bug-gnu-utils
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: 3 make check errors


From: Aharon Robbins
Subject: Re: 3 make check errors
Date: Tue, 17 Feb 2009 21:16:40 +0200

Hi.

In article <address@hidden>,
Hermann Peifer  <address@hidden> wrote:
>Hi Arnold,
>
>Since just about ever, I am getting these 2 errors while running make check
>
>./fnarydel.ok _fnarydel differ: byte 12, line 2
>./fnparydl.ok _fnparydl differ: byte 26, line 2
>
>Could it be that my environment variable WHINY_USERS=1 is causing these 
>"errors"  ?

The answer is that this is indeed due to having WHINY_USERS set.
So they are not a problem.

>If yes: should the different order of array elements really 
>be reported as errors?

Yes, since the test suite is designed for the default case where
WHINY_USERS is not set. I'm not about to start mucking about with the test
suite for this.

>By the way: WHINY_USERS gets my vote for the best gawk extension *ever*. 
>I think it should be more documented and advertised.

It is documented, just not in the usual way. Use The Source, Luke. :-)

>After Ed Morton 
>mentioned it once, it took me weeks to find out that it wasn't just a 
>joke. I don't know how many subscripts I have constructed based on 
>country codes, etc. and later, while looping through the array, 
>everything comes in a nice alphabetic order, without any special effort.

asort() and asorti() are the documented, supported mechanisms for
doing this kind of thing.

There are several reasons why the behavior you like is not the default:

1. Sorting an array is potentially expensive. All users should not have
   to pay that price every time an array is traversed.

2. There is currently only one sort order. As soon as sorted traversal
   became the default, people would start yelling for ways to control
   the sorting (forward, reverse, string, number, locale based, character
   set based, you name it).  That way lies perl.

3. It is not compatible with previous versions or Unix awk.

And others I don't remember at the moment. :-)

>This is a 3rd make check error, which appeared recently. I am not quite 
>sure, what the real issue is here.
>
>./nofile.ok _nofile differ: byte 20, line 1

Also due to WHINY_USERS.  This variable controls several, unrelated
things within gawk.

Thanks,

Arnold




reply via email to

[Prev in Thread] Current Thread [Next in Thread]