I tried:
http://www.hotlinkfiles.com/files/526004_qpvr0/pcre-fpc.tar.gz

Can't get it to work.

There is also:

http://www.renatomancuso.com/software/dpcre/dpcre.htm

(By the way, why all these people wrap PCRE in a class is beyond me. Object oriented PCRE.. defeats the bloody purpose of regular expressions in the first place - quick and dirty access to regexes must not be stored in a class. I repeat, must not. So I think we'll use these downloads as ideas and implement a proper PCRE wrapper unit that's easy to use quickly and dirtily (if that was a word))

I tried: http://www.hotlinkfiles.com/files/526004_qpvr0/pcre-fpc.tar.gz Can't get it to work.

Doing a dumb header port is not that hard. If you can't do it, start making tests, and I'll do it.

then try this: http://www.renatomancuso.com/software/dpcre/dpcre.htm

I've used the above wrapper in a news ticker application without any problems. AFAIK, it still being used in 11 news channels on a daily basis.

Cheers,
-Krishna

Doing a dumb header port is not that hard. If you can't do it, start making tests, and I'll do it.

L505 loves flamebaits. They are what I eat for breakfast.

I personally like to spend my time doing useful things, and if someone has already made the translation then I'm not going to waste time on it. Just like I'm not going to rewrite a stringlist if there is already a stringlist in the RTL. Waste of time. I'd rather write a new data structure and make my time useful. Or I'd rather convert a header that hasn't already been converted.

So why use the JCL? Just do all the translations yourself.. forget JCL.

And yes, that means that not all programmers want to roll their own sysutils or system unit. Lots of them like to make use of something that's already been written for them. And as an API designer, I know this well.

L505

> Doing a dumb header port is not that hard. If you can't do it, start making > tests, and I'll do it.

(skip nonsense stuff)

So why use the JCL? Just do all the translations yourself.. forget JCL.

- JCL's license is MPL and thus not suitable for inclusion in FPC http://wiki.freepascal.org/licensing
- JCL is not in a runnable state for most FPC targets.
- translated Delphi headers are typically not clean for cross platform (most notably 64-bit) use.

After making a few tiny changes to the Delphi source, I got this to compile and work:

http://www.renatomancuso.com/software/dpcre/dpcre.htm

The author says that he had to modify the C source and recompile in order to give Pascal access to all pcre features.

__________________________________________________ Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com

Unfortunately, it's slightly slower than my fastest program that uses the library that comes with FPC. I wonder if the dll was compiled with all of the C compiler's optimiaztions turned on.

{$mode objfpc}
uses pcre; // from www.renatomancuso.com/software/dpcre/dpcre.htm

const
patterns : array[1..9] of pchar =
(
'agggtaaa|tttaccct',
'[cgt]gggtaaa|tttaccc[acg]',
'a[act]ggtaaa|tttacc[agt]t',
'ag[act]gtaaa|tttac[agt]ct',
'agg[act]taaa|ttta[agt]cct',
'aggg[acg]aaa|ttt[cgt]ccct',
'agggt[cgt]aa|tt[acg]accct',
'agggta[cgt]a|t[acg]taccct',
'agggtaa[cgt]|[acg]ttaccct'

    );
  replacements : array[0..10,0..1] of string[15] =
  (
    ('B', '(c|g|t)'), ('D', '(a|g|t)'), ('H', '(a|c|t)'), ('K', '(g|t)'),
    ('M', '(a|c)'), ('N', '(a|c|g|t)'), ('R', '(a|g)'), ('S', '(c|t)'),
    ('V', '(a|c|g)'), ('W', '(a|t)'), ('Y', '(c|t)')
  );

// Append 2 strings to an ansistring rapidly.  Note: the ansistring's
// length will be increased by a more than sufficient amount.
function append2( var dest: ansistring; len0: longint;
                  s1: pchar; len1: longint;
                  s2: pchar; len2: longint): longint; inline;
const  quantum = 599000;
var  newlength: longint;
begin
  newlength := len0 + len1 + len2;
  // Since setlength() is somewhat costly, we'll do it less
  // often than you would think.
  if length( dest ) < newlength then
    setlength( dest, newlength + quantum );
  move( s1^, dest[len0 + 1], len1 );
  move( s2^, dest[len0 + 1 + len1], len2 );
  exit( newlength );
end;
procedure replace_matches( const str: ansistring;  var dest: ansistring );
var
  rx : iregex;
  capture: imatch;
  start, index, truelength, i : longint;
  target, repl, found: string[255];
begin
  target := '[';
  for i := 0 to high(replacements) do
    target += replacements[i,0];
  target += ']';
  rx := regexCreate( target );
  dest := '';   truelength := 0;   start := 0;
  repeat
    capture := rx.match( str, start );
    if not capture.success then  break;
    index := capture.groups[0].index;
    found :=  str[ index + 1 ];
    repl := replacements[ pos(found,target)-2, 1 ];
    truelength := append2(
      dest, truelength, @str[start+1], index-start, @repl[1], length(repl) );
    start := index + 1;
  until false;
  setlength( dest, truelength );
  dest := dest + Copy( str, start+1, length(str) - start + 1);
end;
function count_matches( const needle, haystack: ansistring ):longint;
var
  rx : iregex;
  capture: imatch;
  where: longint;
begin
  rx := regexCreate( needle );
  result := 0;
  where := 0;
  repeat
    capture := rx.match( haystack, where );
    if not capture.success then  break;
    inc( result );
    where := capture.groups[0].index + capture.groups[0].length;
  until false;
end;
var
  pattern : pchar;
  sequence, new_seq : ansiString;
  line: string[255];
  i, count, init_length, clean_length: longint;
  inbuf : array[1..64*1024] of char;
begin
  settextbuf(input,inbuf);
  sequence := '';
  init_length := 0;
  while not eof do
  begin
    readln( line );
    init_length += length( line ) + 1;
    if line[1]  '>' then
      sequence := sequence + line;
  end;
  clean_length := length(sequence);
  for i := low(patterns) to high(patterns) do
  begin
    pattern := patterns[i];
    count := count_matches( pattern, sequence );
    writeln( pattern, ' ', count);
  end;
  replace_matches(sequence, new_seq);
  writeln;
  writeln( init_length );
  writeln( clean_length );
  writeln( length(new_seq) );
end.

