Alright, you're right, I haven't mentioned a few, but here we go.

First of all, lines that begin with C or M are ignored (but in the case of C it specifies the beginning of the block, so in that case everything after the C should be ignored and finally written as <c>)

Then, lines that begin with S, are written as: <set name="temant"> </set>, after the <t>.

The # are comments, and should be written as <!-- whatever -->, but this is already done, as you probably noticed.

One more thing, sometimes, a block entry might have many K's (or 1, doesn't matter), but no S or T (and then R's). Please note that S and T are *rarely* within the same block, if there is no T then theres usually an S instead. I've done all this successfully in the last code I sent you, so you can check that from there.

Hope this is clear and thanks for all the help you've given so far :)

Kind Regards,
Marcos

Updated: Added explanation.

Here is everything altogether:

For instance, lets say that the file reads:
C 100
K Perl
K Monks
K Is
K Cool
F 1234
R Whatever
R More
R BlaBla

This will equal (whatever comes after C in the same line is ignored):
<c>
<p>%Perl,%Monks,%Is,%Cool</p>
<t>
<setflag param="1234"/>
<r>
<li>Whatever</li>
<li>More</li>
<li>BlaBla</li>
</r>
</t>
</c>

This is very simple and has been implemented already (see the code at the end), the problem comes when the block begins with an I (capital 'i'), for example:
I 1010
K _EXE
F 1234
R Whatever
R More
R BlaBla

<c>
<p>_EXE</p>
<t>
<ifflag param="1010">
<then>
<setflag param="1234"/>
<r>
<li>Whatever</li>
<li>More</li>
<li>BlaBla</li>
</then>
</ifflag>
</r>
</t>
</c>

Additionally, lines that begin with C or M are ignored (but in the case of C it specifies the beginning of the block, so in that case everything after the C should be ignored and finally written as <c>)

Then, lines that begin with S, are written as: <set name="temant"> </set>, after the <t>.

The # are comments, and should be written as <!-- whatever -->, but this is already done, as you probably noticed.

One more thing, sometimes, a block entry might have many K's (or 1, doesn't matter), but no S or T (and then R's). Please note that S and T are *rarely* within the same block, if there is no T then theres usually an S instead. I've done all this successfully in the last code I sent you, so you can check that from there.

The two problems are: 1) Getting the </then>\n</ifflag> properly added at the end of such blocks,
2) Getting the '%' ignored *only* in such blocks (instead of <p>%_EXE</p>, do <p>_EXE</p>)

Ok, like I said, thats pretty much it, thanks again.

Kind regards,
Marcos

K One
K Two
[download]

<c>
<p>_EXE</p>
<t>
<ifflag param="1010">
<then>
<setflag param="1234"/>
<r>
<li>Whatever</li>
<li>More</li>
<li>BlaBla</li>
</then>
</ifflag>
</r>
</t>
</c>
[download]

Regarding your first question, yes that was my mistake, I forgot to include the "One" and "Two" in the output. I did it manually so.. =P

As for this question, I know it looks unbalanced and I perfectly understand, but thats the way it should be.

Thanks again.

Kind Regards,
Marcos

#!/bin/perl5

use strict;
use warnings;

my ( @K_array, @R_array ) = ();
my $I = '';
my $out = qq(<c>\n);

while (my $l = <DATA>) {
  my ( $code, $value ) = $l =~ /^(\w|\#)\s+(.*)\n$/;
  $I = $value if $code eq 'I';
  if ( $code eq 'K' ){
    $out .= "<p></p>\n<t>\n" unless @K_array;
    push @K_array, $value
  }
  elsif ( $code eq 'R' ){
    $out .= "<r></r>\n" unless @R_array;
    push @R_array, "<li>$value</li>\n"
  }
  $out .= qq(<set name="temant">$value</set>\n) if $code eq 'S';
  $out .= qq(<setflag param="$value"/>\n)       if $code eq 'F';
  $out .= "<!-- $value -->\n"                   if $code eq '#'
}

my $K;
if ( $I ){
  $K = $K_array[0] # assumes only 1 K with I
}
else{
  $K .= '%' . $_ . ',' for @K_array;
  chop $K
}
$out =~ s|<p></p>|<p>$K</p>|;

my $R = join '', @R_array;
$out =~ s|<r></r>|<r>\n$R</r>|;
$out .= qq(</t>\n</c>);
if ( $I ){
  $out =~ s|<t>|<t>\n<ifflag param="$I">\n<then>|;
  $out =~ s|</r>|</then>\n</ifflag>\n</r>|;
}

print "$out\n";

