expand doesn't support multibyte characters (utf-8)

[Herbert Thielen]

Binary package hint: coreutils

'expand' doesn't replace tab characters by the correct number of spaces in case the input data contains multibyte characters, e.g. german umlauts when using the locale UTF-8.

A small example test file is attached.

I've attached a patch to coreutils 8.4 which works for me, but I'm not experienced in wide character programming, so there might still be some problem - more tests and review needed.

Thanks for the work, best regards

Herbert.

[Herbert Thielen (thielen)]

sample input file attached

[Herbert Thielen (thielen)]

expected output file attached

[Herbert Thielen (thielen)]

upstream bug report is filed at savannah.gnu.org #29138
https://savannah.gnu.org/bugs/index.php?29138

[C de-Avillez (hggdh2)]

Marking as Confirmed/Low; added upstream task. Generically, coreutils does not work with multi-byte strings.

@Herbert: thank you for opening the bug upstream.