[PATCH 4 of 5 RFC] patchbomb: lay test groundwork for some new functionality

[Denis Laxalde]

Augie Fackler a écrit :
> # HG changeset patch
> # User Augie Fackler <augie at google.com>
> # Date 1497385899 14400
> #      Tue Jun 13 16:31:39 2017 -0400
> # Node ID c98f63a3e8ee15645646539ba08b7b7682237cab
> # Parent  218470bf601bb0baaae333c73ccc0aab615d8f40
> patchbomb: lay test groundwork for some new functionality
>
> These email blobs are so big, it'll make things easier to review to
> add this, then implement the feature so the behavior variation is
> obvious.
>
> diff --git a/tests/test-patchbomb.t b/tests/test-patchbomb.t
> --- a/tests/test-patchbomb.t
> +++ b/tests/test-patchbomb.t
> @@ -3028,3 +3028,155 @@ multiple heads are missing at public
>    [255]
>
>  #endif
> +
> +Test hg email defaults settings:
> +  $ cat >> .hgemaildefaults <<EOF
> +  > [devel]
> +  > to = mercurial-devel at example.com
> +  > [devel.setsubscriptions]
> +  > c or d = c-or-d at example.com
> +  > a or b = a-or-b at example.com
> +  > EOF
> +This change modifies file d
> +  $ hg email --date '1970-1-1 0:2' --to=devel --test -r 10 --config patchbomb.publicurl=

I wonder if abusing the --to flag to select an alias (i.e. the "devel"
section in .hgemaildefaults, not sure how to name this) is really a good
idea. It gets confusing because, depending on the value of the --to
flag, the behavior is significantly different.
Maybe consider a dedicated flag (e.g. --alias=devel, suggesting "alias"
in the sense of email alias)?
That would also allow the command to work without a "to=" option under
"[devel]" section in .hgemaildefaults (e.g. one might only want a "cc=").