Jeremy Katz wrote:

On Tuesday, June 02 2009, Allen Kistler said:

[snip]

...

1. Has that been a conscious design choice?

   Well, no, apparently it wasn't. It simply resulted because anaconda switched in F11 to using parted for partitioning. How much anaconda should do vs. how much parted should do is being discussed among the maintainers. The rest of us get to wait and watch.

For clarification, anaconda has been using parted for partitioning for eight years. The code just got reworked for the Fedora 11 cycle and the swap flag setting got lost in the shuffle

Okay. Sorry if I publicly misinterpreted what I thought I read.

...

2. Does it introduce inter-operation problems with other distros (i.e., dual-boot, etc.) and non-Fedora apps?

   On that, no one seems to know for sure, but there is concern that it might. From the comments in the bug, I'd say type 82 is likely to make a return. Whether it does so by F11 GA is another matter.

With a fairly high degree of certainty I can say it won't. There might be problems with something like PartitionMagic (I haven't looked at their behavior with various odd cases in a couple of years) but the Linux world has basically stopped caring since with non-msdos partition tables, you don't have any partition ids to key off of.

Won't be problem for other distros? Other apps? Won't come back? Won't make it into the GA?

Whatever the intent, fdisk still let's you change the type after installation if you want. If you don't want, that works for swap, too. (As I said, I hadn't seen any problems.)

I was under the impression that the RAID autostart type mattered, though. I can easily remember when it was crucial. Hmm... Maybe it's time to play.