| 1 |
On Sunday 07 October 2007, Roy Marples wrote: |
| 2 |
> On Sun, 2007-10-07 at 01:09 -0400, Mike Frysinger wrote: |
| 3 |
> > On Tuesday 02 October 2007, George Shapovalov wrote: |
| 4 |
> > > Tuesday, 2. October 2007, Roy Marples Ви написали: |
| 5 |
> > > > And here it is |
| 6 |
> > > |
| 7 |
> > > - if [[ "${FORTRANC}" = "gfortran" ]]; then |
| 8 |
> > > + if [ "${FORTRANC}" = "gfortran" ]; then |
| 9 |
> > > |
| 10 |
> > > You know, it is funny to see these lines after all those cries about |
| 11 |
> > > how [ is evil and we should really never ever use it but rather always |
| 12 |
> > > use [[ form :). |
| 13 |
> > > |
| 14 |
> > > Oh, btw, does the compliant sh correctly process the most cited |
| 15 |
> > > composition? : [ -n $foo ] && [ -z $foo ] && echo "huh?" |
| 16 |
> > > (straight from devmanual, the primary reason why [ is considered bad) |
| 17 |
> > |
| 18 |
> > [...] will crap a brick if foo expands to multiple arguments |
| 19 |
> |
| 20 |
> Only if unquoted. |
| 21 |
> |
| 22 |
> > which is why |
| 23 |
> > [[...]] is preferred ... it'll do the sane thing and work correctly |
| 24 |
> |
| 25 |
> Which encourages developers not to quote in my view. |
| 26 |
|
| 27 |
addressing one issue (quoting) while ignoring others (extended syntax) leads |
| 28 |
to crappy inconsistencies. if the standard you want (POSIX) cant do it all, |
| 29 |
then nitpicking details is a waste of time. |
| 30 |
-mike |
Archives