OK... here goes then: I begin by making a list of all of the "mode selection" options of dch:
-a Append
-e Edit
-r Release
-i Increment
-v Explicit version
-d From dir name
-n NMU
--bin-nmu BinNMU
-R Rebuild (-lbuild)
-q QA
-s Security
--bpo BPO
-l Local (suffix)
then I discounted those ones for which dch actually overrides the distributor back to Debian if they are used:
-n NMU
--bin-nmu BinNMU
-q QA
--bpo BPO
and discounted the ones which don't add a new version stanza to the changelog:
-a Append
-e Edit
-r Release
and discounted the ones which explicitly specify an exact version:
-v Explicit version
-d From dir name
and discounted the one already explicitly excluded from the "ubuntu1" block:
-R Rebuild (-lbuild)
which leaves:
-i Increment
-s Security
-l Local (suffix)
of which -i and -s can and should exhibit "ubuntu1" behaviour and -l never should.
... at which point I prove to myself that you are right and my changes as-is do in fact break the ubuntu-desired behaviour for -s. :-/
So I'll be fixing that, then.
Once I have fixed that, the logical justification proceeds as follows: Since the "ubuntu1" behaviour is desired only in the case of -i and -s, provided every preceding conditional in the long conditional EITHER refers to a mutually exclusive mode option OR explicitly tests for $distributor ne 'Ubuntu', it will be OK to move the "ubuntu1" logic into the !$opt_news block.
OK... here goes then: I begin by making a list of all of the "mode selection" options of dch:
-a Append
-e Edit
-r Release
-i Increment
-v Explicit version
-d From dir name
-n NMU
--bin-nmu BinNMU
-R Rebuild (-lbuild)
-q QA
-s Security
--bpo BPO
-l Local (suffix)
then I discounted those ones for which dch actually overrides the distributor back to Debian if they are used:
-n NMU
--bin-nmu BinNMU
-q QA
--bpo BPO
and discounted the ones which don't add a new version stanza to the changelog:
-a Append
-e Edit
-r Release
and discounted the ones which explicitly specify an exact version:
-v Explicit version
-d From dir name
and discounted the one already explicitly excluded from the "ubuntu1" block:
-R Rebuild (-lbuild)
which leaves:
-i Increment
-s Security
-l Local (suffix)
of which -i and -s can and should exhibit "ubuntu1" behaviour and -l never should.
... at which point I prove to myself that you are right and my changes as-is do in fact break the ubuntu-desired behaviour for -s. :-/
So I'll be fixing that, then.
Once I have fixed that, the logical justification proceeds as follows: Since the "ubuntu1" behaviour is desired only in the case of -i and -s, provided every preceding conditional in the long conditional EITHER refers to a mutually exclusive mode option OR explicitly tests for $distributor ne 'Ubuntu', it will be OK to move the "ubuntu1" logic into the !$opt_news block.