__DATA__
C 1010
K Perl
K is
K cool
F 1234
S Alright
R Whatever
R More
R BlaBla
[download]

While this appears to have potential, right now it only prints out lines that begin with an R, S or F, and completely ignoring the block structure itself.

By the way, some blocks don't even have an S or T, so, that shouldn't be depended on.

I also like the idea of hard-coding it all into the application, it avoids a whole lot of problems.

Thanks.

Kind Regards,
Marcos

Loading the input into a hash may be the way to go.

Ok,
Lines with M are *completely* ignored.

Lines that begin with a # can be found pretty much anywhere, be it at the beginning, middle or end of a block. But they should be enclosed in <!-- -->. Think of these as individual lines, that don't have any effect in the block structure itself.

There are no T's, I confused them with the F's. Sorry about that.
Now, as for the S and F's, here are some examples:
1) Input:
K One
K Two
S Alright
R Whatever?
R Blablabla
R And even more
1) Output:
<c>
<t>
<set name="temant>Alright</set>
<r>
<li>Whatever?</li>
<li>Blablabla</li>
<li>And even more</li>
</r>
</t>
</c>

2) Input:
K Perl
K Monks
S Theme
F 1234
R Examples
R More
R Yes
2) Output:
<c>
<p>%Perl,%Monks</p>
<t>
<set name="temant">Theme</set>
<setflag param="1234"/>
<r>
<li>Examples</li>
<li>More</li>
<li>Yes</li>
</r>
</t>
</c>

On a seperate note, there may be special blocks that don't have either an S or an F. Example:
K Special
K Case
R One
R Two
R Three
Here, it translates to:
<c>
<p>%Special,%Case</p>
<t>
<r>
<li>One</li>
<li>Two</li>
<li>Three</li>
</r>
</t>
</c>

Hopefully this should clear things up. If you have any other questions, don't hesitate to ask, thanks.

Kind Regards,
Marcos

Input with C:

C 1010
K Perl
K is
K cool
F 1234
S Alright
R Whatever
R More
R BlaBla
[download]

<c>
<p>%Perl,%is,%cool</p>
<t>
<setflag param="1234"/>
<set name="temant">Alright</set>
<r>
<li>Whatever</li>
<li>More</li>
<li>BlaBla</li>
</r>
</t>
</c>
[download]

I 1010
K _EXE
F 1234
R Whatever
R More
R BlaBla
[download]

<c>
<p>_EXE</p>
<t>
<ifflag param="1010">
<then>
<setflag param="1234"/>
<r>
<li>Whatever</li>
<li>More</li>
<li>BlaBla</li>
</then>
</ifflag>
</r>
</t>
</c>
[download]

K Special
K Case
R One
R Two
R Three
[download]

<c>
<p>%Special,%Case</p>
<t>
<r>
<li>One</li>
<li>Two</li>
<li>Three </li>
</r>
</t>
</c>
[download]

Yeah, when you set the DATA as individual blocks such as the ones you demo'ed, then it works flawlessly, but try setting DATA to say..

C 90
K One
K Two
F 1234
R This
R Is
R The
R First
R Example
#
K Second
F 4321
R Second
R Example
#
# Some
# Comments
#
K Last
F 0101
F 0202
F 0303
R This
R Is
R The
R Third
R And
R Last
R Example

Then you'll realize that the problem is the program interprets the whole input as one block, when in reality theres dozens and dozens of them.

Hope its a bit clearer now.

Thanks.

Kind Regards,
Marcos

Please John, don't give up now, I know you're close.

Whats the point of giving up right now..?

I really appreciate all the help you've given so far, honestly.

Best Regards,
Marcos