bug wrangling: strict ordering of 'affects' during merges?

[James Addison <jay@jp-hosting.net>]

Hi folks,

I'm learning the Debian BTS and control operations, and am trying to
merge three RC bugs that relate to a common cause (#1031293, #1031443,
#1031449).

It seems difficult to get the order of the 'affects' list to match in
each of the bugs, seemingly a requirement for them to merge
successfully - I've spent a little over an hour attempting various
combinations of +, - and = operations using "Control: affects" headers
to try to get a match.

The latest error from BTS was:

  Unable to merge bugs because:
  affects of #1031293 is
'src:python-scrapy,src:python-zstandard,src:rpmlint,src:libzstd' not
'src:rpmlint,src:python-scrapy,src:python-zstandard,src:libzstd'

Is it necessary to use a strict ordering check for the 'affects'
field?  And/or am I doing something stupid?  (other than trying to
help with bugs that aren't marked 'help', which I admit may be causing
nuisance)

Thanks,
James