PureBasic Forums - English

http://www.purebasic.com
https://www.purebasic.fr/english/

[Implemented] Regular Expressions

https://www.purebasic.fr/english/viewtopic.php?t=3209

Page 2 of 3

Posted: Wed Jul 03, 2002 11:28 am
by BackupUser
Restored from previous forum. Originally posted by Danilo.

>But you didn't tell the compiler the value of #PB_EventCloseWindow and you're
>more than happy to use it... see what I mean?

This constant is not declared at runtime nor
at compile time. Its a fixed constants.
See what i mean ??

cya,
...Danilo

(registered PureBasic user)

Posted: Wed Jul 03, 2002 1:08 pm
by BackupUser
Restored from previous forum. Originally posted by Schlowski.

@Iontichy: This link only leads to an exact pattern match, nothing to do with regular expressions...

@Danilo: Is it only me or are your answers have a rude sound? Franco asked IMHO a simple question...

And no, I can't write these reg-ex stuff , mainly because of lack of interest for me and due to a great lack of time...

Greetings
Schlowski

Posted: Wed Jul 03, 2002 1:25 pm
by BackupUser
Restored from previous forum. Originally posted by Iontichy.

Ok, just another try...

http://www.stinga.com/downloads/sgregexp10.zip

This is a complete RegExp object library for use with
Visual Basic (or any other COM capable language, i think).
This library is completly free and includes the complete
source code (Visual Studio C++ project).

---------------
star traveler

Posted: Wed Jul 03, 2002 3:33 pm
by BackupUser
Restored from previous forum. Originally posted by fred.
The real Problem:
The PB compiler is very simple and i dont think Fred
will add this. The Compiler doesnt optimize the output,
the compiler cant find procedures itself (you have to
use "Declare" for that), so do you really think Fred
will add this ??
The fact than the PB compiler can't find procedure at compile time has nothing to do with optimizations. It's simply because PBCompiler is a one pass compiler. All the produced code is of course optimized (even if the coder doesn't see it). May be I should put a useless 'disable optimizations' flags to let the user see them...

Fred - AlphaSND

Posted: Wed Jul 03, 2002 5:24 pm
by BackupUser
Restored from previous forum. Originally posted by Franco.
...May be I should put a useless 'disable optimizations' flags to let the user see them...

Fred - AlphaSND
NO please, my little brain is already full...

I trust you, without knowing what's going on in the compiler


Have a nice day...
Franco

Sometimes you have to go a lonely way to accomplish genius things.

Posted: Wed Jul 03, 2002 5:27 pm
by BackupUser
Restored from previous forum. Originally posted by PB.

> This constant is not declared at runtime nor at compile time. Its a fixed
> constants.

You have a good point there, which I didn't consider.

> See what i mean ??

You being sarcastic? Let's not fight -- it was just an idea.


PB - Registered PureBasic Coder

Posted: Wed Jul 03, 2002 5:32 pm
by BackupUser
Restored from previous forum. Originally posted by Franco.
@Franco:
> Why not
> THIS library works only with STRINGS not with OS related stuff...

You were talking about ActiveX in your posting.
No word about Regular Expressions there..

Have fun,
...Danilo
Nop, I was talking about a real asm Library for PureBasic NOT ActiveX.
But I understood what you meant.

And at this time I don't have any interest in regular expressions, that's why I don't touch it. It's too much time consuming.



Have a nice day...
Franco

Sometimes you have to go a lonely way to accomplish genius things.

Posted: Wed Jul 03, 2002 5:59 pm
by BackupUser
Restored from previous forum. Originally posted by Franco.

At least 2 PureBasic Commands can be used for pattern matching:
FindString and ReplaceString.

But Regular Expression goes further.
It uses special characters for pattern matching.
The special characters are '\$', '^', '.', '*', '+', '?', '[', ']' and '\'.
And every character has a special function.
This characters are already used by PureBasic. (except '^' and '?')

I suppose it would be a huge task for Fred to allow that these characters can also be used by Regular Expression stuff.

For this you can use other languages like Perl or tools like grep.


Have a nice day...
Franco

Sometimes you have to go a lonely way to accomplish genius things.

Posted: Thu Jul 04, 2002 1:11 am
by BackupUser
Restored from previous forum. Originally posted by Sasa.

