> Handle multiple rdloaddriver= parameters in the cmdlineDict. In cases
> where we already have a key, convert the value to a set and add the new
> value to it. Start new keys with just the value.
>
> In the boot loader code, take cmdlineDict values that are sets, convert
> them to lists, and join them to a single string separated by commas
> before writing them to the argument list.
This set looks fine, except for comments below.
> @@ -174,7 +174,10 @@ class KernelArguments:
>
> val = flags.cmdline.get(arg, "")
> if val:
> - newArgs.append("%s=%s" % (arg, val))
> + if type(val) == type(set()):
> + newArgs.append("%s=%s" % (arg, ','.join(list(val)),))
> + else:
> + newArgs.append("%s=%s" % (arg, val,))
> else:
> newArgs.append(arg)
>
You don't need the trailing commas in your tuples. The only time you
need to use a trailing comma is if you're creating a singleton tuple
because python syntax is a little weird there. Makes it a little easier
to read, I think.
> diff --git a/flags.py b/flags.py
> index e01ec63..afb56e2 100644
> --- a/flags.py
> +++ b/flags.py
> @@ -61,7 +61,19 @@ class Flags:
> key = i
> val = None
>
> - cmdlineDict[key] = val
> + if key.lower() == "rdloaddriver":
> + key = key.lower()
> +
> + if cmdlineDict.has_key(key):
> + if type(cmdlineDict[key]) == type(set()):
> + cmdlineDict[key].add(val)
> + else:
> + tmpset = set()
> + tmpset.add(cmdlineDict[key])
> + tmpset.add(val)
> + cmdlineDict[key] = tmpset
> + else:
> + cmdlineDict[key] = val
>
> return cmdlineDict
I don't really like this, but I think it will be okay. We don't appear
to allow for multiple instances of a command line option anywhere else
in anaconda, so you shouldn't be breaking existing behavior by using a
set when that comes up.
- Chris
_______________________________________________
Anaconda-devel-list mailing list
Anaconda-devel-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/anaconda-devel-list
