Not sure how savannah reports are followed, so small headups of a thing I
experienced which probably is a bug. I made small bug report at
https://savannah.gnu.org/support/?108110 - though obviously the behavior as
described for '==' is valid (rhs is pattern).

So what I suspected was that [[ <expression> ]] treated both = and == in the
same way - and this seemed invalid, as for = it should be treated as a normal
string - in the same way as for [ <expression> ] and test.

If my reasoning is fine, following patch should address the problem. If not,
then just ignore it.

I also added minor man update, to emphasize [[ and [ difference in rhs argument.

Michal Soltys (1):
  execute_cmd.c: don't treat rhs of = as pattern in conditional
    expressions

 doc/bash.1    |    5 ++++-
 execute_cmd.c |    7 +++----
 2 files changed, 7 insertions(+), 5 deletions(-)

-- 
1.7.10.4


Reply via email to