=head2 Is there a way to talk to mSQL-1 and mSQL-2 within the same program?

No.



=head2 How should I install Msql, so that I can work with mSQL-1 and
mSQL-2 on demand?

You have to store the two modules in different places and take care
which one you want to use. You could do it with the following shell
script:

    for L in /usr/local/Minerva /usr/local/Hughes ; do
      echo ------------------------------------
      echo Configuring for $L
      echo ------------------------------------
      perl Makefile.PL INSTALLSITELIB=$L/perl INSTALLSITEARCH=$L/perl \
	INSTALLSCRIPT=$L/bin
      make
      make test
      make install
    done

From then you can talk to mSQL-1 by running

    perl -I/usr/local/Minerva/perl yourprogram

and you talk to mSQL-2 by running

    perl -I/usr/local/Hughes/perl yourprogram

The invocations differ in the @INC path (see man perlvar) such that
only the correct library for the desired version can get loaded. Once
again: you have to make sure yourself, that the correct library is
loaded, otherwise you cannot connect to the database.

Perl decides to load a library at compile time of the statement
'use Msql' (see man perlfunc/use). Your @INC path at that moment
decides upon which version you get.

Alternatively to the above you could insert the library location in
your own program with

    use lib '/usr/local/Minerva/perl'; # change @INC at compile time

After the above descibed steps the two installed pmsql(1) programs are
identical. You could insert such a 'use lib' statement in each of them
to have them talk to the appropriate server. Or you call them as perl
programs the same way as "yourprogram" above.



=head2 I get: Can't call method ... without a package or object reference

Run your script with the -w switch or test your connect and query
statements on success or failure.

    #WRONG:
    $db = Msql->connect();

    #RIGHTish:
    $db = Msql->connect or die "Couldn't connect to....: $Msql::db_errstr";

    #WRONG:
    $sth = $db->query("some query");

    #RIGHTish:
    $sth = $db->query("some query") or die "Query failed: $Msql::db_errstr";



=head2 I don't understand how you retrieve the data from a SELECT Query with Msql Perl.

Precanned reply:

    read the README

    read the manpage

    read the FAQ that comes with MsqlPerl

    read the tutorial in t/msql.t

subscribe to the msqlperl mailing list by sending mail to
msqlperl-request@franz.ww.tu-berlin.de with "subscribe msqlperl
your.name@your.domain" in the body of the message. Ask the question to
the mailing list by writing to msqlperl@franz.ww.tu-berlin.de. Wait
for the answer. Keep subscribed to the mailing list and answer the
same question as soon as somebody else asks it.



=head2 How do I get the names of the fields after a listfields?

Please, read the manpage. Type 'man Msql' and if that doesn't help,
try 'perldoc Msql'.



=head2 Where can I find more documentation?

Msql.pm IS the manpage in pod format. If you're on UNIX you should be
able to read it with 'man Msql'. On other operating systems try
'perldoc Msql'. Also, the test script, t/msql.t might help you to sort
things out. The most recent addition is the pmsql program which may
also help you to learning by example.



=head2 HPUX compilation goes wrong!!!

     From: Yiorgos Adamopoulos <Y.Adamopoulos@noc.ntua.gr>

     I successfully compiled MsqlPerl for a HP817 (HP-UX 9.04/s800) using Perl5.002,
     msql-1.0.14 and gcc-2.7.2.  I added CC_ONLY = -fPIC to site.mm after running
     setup.



=head2 How do I retrieve fields by name?

The internal fetchhash routine is quite good. If you need speed, you
should hardcode your column names into a hash slice. Everything else
is a terrible waste.

		     closure:  8 secs ( 7.47 usr  0.03 sys =  7.50 cpu)
	  hashslice via name: 10 secs ( 9.51 usr  0.04 sys =  9.55 cpu)
	  internal_fetchhash:  7 secs ( 7.05 usr  0.03 sys =  7.08 cpu)
       brute force fetchhash: 15 secs (13.90 usr  0.05 sys = 13.95 cpu)
     hashslice via constants:  6 secs ( 5.57 usr  0.02 sys =  5.59 cpu)

     #!/usr/bin/perl

     use Msql;
     my $db = Msql->Connect("","mod");
     $sth = $db->Query("select * from mods");


     sub Msql::Statement::brute_fetchhash {
	 my $self = shift;
	 my %hash;
	 @hash{ $self->name } = $self->FetchRow and \%hash;
     }

     sub Msql::Statement::gimme_a_fetchhash_routine {
	 my $self = shift;
	 my @z = $sth->name;
	 sub { my %hash; @hash{@z} = $self->FetchRow and \%hash;};
     }

     use Benchmark;

     timethese(20, {
     'hashslice via constants' => q{
	 $sth->DataSeek(0);
	 while (@row{qw/modid userid maillistid chapterid seqdummy statd stats statl stati description changed changedby/} = $sth->FetchRow){
	     $X = $row{modid};
	     print "$X\n" ;
	 }
     },

     '     hashslice via name' => q{
	 $sth->DataSeek(0);
	 while (@row{$sth->name} = $sth->FetchRow){
	     $X = $row{modid};
	     print "$X\n" ;
	 }
     },

     '  brute force fetchhash' => q{
	 $sth->DataSeek(0);
	 while ($hashref = $sth->brute_fetchhash){
	     $X = $hashref->{modid};
	     print "$X\n" ;
	 }
     },
     '                closure' => q{
	 $sth->DataSeek(0);
	 $clos = $sth->gimme_a_fetchhash_routine;
	 while ($hashref = &$clos){
	     $X = $hashref->{modid};
	     print "$X\n" ;
	 }
     },
     '     internal_fetchhash' => q{
	 $sth->DataSeek(0);
	 while (%hash = $sth->fetchhash){
	     $X = $hash{modid};
	     print "$X\n" ;
	 }
     },
     });



=head2 Dynamic Loading

Hello I am running XXX and would really like to use Msqlperl but I am
getting this error message.

 Can't load module Msql, dynamic loading not available in this perl.
   (You may need to build a new perl executable which either supports
	dynamic loading or has the Msql module statically linked into it.)
  at /usr/lib/perl5/site_perl/Msql.pm line 56
 BEGIN failed--compilation aborted at ./sql.pl line 5.

There is a manpage dedicated to building and installing modules:
ExtUtils::MakeMaker. The long answer is in there. The short answer is:
run 'make perl' in the MsqlPerl build directory. The third answer is,
get an operating system that supports dynamic loading. With Linux it's
already quite standard to have dynamic loading, and so is with many
other OSs. Check out the newgroups for your OS and consult the FAQs
there.



=head2 How can I construct an insert statement from an array?

The icc routine below should do what you want. It's hardly tested and
I'm reluctant to add it to the distribution. But feel free to convince
me that including it is a good thing.

    #!/usr/bin/perl -w
    use Msql;
    use strict;

    sub Msql::icc {
	my($db,$table,@fields)=@_;
	my($s,@types)=$db->listfields($table) or return;
	if (@fields){
	    # The user wants to input the fields in a different order
	    my(@tfields,@ttypes,%tfields);
	    @tfields = $s->name;
	    @ttypes = $s->type;
	    @tfields{@tfields} = @ttypes;
	    @types = @tfields{@ttypes};
	} else {
	    # They are gonna give the arguments in table's order
	    @fields = $s->name;
	    @types = $s->type;
	}
	# return a subroutine reference.
	sub {
	    my(@arr)=@_;
	    return join " ",
	    "insert into $table values (",
	    join(
		 ",",
		 map {
		     defined $arr[$_] ?
			 $types[$_] == &Msql::REAL_TYPE
			     ||
			 $types[$_] == &Msql::INT_TYPE ?
			     $arr[$_]+0 :
				 Msql->quote($arr[$_]) :
				     "NULL"
				 }
		 0..$#types
		),
		     ")\n";
	}
    }

    # Example for usage:
    {
	my $dbh=Msql->connect("","test");
	my $f=$dbh->icc("pwent");
	my(@info,$query,$sth);
	while (@info = getpwent()) {
	    print $query =&$f(@info);
	    $sth = $dbh->query($query) or die Msql->errmsg;
	}
    }

We also got email from Doug Wilson who pounded on the icc method:

    Date: Thu, 1 May 1997 14:23:31 -0500 (CDT)
    From: Doug Wilson <doug@hurl.net>
    To: msqlperl@franz.ww.tu-berlin.de

    After debugging the sample insert function from the installation FAQ,
    I added an extra feature; now not only are strings automatically quoted
    correctly (as before), they are also truncated to their database length.

