----END PGP SIGNATURE-----
Received on Sun May 10 21:46:02 2009