__________________________________________________ Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com

Unfortunately, it's slightly slower

Why? This is great if our library is fast :)

No, not nonsense. Jeff P. (yetanothergeek) had a header for freepascal not using any JCL, AFAIK. Which is the original link I provided, which wasn't working. Which is why I posted this to the mailing list.

He did it for us already, AFAIK. No sense repeating Jeff's work if he already did it.

Because of the ridiculous GPL license which freepascal uses, which isn't as free as a BSD/NRCOL license.
That's not JCL's problem, that is freepascal and GPL's problem. Blame it on them.

- JCL is not in a runnable state for most FPC targets. - translated Delphi headers are typically not clean for cross platform (most notably 64-bit) use.

Jeff's unit may have been, for all you know, which is the unit that I supplied a uRL to in my first email. Which I cannot download, and am waiting for Jeff to reply. In the mean time, S. Fisher has already got a unit working while we were yapping.

L505

Unfortunately, it's slightly slower than my fastest program that uses the library that comes with FPC. I wonder if the dll was compiled with all of the C compiler's optimiaztions turned on.

May have to make an .O file..

For example static linking.. instead of dynamic.

AFAIK one can link directly to a .O (object) GCC file..

L505

No, not nonsense. Jeff P. (yetanothergeek) had a header for freepascal not using any JCL, AFAIK. Which is the original link I provided, which wasn't working. Which is why I posted this to the mailing list.

He did it for us already, AFAIK. No sense repeating Jeff's work if he already did it.

Because of the ridiculous GPL license which freepascal uses, which isn't as free as a BSD/NRCOL/MIT license.

That's not JCL's problem, that is freepascal and GPL's problem. Blame it on them.

Freedoom can go both ways. Although I think you are a BSD fan and I am patronizing anyway.

- JCL is not in a runnable state for most FPC targets. - translated Delphi headers are typically not clean for cross platform (most notably 64-bit) use.

Jeff's unit may have been, for all you know, which is the unit that I supplied a uRL to in my first email. Which I cannot download, and am waiting for Jeff to reply. In the mean time, S. Fisher has already got a unit working while we were yapping.

L505

For everbody who wants pcre, here's how I got it working.

From http://www.renatomancuso.com/software/dpcre/dpcre.htm,

download "PCRE 6.7 DLL".

To the file PCRE.pas, apply the attached patch file. If you don't, there will be errors during compiling.

I couldn't find any documentation. This quick and dirty little program may help.

{$mode delphi}
uses pcre;
function count_matches( const needle, haystack: ansistring ):longint;
var
  rx : iregex;
  capture: imatch;
  where: longint;
begin
  rx := regexCreate( needle );
  result := 0;
  where := 0;
  repeat
    capture := rx.match( haystack, where );
    if not capture.success then  break;
    inc( result );
    where := capture.groups[0].index + capture.groups[0].length;
  until false;
end;
var
  rx : iregex;
  capdata: imatch;
  md : idfamatchCollection;
  str: ansistring;
  i : integer;
begin
  str := 'Come back, Mac.';
  rx := regexCreate( '(\w*)a(\w*)' );
  capdata := rx.match( str );
  if capdata.success then
  begin
    writeln( 'found' );
    writeln( capdata.groups.count, ' captures' );
    for i:= 0 to capdata.groups.count - 1 do
      writeln( ' ', capdata.groups[i].index, ' ', capdata.groups[i].length,
        '  value: ', capdata.groups[i].value );
    writeln;
  end
  else
    writeln( 'not' );
  writeln;
  md := rx.dfaMatch( str );
  writeln( 'count: ', md.count );
  if md.count > 0 then
  begin
    writeln( 'dfamatch won' );
    writeln( md[0].index, ' ', md[0].length,' ', md[0].value );
  end
  else
    writeln( 'dfamatch failed' );
  writeln;
  writeln('counting matches');
  writeln( count_matches( 't', 'try to do it' ));
end.

__________________________________________________ Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com

Because of the ridiculous GPL license which freepascal uses, which isn't as free as a BSD/NRCOL/MIT license.

Please take your license flaming elsewhere, it is completely off topic on this list.

Jonas
FPC lists moderator