On Tue, 2008-11-11 at 10:57 +0100, Filippo Giunchedi wrote:
> On Sun, Nov 09, 2008 at 08:13:26PM +0000, Adam D. Barratt wrote:
> > Looking at an earlier set of results, there are still a couple of false
> > positives but that's 97% less than before and libtool is full of
> > horrible shell:
> >
> > libtool_1.5.26-4_i386.deb:possible bashism in ./usr/bin/libtool line 4697
> > ($'...' should be "$(printf '...')"):
> > libtool_1.5.26-4_i386.deb- reload_objs="$objs$old_deplibs "`$echo
> > "X$libobjs" | $SP2NL | $Xsed -e '/\.'${libext}$'/d' -e '/\.lib$/d' -e
> > "$lo2o" | $NL2SP`" $reload_conv_objs" ### testsuite: skip nested quoting
> > test
> >
> > openc++_2.8-9_i386.deb:possible bashism in ./usr/bin/occ2 line 151 ($'...'
> > should be "$(printf '...')"):
> > openc++_2.8-9_i386.deb- , '\''$'\'' => '\''dlr'\''
>
> these are indeed trickier, I have no idea how to exclude these ATM :(
I've applied your patch anyway; thanks (and an equivalent for $"").
If someone comes up with a useful way of excluding the cases above then
I'll happily apply it but given the improvement over the previous
attempt at catching this particular bashism I'm quite happy to go with
the patch and see if anyone moans. :)
Adam
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
