bug#29668: grep: Fatal problem with (big) file

[Norihiro Tanaka]

On Tue, 12 Dec 2017 16:28:09 -0800
Paul Eggert <address@hidden> wrote:

> On 12/11/2017 03:36 PM, Norihiro Tanaka wrote:
> > Perhaps, characters not to be able to recognize in your locale included
> > in Tieliikenne 5.0.csv and volvot.csv are included.
> 
> Yes, that's the problem. The original 'grep' output ended in "Binary file 
> Tieliikenne5.0.csv matches" but the user didn't see that. Perhaps we should 
> send that diagnostic to stderr as well.

I don't seem that that's problem.  the user pass output of grep to wc -l,
so `Binary file ... matches' line is also counted by `wc' as one line.

$ env LC_ALL=C grep 'Volvo' Tieliikenne\ 5.0.csv | wc -l
266175
$ env LC_ALL=en_US.utf8 grep 'Volvo' Tieliikenne\ 5.0.csv | wc -l
241264
$ env LC_ALL=en_US.utf8 grep 'Volvo' Tieliikenne\ 5.0.csv | tail -1
Binary file Tieliikenne 5.0.csv matches

$ env LC_ALL=C grep N3 volvot.csv | wc -l
17822
$ env LC_ALL=en_US.utf8 grep N3 volvot.csv | wc -l
11741
$ env LC_ALL=en_US.utf8 grep N3 volvot.csv | tail -1
Binary file volvot.csv matches