Re: amendments to backtick-removing series

[Bernhard Voelker]

On 04/04/2012 04:13 PM, Jim Meyering wrote:
> Andreas Schwab wrote:
> 
>> Jim Meyering <address@hidden> writes:
>>
>>> diff --git a/tests/misc/stty b/tests/misc/stty
>>> index 97020e8..fa66a1a 100755
>>> --- a/tests/misc/stty
>>> +++ b/tests/misc/stty
>>> @@ -70,7 +70,7 @@ for opt in $options; do
>>>
>>>    # Likewise, 'stty -cread' would fail, so skip that, too.
>>>    test $opt = cread && continue
>>> -  rev=$(eval echo "\$REV_$opt")
>>> +  eval \$REV_$opt
>>
>> That's not what I wrote.
> 
> Indeed not.
> Send a patch next time, and I'll be less likely to botch it.

I think this code hasn't been run for a very long time,
because RUN_LONG_TESTS is not set anywhere.

  if test -n "$RUN_LONG_TESTS"; then
    # Take them in pairs.

So shouldn't this ~20sec part of the test be moved into
a very_expensive_ guard?

And if that part runs, then the test fails (at least here)
because of parenb and cread options.

Do they have to be excluded?

diff --git a/tests/misc/stty b/tests/misc/stty
index 650231f..ef3403a 100755
--- a/tests/misc/stty
+++ b/tests/misc/stty
@@ -79,8 +79,10 @@ done
 if test -n "$RUN_LONG_TESTS"; then
   # Take them in pairs.
   for opt1 in $options; do
+    case $opt1 in parenb|cread) continue;; esac
     echo .|tr -d '\n'
     for opt2 in $options; do
+      case $opt2 in parenb|cread) continue;; esac

       stty $opt1 $opt2 || fail=1

Have a nice day,
Berny