Hey, hey,

nice that this thread got rolling.

Well, as I looked on freshmeat.net and sourceforge.net there are lots of C, CPP, Java and other open source libraries. This shouldn't be the problem. To mention a few:

Grep of course is open source: http://www.gnu.org/software/grep/grep.html

If you filter for regular expressions AND Programming Language :: C AND Topic :: Software Development :: Libraries
you get: http://freshmeat.net/search/?q=regular expressions&section=projects&filter=164,809

There are 4 libraries written in C there. And that way I found the source for the mother of all libraries: PCRE - Perl Compatible Regular Expressions

http://www.pcre.org

You can do the same thing on sf.net of course. Get the idea?

So, there is enough source out there, but surely there would have to be a statement, that this is going to be implemented sometime. Maybe for 4.0 or 4.5 or 11.0 :wink:

I know this is a lot to ask for and this is why I won't ask for it :wink:

But I sure will hope for it. I will try to read through PCRE Code a little, but I am afraid, with my little knowledge of C I will have my difficulties.

All the best,
Sasa

Having no sig really is the best sig, isn't it?

Posted: Thu Jul 04, 2002 6:24 am
by BackupUser
Restored from previous forum. Originally posted by Danilo.

> The fact than the PB compiler can't find procedure at
> compile time has nothing to do with optimizations.
> It's simply because PBCompiler is a one pass compiler.

You can do that also with a 1-pass-compiler.
But hey, thats not a real problem.

>All the produced code is of course optimized
>(even if the coder doesn't see it).

I said "The Compiler doesnt optimize the output"
and thats true.

You want to know what i mean ??
OK, lets go:

Code: Select all

; file TEST2.PB
MessageRequester("INFO","Hello World!",0)
; end
pbcompiler.exe test2.pb /COMMENTED

Code: Select all

; result PureBasic.asm
; 
; PureBasic x86 3.20 generated code
; 
; (c) 2002 Fantaisie Software
; 
; The header must remain intact for Re-Assembly
; 
; Requester
EXTERN _ExitProcess@4
EXTERN _HeapAlloc@12
EXTERN _HeapCreate@12
EXTERN _HeapDestroy@4
EXTERN _HeapFree@12
EXTERN _GetModuleHandleA@4
; 
EXTERN _PB_MessageRequester@12
EXTERN _GetModuleFileNameA@12
EXTERN _SetCurrentDirectoryA@4
EXTERN _RtlZeroMemory@8
GLOBAL _PB_Instance
GLOBAL _PBV_ExecutableType
GLOBAL _PB_FullPath
GLOBAL _PB_DDrawBase
GLOBAL _PB_DirectInput
GLOBAL _PB_DirectSound
GLOBAL _PB_MemoryBase
GLOBAL  PB_Instance
GLOBAL  PB_MemoryBase
GLOBAL _PB_EndFunctions
GLOBAL _PB_StringBase
GLOBAL  PB_StringBase
GLOBAL _WinMain@16
; 
SEGMENT .text USE32 CLASS=CODE
; 
_WinMain@16
; 
  PUSH   dword I_BSSEnd-I_BSSStart
  PUSH   dword I_BSSStart
  CALL  _RtlZeroMemory@8
  CALL  _GetModuleHandleA@4
  MOV    [_PB_Instance],eax
  PUSH   dword 0
  CALL  _GetModuleHandleA@4
  MOV    [_PB_Instance],eax
  PUSH   dword 0
  PUSH   dword 4000
  PUSH   dword 0
  CALL  _HeapCreate@12
  MOV    [PB_MemoryBase],eax
  PUSH   dword 500
  PUSH   dword _PB_FullPath
  PUSH   dword [_PB_Instance]
  CALL  _GetModuleFileNameA@12
  MOV    eax, _PB_FullPath
  MOV    ecx,eax
PB_GetPathLoop
  MOV    dh,[eax]
  INC    eax
  CMP    dh,'\'
  JNE    PB_GetPathSkip
  MOV    ecx,eax
PB_GetPathSkip
  CMP    dh,0
  JNE    PB_GetPathLoop
  MOV    byte [ecx],0
  PUSH   dword PB_DataSectionStart
  POP    dword [PB_DataPointer]
; 
; InitString()
; 
  PUSH   dword 64000
  PUSH   dword 8
  PUSH   dword [PB_MemoryBase]
  CALL  _HeapAlloc@12
  MOV    [PB_StringBase],eax
; 
; MessageRequester("INFO","Hello World!",0)
  PUSH   dword 0
  PUSH   dword _S2
  PUSH   dword _S1
  CALL  _PB_MessageRequester@12
_PB_EOP
  CALL  _PB_EndFunctions
; 
; End Of Program
; 
  PUSH   dword [PB_MemoryBase]
  CALL  _HeapDestroy@4
  POP    ebx
  PUSH   dword 0
  CALL  _ExitProcess@4
_PB_EndFunctions
  RET
; 
; 
PB_AllocateString
  PUSH   ecx
  PUSH   edx
  MOV    eax,[ecx]
  CMP    eax,0
  JE     PB_AllocateStringSkip
  PUSH   eax
  PUSH   dword 0
  PUSH   dword [PB_MemoryBase]
  CALL  _HeapFree@12
PB_AllocateStringSkip
  MOV    eax,[PB_StringBase]
  POP    edx
  SUB    eax,edx
  PUSH   edx
  INC    eax
  PUSH   eax
  PUSH   dword 0
  PUSH   dword [PB_MemoryBase]
  CALL  _HeapAlloc@12
  POP    edx
  POP    ecx
  MOV    [ecx],eax
  MOV    [PB_StringBase],edx
PB_AllocateStringLoop
  MOV    cl,[edx]
  INC    edx
  MOV    [eax],cl
  INC    eax
  CMP    cl,0
  JNE    PB_AllocateStringLoop
  RET
; 
PB_FreeString
  CMP    edx,0
  JE     PB_FreeStringEnd
  PUSH   eax
  PUSH   edx
  PUSH   dword 0
  PUSH   dword [PB_MemoryBase]
  CALL  _HeapFree@12
  POP    eax
PB_FreeStringEnd
  RET
; 
; 
PB_CopyString
  MOV    ecx,[PB_StringBase]
  CMP    edx,0
  JE     PB_CopyStringEnd_Null
PB_CopyStringLoop
  MOV    al,[edx]
  INC    edx
  MOV    [ecx],al
  INC    ecx
  CMP    al,0
  JNE    PB_CopyStringLoop
  DEC    ecx
  MOV    [PB_StringBase],ecx
  RET
PB_CopyStringEnd_Null
  MOV    [ecx],edx
  RET
; 
SEGMENT .data CLASS=DATA
; 
_PBV_ExecutableType: dd 0
_S1: db "INFO",0
_S2: db "Hello World!",0
PB_NullString: db 0
; 
SEGMENT .bss CLASS=DATA2
; 
I_BSSStart
_PB_MemoryBase:
PB_MemoryBase: RESD 1
_PB_StringBase:
PB_StringBase: RESD 1
_PB_Instance:
PB_Instance: RESD 1
_PB_DDrawBase: RESD 1
_PB_DirectInput: RESD 1
_PB_DirectSound: RESD 1
PB_FPUControl: RESW 1
_PB_FullPath: RESB 500
; 
  ALIGNB 4
PB_DataPointer RESD 1
  ALIGNB 4
I_BSSEnd
SEGMENT .data CLASS=DATA:
PB_DataSectionStart:
You dont need all that init stuff in this example.
You dont need the String stuff, you dont need _RtlZeroMemory@8,
you dont need the Path things, you dont need the Heap stuff, ...

At least 25% bull in the output.
I really dont call that optimized output, but i could live
with it. Its ~1k init-stuff, OK.

Frederic: thats not all, sorry.

