Re: Mass Bug Filing: Missing Build-Depends: graphviz
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On Tue, Mar 22, 2016 at 09:26:45PM +0100, Santiago Vila wrote:
> I think the issue is not really whether HAVE_DOT=yes is good or not
> in general, but whether this is an issue that should be decided on a
> per package basis or not.
I agree, that is what we are discussing. I'm saying that if a package does not
specify what it should do, we should generate better documentation. That means
setting HAVE_DOT=yes. I don't know upstream's reasoning for not making that
the default. It would be good to ask them; they could have a good reason.
They may also have a reason that is not relevant for Debian (such is "on
Windows dot is hard to install and we want things to work easily everywhere").
> Packages using doxygen usually have a file in which they put options
> like this. From 57 packages I checked, 9 of them deliberately set the
> option to yes. The upstream default is no, so this means the other 48
> (sort of) deliberately decided not to use dot (by not overriding the
> variable from its upstream default).
I disagree with this conclusion. I have some (private) packages which generate
doxygen documentation (and they depend on graphviz), and I used the defaults as
initial values for my configuration. I did this on a Debian system, so it uses
Debian's defaults.
Your assumption is that most upstreams use a non-Debian (based) release of
doxygen. I think it is more likely that that assumption is wrong, than that so
many upstreams think inheritance graphs are useless.
> But we are currently overriding all of them.
Just the ones that don't specify anything, right? I think it is reasonable
that Debian makes the choice for them, and generating better documentation
seems like a good choice to me.
Thanks,
Bas
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iQIcBAEBAgAGBQJW8cBmAAoJEJzRfVgHwHE6rHYP/A2jcgLASXeSfJZ9maMoUrke
YNT7Qqh9kCWj8OOXVZnTIAc9sTRzAcpfR4k+myw4UwliivFFzhVMqcPfYnnRLfAs
zixCdPzMtFlNCrjX9AUeqXHu73u9+6B9sJQqer4jv7Aa0xPEC7sI1fMJh8Zvck8J
hRBLspgUl+4nechFieIotkpLrNRQoD0DJlpRByjnZfA5660E71fdEeLxvUR9cHKG
6cEH4QjCw3qHdwxJhxSNsB08oUOL0YjFoFT6tui8YVAO+AvnZJjwF4VNk81ZI8WX
MdAybwQWgERLcbjUZ9Gn6kQwuF5y/e/zaqRmYDWCzFxU+WErNy+p93Qfk1JFGAA8
tpkaMEJBDPK3uNJm9jmFwfzE8z0SNFKrkJcHTidGSFkqRRsvjomxkgmHAJYa4KAK
sYJ65mLwZkrvp5aUTqTJrRoK7ewpus8mCAZHLAKP/kkgqVThNaXEWmkBzSXdiOBx
dkv1zSkmyTY2RjPUmNt8jvKj2rh5YvDNm+0fPBkt+LXGmKp5t/Ux+/omGWHJy/Hj
GXx78nhLGIQP2fnBBM7n1KB5LGszqpMYIs8YhIEYwHSDWrzoPUFvyb9WI+3K+Sdg
+WKAVWgJ/sODjHngzQdpkHYjfORY2NSBMjQQvqEEYCro3YJriubL6AIY9q3QgUvo
Qc26EKLaDUrysP+qqf2s
=6lUh
-----END PGP SIGNATURE-----
Reply to: