Bug#253012: Bug#253498: non-posixisms
On Wed, Jun 09, 2004 at 04:11:39PM -0400, Clint Adams wrote:
> These RE's could be tweaked, I imagine, but they're better than the current
> situation.
>
> --- /usr/share/lintian/checks/scripts 2004-04-25 17:53:59.000000000 -0400
> +++ checks/scripts 2004-06-09 15:59:57.703351772 -0400
> @@ -422,7 +422,15 @@
> '[^\\\]\{([^\s]+?,)+[^\\\}\s]+\}', # brace expansion
> '(?:^|\s+)\w+\[\d+\]=', # bash arrays, H[0]
> '\$\{\#?\w+\[[0-9\*\@]+\]\}', # bash arrays, ${name[0|*|@]}
> - '(?:^|\s+)(read\s*(?:;|$))' # read without variable
> + '(?:^|\s+)(read\s*(?:;|$))', # read without variable
> + '(test|\[) .+-[ao]', # test/[ -a/-o binary operators
(test|\[).+\s-[ao]\s perhaps?
This has still the problem of mathing stuff in " " but we can probably
ignore that.
> + 'kill\s+-[^s]', # kill -[0-9] or -[A-Z]
I would suggest either kill\s+-[^sl] or kill\s+-[0-9A-Z]
> + 'trap\s+".*"\s+.*[1-9]', # trap with signal numbers
hmm, perhaps better trap\s+["']?.*["']?\s+[1-9]
> + 'local', # local scoping of variables
local\s
> + 'command\s+-v', # UPism
> + 'type .+', # XSIism
type\s+.+ just to be consitent
> + 'echo\s+-[neE]', # technically -n is permitted
> + '&>' # technically a cshism
> );
>
> for my $re (@bashism_regexs) {
And we need an common RE before the commands to try to guess if they are
really commands.
Will look further into it, thanks for the patch.
Gruesse,
--
Frank Lichtenheld <djpig@debian.org>
www: http://www.djpig.de/
Reply to: