RE: htags improvements.

[Bakken, Luke]

> Thank you for your patch. It's a good job.
> 
> By the way, one month age, I added the following comment to the top
> of htags.in in CVS repository. Sorry for my late comment.
> 
> > # COMMENT FOR HACKERS:
> > #
> > # Htags is written so that it run even with perl4.
> http://savannah.gnu.org/cgi-bin/viewcvs/global/global/htags/ht
ags.in.diff?r1=1.168&r2=1.169
> 
> I agree with refactoring and I want also to move to new perl early
> in the mind.  But I'm hesitant to rewrite htags in perl5 style when
> thinking about the caused troubles.
> 
> There are many version of perl in the world. And I know many people
> were annoyed by the compatibility problem of perl.
> 
> For example, FreeBSD used perl4 up to a comparatively recent date
> and even now it has only perl-5.005_03 which is comparatively old.
> 
>       FreeBSD 2.2.5(1997/10) - 2.2.8(1998/12)         perl-4.036
>       FreeBSD 3.0(1998/10) - 3.1(1999/2)              perl-5.005_02
>       FreeBSD 3.2(1999/5) - 4.9(2003/10)              perl-5.005_03
> 
> In FreeBSD project, in a certain period of time, hackers wrote some
> commands using perl. It was thought to be a good idea. But in fact,
> they were annoyed by the compatibility problem of perl. At last,
> they removed perl from their core system and many utilities have
> been rewritten as shell scripts or C programs.
> 
> I think we should rather move to C language for the performance and
> the extendibility in the future. (It takes time though.)
> Till then, it is safe to keep perl4 syntax, I think. I know that
writing
> program using just perl4 facilities, it works well in any environment.

Thanks for your reply,

I can see not changing htags because you view it as a dead-end written
in perl. However, perl4 is a dead language. As you pointed out, it was
last used in freebsd in 1998 - 5 years ago. This is practically
"forever" in terms of perl, or any open-source language. I program perl
professionally and have never used perl4 syntax.

I will keep my patch and apply it when I upgrade global and htags. I am
also going to work on htags for further speed improvements, because I
see many areas in which it can be sped up. Perhaps this will change your
mind about moving to C - has anyone started that project yet?

Thanks,
Luke