    my sample:
    #!/usr/bin/perl
    # tabstop=4

    use lib '/usr/local/Minerva/perl';
    use Msql;

    sub Msql::icc
    {
	my($db,$table,@fields)=@_;
	my($s,@types)=$db->listfields($table) or die Msql->errmsg;
	my $tfld_idx = 0;
	if (@fields){
	    # The user wants to input the fields in a different order
	    my(@tfields,@ttypes,@tlengths,%tmptypes,%tmplengths,$fieldname);
	    @tfields = $s->name;
	    @ttypes = $s->type;
	    @tmptypes{@tfields} = @ttypes;
	    @tlengths = $s->length;
	    @tmplengths{@tfields} = @tlengths;
	    foreach $fieldname (@fields) {
		$types[$tfld_idx]=$tmptypes{$fieldname};
		$lengths[$tfld_idx++]=$tmplengths{$fieldname};
	    }
	} else {
	    # They are gonna give the arguments in table's order
	    @fields = $s->name;
	    @types = $s->type;
	    @lengths = $s->length;
	}
	# return a subroutine reference.
	sub {
	    my(@arr)=@_;
	    return join " ",
	    "insert into $table",
		$tfld_idx > 0 ?
		join(" ", "(", join(",", @fields), ")") : "",
		"values (",
		join(",",
		     map {
			 defined $arr[$_] ?
			     $types[$_] == &Msql::REAL_TYPE
				 ||
			     $types[$_] == &Msql::INT_TYPE ?
				 $arr[$_]+0 :
				     Msql->quote(
					substr($arr[$_],0,$lengths[$_]))
			    :
			     "NULL"
				     }
		     0..$#types
		    ),
		    ")\n";
	}
    }

    #Sample Usage
    my ($regfile) = "7094.reg";
    my(%regvars,%dbvars);

    #Read in keys & values
    if (open (REGFILE, "<$regfile"))
    {
	while (<REGFILE>)
	{
	    $line = $_;
	    chomp($line);
	    ($key, $value) = split(/=/, $line);
	    $regvars{$key} = $value;
	}

	close (REGFILE);
    }

    #Do the file to database field translation
    $dbvars{user}=$regvars{Pppname};
    $dbvars{lname}=$regvars{LastName};
    $dbvars{fname}=$regvars{FirstName};
    $dbvars{address1}=$regvars{Address};
    $dbvars{address2}=$regvars{Address1};
    $dbvars{city}=$regvars{City};
    $dbvars{state}=$regvars{State};
    $dbvars{zip}=$regvars{Zip};
    $dbvars{country}=$regvars{Country};
    $dbvars{hphone}=$regvars{Phone};
    $dbvars{email}=$regvars{EmailName};
    $dbvars{epasswd}=$regvars{EmailPassword};
    $dbvars{ppp}=$regvars{PppName};
    $dbvars{ppasswd}=$regvars{PppPassword};
    $dbvars{regcode}=$regvars{RegCode};
    $dbvars{serial}=$regvars{RegNumber};

    $dbh=Msql->connect("","ispreg");
    my @fnames = keys(%dbvars);
    my $f=$dbh->icc("customers", @fnames);
    my(@info,$query,$sth);
    @info = values(%dbvars);

    $query = &$f(@info);
    $sth = $dbh->query($query) or die Msql->errmsg;
    ----------------------------------------------------------------------
    To unsubscribe from this list send a message containing "unsubscribe"
    to msqlperl-request@franz.ww.tu-berlin.de



=head2 What can I do if I can't connect to local msql server

    Date: Sun, 27 Apr 1997 15:52:16 -0400 (EDT)
    From: Mark-Jason Dominus <mjd@plover.com>
    To: msql-list@Bunyip.Com
    Subject: Re: [mSQL] linux: can't connect to local msql server

    On Sun, 27 Apr 1997, Rob Hill wrote:
    > I have the msql.sock devices in my /usr/local/Hughes dir but I cannot
    > connect to msql locally.

    When I installed msql2, it put the socket into
    /usr/local/Hughes/msql.sock, but the perl Msql.pm module wanted to
    find it under /dev/msql.  I made a symbolic link from /dev/msql to
    /usr/local/Hughes/msql.sock.

    I did

	    strace -o /tmp/msqlperl perl testprogram
	    grep connect /tmp/msqlperl 

    to find out where Msql was looking for the socket file.

    mjd@pobox.com                                             Mark-Jason Dominus
    mjd@plover.com                              Plover Systems, Philadelphia, PA