On 08/02/15 19:40, Bernhard Voelker wrote:
> * Makefile.am (sort-THANKS.in): Add target to sort the name/email pair
> list in 'THANKS.in'.
> * THANKS.in: Sort via the above new target.
> * cfg.mk (sc_THANKS_in_sorted): Add rule to ensure that 'THANKS.in'
> remains sorted.
> ---
> Makefile.am | 15 +++++++++++++++
> THANKS.in | 56 ++++++++++++++++++++++++++++----------------------------
> cfg.mk | 11 +++++++++++
> 3 files changed, 54 insertions(+), 28 deletions(-)
>
> diff --git a/Makefile.am b/Makefile.am
> index cf0c8a0..e3a4354 100644
> --- a/Makefile.am
> +++ b/Makefile.am
> @@ -172,6 +172,21 @@ THANKS: THANKS.in Makefile.am .mailmap thanks-gen
> .version
> printf ';; %s\n' 'Local Variables:' 'coding: utf-8' End:; \
> } > $@-t && chmod a-w $@-t && mv $@-t $@
>
> +# Sort the contributor name/email pair list in 'THANKS.in'.
> +# See sc_THANKS_in_sorted ... which uses the same sort options.
> +.PHONY: sort-THANKS.in
> +sort-THANKS.in: Makefile.am
> + $(AM_V_GEN)rm -f $@-t; \
> + { \
> + sed '/^$$/q' THANKS.in \
> + && { sed '/^$$/,/^$$/!d;/^$$/d' THANKS.in \
> + | LC_ALL='en_US.UTF-8' sort -f -k1,1; } \
> + && echo \
> + && sed '0,/^$$/d;0,/^$$/d' THANKS.in; \
> + } > $@-t \
> + && test "$$(wc < THANKS.in)" = "$$(wc < $@-t)" \
> + && mv $@-t THANKS.in
This separate target is probably overkill,
as THANKS.in should remain sorted due to the sc_ rule.
> diff --git a/cfg.mk b/cfg.mk
> index f5be6de..26df247 100644
> --- a/cfg.mk
> +++ b/cfg.mk
> @@ -611,6 +611,17 @@ sc_THANKS_in_duplicates:
> && { echo '$(ME): remove the above names from THANKS.in' \
> 1>&2; exit 1; } || :
>
> +# Ensure the contributor list stays sorted. The sort order can be fixed
> +# with 'make sort-THANKS.in' ... which uses the same sort options.
> +sc_THANKS_in_sorted:
> + @sed '/^$$/,/^$$/!d;/^$$/d' THANKS.in > $@.1; \
> + LC_ALL='en_US.UTF-8' sort -f -k1,1 $@.1 > $@.2
> + @diff -u $@.1 $@.2; diff=$$?; \
> + rm -f $@.1 $@.2; \
> + test "$$diff" = 0 \
> + || { echo '$(ME): THANKS.in is unsorted; please run' \
> + 'make sort-THANKS.in' 1>&2; exit 1; }
> +
I don't like the sorting in this locale at all.
Actually the locale is OK, just the sorting in the patch seems
to have been done in LC_ALL=C ?
$ printf '%s\n' Duncan Dániel Gaël Gary Ørn Oskar | LC_ALL=en_US.utf8 sort
Dániel
Duncan
Gaël
Gary
Ørn
Oskar
That locale doesn't work correctly with the only cyrillic entry though:
$ printf '%s\n' Марк Michal | LC_ALL=en_US.utf8 sort
Michal
Марк
I'm thinking we should convert that to latin1 (Mark Korenberg) like all other
entries.
cheers,
Pádraig.
