| 1 |
On Saturday 9 May 2009, 12:15, Stroller wrote: |
| 2 |
> On 8 May 2009, at 14:38, Stroller wrote: |
| 3 |
> > ... |
| 4 |
> > if echo hello|grep --color=auto l >/dev/null 2>&1; then |
| 5 |
> > export GREP_OPTIONS='--color=auto' GREP_COLOR='1;32' |
| 6 |
> > fi |
| 7 |
> |
| 8 |
> I'm afraid this thread has run away from me. I'm drinking the day's |
| 9 |
> first cup of tea & rubbing my eyes furiously in confusion. Wha? |
| 10 |
> I'm sure I'll comprehend the discussion better when I re-read later. |
| 11 |
> However, is there actually any need to parse whether the grep supports |
| 12 |
> colour before setting it? |
| 13 |
> |
| 14 |
> Let's say we use BSD grep or Schilling grep or whatever - is there |
| 15 |
> actually any harm in exporting GREP_OPTIONS='--color=auto' in this |
| 16 |
> case? |
| 17 |
|
| 18 |
Yes, because if the grep implementation in question supports GREP_OPTIONS |
| 19 |
but doesn't support --color, you'll get errors when it's run. |
| 20 |
|
| 21 |
(The assumption the author made is that if --color is supported, then |
| 22 |
GREP_OPTIONS is too, which is reasonable and is what happens for GNU |
| 23 |
grep, although I cannot speak for other implementations). |
Archives