Re: [PATCH] - more save to bugzilla error checking

[Jeremy Katz <katzj redhat com>]

On Fri, 2008-10-31 at 15:47 -0400, Chris Lumens wrote:
> Used to be that "ppc" was not a valid platform in bugzilla but "powerpc"
> was.  Of course, rpmUtils.arch.getBaseArch returned the former.  Well
> that's been corrected in our bugzilla now, but this patch is still valid
> as far as I'm concerned.
> 
> This checks the platform and sets it to All if rpmUtils returns
> something that bugzilal doesn't understand, and checks that
> product.productName also exists in bugzilla.  It's up to the install
> class to give us a useful backup there, just like it is with the
> version.

Looks okay -- one minor nit

> diff --git a/filer.py b/filer.py
> index e8431ea..74e8461 100644
> --- a/filer.py
> +++ b/filer.py
> @@ -310,6 +322,11 @@ class BugzillaFiler(AbstractFiler):
>                  whiteboards.append((wb, val))
>                  kwargs.pop(key)
>  
> +            if key == "platform":
> +                platformLst = self.__withBugzillaDo(lambda b: b._proxy.Bug.legal_values({'field': 'platform'}))
> +                if not val in platformLst:
> +                    kwargs[key] = "All"

Is All guaranteed to be in the list?  Obviously it is in our bugzilla,
but maybe we should fall back to gettin ga list of platforms (is that
even possible) and just chucking in the first one?  We should at least
verify that All is valid and give a nice error like we do for other
cases

Jeremy