[Libosinfo] [PATCH 4/8] installer: API to query supported device driver formats

[Christophe Fergeau]

On Mon, Jan 28, 2013 at 11:24:55PM +0200, Zeeshan Ali (Khattak) wrote:
> On Mon, Jan 28, 2013 at 7:28 PM, Christophe Fergeau <cfergeau at redhat.com> wrote:
> > On Mon, Jan 28, 2013 at 04:40:06PM +0200, Zeeshan Ali (Khattak) wrote:
> >> On Mon, Jan 28, 2013 at 12:27 PM, Christophe Fergeau
> >> <cfergeau at redhat.com> wrote:
> >> > On Mon, Jan 28, 2013 at 05:18:43AM +0200, Zeeshan Ali (Khattak) wrote:
> >> >> From: "Zeeshan Ali (Khattak)" <zeeshanak at gnome.org>
> >> >> diff --git a/osinfo/osinfo_install_script.h b/osinfo/osinfo_install_script.h
> >> >> index d91751e..82486ef 100644
> >> >> --- a/osinfo/osinfo_install_script.h
> >> >> +++ b/osinfo/osinfo_install_script.h
> >> >> @@ -163,6 +165,9 @@ OsinfoPathFormat osinfo_install_script_get_path_format(OsinfoInstallScript *scri
> >> >>  gboolean osinfo_install_script_get_can_pre_install_drivers(OsinfoInstallScript *script);
> >> >>  gboolean osinfo_install_script_get_can_post_install_drivers(OsinfoInstallScript *script);
> >> >>
> >> >> +OsinfoDeviceDriverFormat osinfo_install_script_get_pre_install_driver_format(OsinfoInstallScript *script);
> >> >> +OsinfoDeviceDriverFormat osinfo_install_script_get_post_install_driver_format(OsinfoInstallScript *script);
> >> >
> >> > I don't think assuming that a given installer will support only one driver
> >> > format is expressive enough. For Windows post-install drivers, supporting
> >> > unpacked Windows drivers in addition to running a .exe shouldn't be very
> >> > hard, and this API would not work there.
> >>
> >> I see you point. I can make it a list. Would that be good?
> >
> > I think so. Though concretely why do we need to expose this information? In
> > all cases user of this information will need to drop it to a disk image
> > which will be passed to the VM, no?
> 
> Without this information, apps not only have to copy unnecessary
> driver files but most probably (as is the case with spice-guest-tools
> binary in Boxes) also need to create a redundant disk image to copy
> the files to when drivers are incompatible with scripts. Also same
> drivers can be available in multiple formats so Apps should be able to
> pick one in compatible (with script) format.

At this point, this is all theoritical, isn't it? We support one
post-install format, which must be a .exe supporting the /S switch, and
which has to be copied to an ISO image. Wouldn't it be better to postpone
this API until there's a need for it?

Christophe
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: not available
URL: <http://listman.redhat.com/archives/libosinfo/attachments/20130129/a914d563/attachment.sig>