The second commit, which just scans the output for possible b4_foo or m4_foo, would have helped me in the past diagnose some errors more easily. But to my biggest surprise it caught a few problems.
Akim Demaille (4): style: formatting changes in scan-code.l m4: catch suspicions of unevaluated macros tests: factor iterating over skeletons d: prepare to be able to run LAC tests TODO | 12 ++++++++---- data/skeletons/bison.m4 | 7 +++++++ data/skeletons/d.m4 | 4 ++-- data/skeletons/glr.c | 9 ++------- data/skeletons/lalr1.d | 3 +++ data/skeletons/location.cc | 17 ++++++++++------- src/scan-code.l | 6 +++--- src/scan-skel.l | 19 ++++++++++++++++++- tests/input.at | 38 ++++++++++++++++++-------------------- tests/local.at | 29 +++++++++++++++++++++++++++++ tests/regression.at | 10 +++++++++- tests/scanner.at | 2 +- 12 files changed, 110 insertions(+), 46 deletions(-) -- 2.32.0
