Re: porting question

[Paul Eggert]

Andrei Agapi <address@hidden> writes:

> It's OK, I got it sorted out ... 24 and 26 are because
> of using an older 'sed'.

Thanks for reporting this.  What was the problem with the older 'sed'?
I assume that it mishandled the following script somehow (taken from
tests/sets.at), but what exactly went wrong?

        #n
        /^NULLABLE$/ {
           :null
           p
           n
           /^[   ]*$/ !b null
        }
        /^FIRSTS$/ {
           :firsts
           p
           n
           /^[   ]*$/ !b firsts
        }
        /^FDERIVES$/ {
           :fderiv
           p
           n
           /^[   ]*$/ !b fderiv
        }
        /^DERIVES$/ {
           :deriv
           p
           n
           /^[   ]*$/ !b deriv
        }

Is there a simple change to this script that would have made it work
for you, and saved you the trouble of worrying about the issue?


> The other 2 were Minix-related, are not because of bison ...

Again, is there a simple change to the test cases that would have
avoided the problem?  All other things being equal, we'd rather
test Bison than test Minix.

Thanks.