[Phoenix-pm] Running a SQL program within PERL DBI

[Scott Walters]

Hi Bobby,

Well, prototypes let you emulate built-in functions, as that was why they
were created.  They don't do type checking, but there are modules to do that.
One of the things built-ins do is operate on a hash without the user having 
to explicitly pass it by reference.  They also accept code blocks as arguments
without 'sub' in front of them and a few other things.  But prototypes are
useful for letting the user write more natural code that works more like
built-in functions.  Whether or not that's of value, it's of value to not
break it when people use do it ;)

-scott 

On  0, "Metz, Bobby W, WCS" <bwmetz at att.com> wrote:
> 	Well, I was looking for a simple "yes, P5 call without & works",
> but I like the example.  That said, I'm sure I'm opening myself up for
> ridicule, but the way I learned subs was that you pass the reference to
> the hash, not the var itself.
> 
> 	sub foo (\%) { print "Scott foo: ", $_[0]->{foo}, "\n";
> $_[0]->{foo} .= " Scott"; }
> 	my %stuff = (foo => "bar");
> 	foo(%stuff);
> 	&foo(%stuff);
> 
> 	sub foo2 { print "Bob foo: ", $_[0]->{foo}, "\n"; $_[0]->{foo}
> .= " Bob"; }
> 	my %stuff = (foo => "bar");
> 	foo2(\%stuff);
> 	&foo2(\%stuff);
> 
> 	Output:
> 	Scott foo: bar
> 	Scott foo: 
> 	Bob foo: bar
> 	Bob foo: bar Bob
> 
> 	Guess this has always made more sense to me since I first cut my
> teeth on ANSI C where you had to pass the ref to modify the values.  So,
> by passing the ref I can read and write.  But, it doesn't make an
> argument for using & since either calling method with the hash ref
> works, so I see the "yes" answer I was looking for and I don' t have to
> change my style (again, ridicule?).
> 
> Bobby
> 
> -----Original Message-----
> From: Scott Walters [mailto:scott at illogics.org]
> Sent: Thursday, March 16, 2006 8:58 AM
> To: Anthony R. Nemmer
> Cc: Metz, Bobby W, WCS; phoenix-pm at pm.org; Loo, Peter # PHX
> Subject: Re: [Phoenix-pm] Running a SQL program within PERL DBI
> 
> 
> Hi Tony,
> 
> In Perl 6, &sub is a reference to the sub, like \&sub in Perl 5.  Or
> something like that.  Typeglobs are going away, of course.  
> 
> Even if you don't use prototypes, other people might, and then & breaks
> the code.
> 
> sub foo (\%) { print "stuff: ", $_[0]->{foo}, "\n"; }
> 
> my %stuff = (foo => "bar");
> 
> foo(%stuff);
> &foo(%stuff);
> 
> When calling functions exported by modules (or not exported), you don't
> know
> if they're using prototypes.  Yes, prototypes are a bit of a hack, but
> only
> insomuch as they aren't complete.  I'd love to be able to prototype
> things
> like how grep $_ eq 5, @numbers or map "$_\n", @things work, where the
> first
> argument is an expression that gets treated like a code block.
> 
> &foo also defaults to sending @_ in the function call, which is
> sometimes handy
> but in other classes makes action-at-a-distance too easy, clobbering @_
> in
> some remote subroutine.
> 
> But most of all, &foo identifies the programmer as someone who learned
> to 
> program by immitating someone who immitated someone else and now writes
> Perl like Matt Wright.
> 
> As for whether or not Perl 4 is inherently bad, no, of course not.  Perl
> 1
> isn't inherently bad.  But Perl 5 offers some major improvements --
> three
> argument open, two argument system (for security), references (no more
> nasty
> glob tricks to pass datastructures), etc.  If you want to write in a
> Perl 4
> style and actually *know* Perl 4 and aren't just doing a Matt Wright
> impression,
> that's fine, but IMO (I AM GOD EVERYTHING I SAY IS RIGHT EVERYONE WHO
> DISAGREES
> IS WRONG) even old Perl 4 programmers should come up to speed on those
> few
> things or else they're writing inferior code by any metric.
> 
> Okay, rant done =)
> 
> -scott
> 
> On  0, "Anthony R. Nemmer" <intertwingled at qwest.net> wrote:
> > What SPECIFICALLY is wrong with using the & to identify a subroutine 
> > call, if I am not using prototypes?  Are you saying that it just makes
> 
> > the code look uglier?  I always thought prototypes were a hack anyway.
> > 
> > Perl has always used sigils to identify various entities...
> > $ for scalars, @ for arrays, % for hashs, \ for references, * for 
> > typeglobs, & for subroutines.  If I am not using prototypes, I don't
> see 
> > what the problem is with using & to identify a subroutine call.
> > 
> > I understand that this may all change in Perl 6.
> > 
> > Tony
> > 
> > Metz, Bobby W, WCS wrote:
> > > Hey thanks for the "&" tip.  Old habit I guess since the folks I
> learned
> > > from used it in all their code and I've never read up on Perl 5
> calls to
> > > find out otherwise.  Thought I knew what I was doing :-)
> > > 
> > > So, just "@results = mysql_mod::execute($my_sql_file);" does the
> trick?
> > > 
> > > Bobby
> > > 
> > > -----Original Message-----
> > > From: Scott Walters [mailto:scott at illogics.org]
> > > Sent: Wednesday, March 15, 2006 7:07 PM
> > > To: Metz, Bobby W, WCS
> > > Cc: Loo, Peter # PHX; phoenix-pm at pm.org
> > > Subject: Re: [Phoenix-pm] Running a SQL program within PERL DBI
> > > 
> > > 
> > >> use mysql_mod;
> > >> $my_sql_file = "blahblahblah"
> > >> @results = &mysql_mod::execute($my_sql_file);
> > > 
> > > Bobby, Peter,
> > > 
> > > I'd like to encourage you to drop the &, though.  You needed it in
> Perl
> > > 4;
> > > in Perl 5, you don't, but gives you a funky Perl 4 compatability
> mode
> > > where prototypes are disabled.  You don't want this.  It breaks
> things.
> > > 
> > >> foreach $line (@results)
> > >>  {
> > >>   # pattern match whatever
> > >>   # call other routines
> > >>   # etc.
> > >>  }
> > >>
> > >> This could save you time cut/pasting or re-writing the sqlplus
> call,
> > > 
> > > Yes, there's absolutely no reason why you can't put SQL in a file
> > > rather than hard-code it into a program.  
> > > 
> > >> redirection, or whatever, each and every time you have a new script
> > >> need.  Again, no point trying to parse the .sql file to perform the
> > >> queries in DBI...just process the output from sqlplus.
> > > 
> > > ... and I posted something that would fork off a process and run it
> > > in the SQL shell.
> > > 
> > >> 	As to leaving the group...give us another chance.  I agree with
> > >> Brock that your original problem statement wasn't well defined and
> I
> > > 
> > > If it makes you feel any better, I make about $5/hour >=)
> > > 
> > > -scott
> > > _______________________________________________
> > > Phoenix-pm mailing list
> > > Phoenix-pm at pm.org
> > > http://mail.pm.org/mailman/listinfo/phoenix-pm
> > > 
> > > 
> > 
> > 
> > -- 
> > 
> > I always have coffee when I watch radar!
> > _______________________________________________
> > Phoenix-pm mailing list
> > Phoenix-pm at pm.org
> > http://mail.pm.org/mailman/listinfo/phoenix-pm