maint: make sc_THANKS_in_duplicates more robust

The previous commit v8.23-124-g7b1ca5f made the above syntax-check rule
fail, because that took the whole content of THANKS.in for comparison.

* cfg.mk (sc_THANKS_in_duplicates): Strip off the header (all before the
first empty line) and the footer (all past the next empty line) from
'THANKS.in' for the check.

diff --git a/cfg.mk b/cfg.mk
index c1a46585f68b00de589760e304c7fe707149e7b3..4803074f818c0968c4bf72a3b104054be9be5a42 100644 (file)
--- a/cfg.mk
+++ b/cfg.mk
@@ -606,7 +606,8 @@ sc_preprocessor_indentation:
 # this rule detects that their pair may now be removed from THANKS.in.
 sc_THANKS_in_duplicates:
        @{ git log --pretty=format:%aN | sort -u;                       \
-           cut -b-36 THANKS.in | sed '/^$$/d;s/  *$$//'; }             \
+           cut -b-36 THANKS.in                                         \
+             | sed '/^$$/,/^$$/!d;/^$$/d;s/  *$$//'; }                 \
          | sort | uniq -d | grep .                                     \
            && { echo '$(ME): remove the above names from THANKS.in'    \
                  1>&2; exit 1; } || :