Now we going deeper... and test how good
pbcompiler really is:

Code: Select all

; file TEST.pb
MessageRequester("INFO","Hello World!",0)


Procedure myBeep()
   beep_(1400,300)
   beep_(1000,200)
EndProcedure


Procedure consoletest()
   OpenConsole()
   For a = 0 To 20
    PrintN("Hello World from Console")
   Next a
   Input()
   CloseConsole()
EndProcedure
; end


pbcompiler test.pb /COMMENTED

Code: Select all

; result PureBasic.asm
; 
; PureBasic x86 3.20 generated code
; 
; (c) 2002 Fantaisie Software
; 
; The header must remain intact for Re-Assembly
; 
; Console
; Requester
EXTERN _ExitProcess@4
EXTERN _HeapAlloc@12
EXTERN _HeapCreate@12
EXTERN _HeapDestroy@4
EXTERN _HeapFree@12
EXTERN _GetModuleHandleA@4
EXTERN _Beep@8
; 
EXTERN PB_CloseConsole
EXTERN PB_Input
EXTERN _PB_MessageRequester@12
EXTERN PB_OpenConsole
EXTERN PB_PrintN
EXTERN _GetModuleFileNameA@12
EXTERN _SetCurrentDirectoryA@4
EXTERN _RtlZeroMemory@8
GLOBAL _PB_Instance
GLOBAL _PBV_ExecutableType
GLOBAL _PB_FullPath
GLOBAL _PB_DDrawBase
GLOBAL _PB_DirectInput
GLOBAL _PB_DirectSound
GLOBAL _PB_MemoryBase
GLOBAL  PB_Instance
GLOBAL  PB_MemoryBase
GLOBAL _PB_EndFunctions
GLOBAL _PB_StringBase
GLOBAL  PB_StringBase
GLOBAL _WinMain@16
; 
SEGMENT .text USE32 CLASS=CODE
; 
_WinMain@16
; 
  PUSH   dword I_BSSEnd-I_BSSStart
  PUSH   dword I_BSSStart
  CALL  _RtlZeroMemory@8
  CALL  _GetModuleHandleA@4
  MOV    [_PB_Instance],eax
  PUSH   dword 0
  CALL  _GetModuleHandleA@4
  MOV    [_PB_Instance],eax
  PUSH   dword 0
  PUSH   dword 4000
  PUSH   dword 0
  CALL  _HeapCreate@12
  MOV    [PB_MemoryBase],eax
  PUSH   dword 500
  PUSH   dword _PB_FullPath
  PUSH   dword [_PB_Instance]
  CALL  _GetModuleFileNameA@12
  MOV    eax, _PB_FullPath
  MOV    ecx,eax
PB_GetPathLoop
  MOV    dh,[eax]
  INC    eax
  CMP    dh,'\'
  JNE    PB_GetPathSkip
  MOV    ecx,eax
PB_GetPathSkip
  CMP    dh,0
  JNE    PB_GetPathLoop
  MOV    byte [ecx],0
  PUSH   dword PB_DataSectionStart
  POP    dword [PB_DataPointer]
; 
; InitString()
; 
  PUSH   dword 64000
  PUSH   dword 8
  PUSH   dword [PB_MemoryBase]
  CALL  _HeapAlloc@12
  MOV    [PB_StringBase],eax
; 
; MessageRequester("INFO","Hello World!",0)
  PUSH   dword 0
  PUSH   dword _S2
  PUSH   dword _S1
  CALL  _PB_MessageRequester@12
; 
; Procedure myBeep()
  JMP    NEAR _EndProcedure0
_Procedure0
  PUSH   ebx
  PUSH   ecx
  PUSH   ebp
  PUSH   esi
  PUSH   edi                                                                                                                                                                                                                                                           
; beep_(1400,300)
  PUSH   dword 300
  PUSH   dword 1400
  CALL  _Beep@8
; beep_(1000,200)
  PUSH   dword 200
  PUSH   dword 1000
  CALL  _Beep@8
; EndProcedure
  XOR    eax,eax
_EndProcedure1
  POP    edi
  POP    esi
  POP    ebp
  POP    ecx
  POP    ebx
  RET
_EndProcedure0
; 
; Procedure consoletest()
  JMP    NEAR _EndProcedure2
_Procedure2
  PUSH   ebx
  PUSH   ecx
  PUSH   ebp
  PUSH   esi
  PUSH   edi
  MOV    esi,esp
  SUB    esp,4
  MOV    eax,esp
  MOV    edx,eax
  ADD    edx,4
_ClearLoop3
  MOV    dword [eax],0
  ADD    eax,4
  CMP    eax,edx
  JNE   _ClearLoop3                                                                                                                      
; OpenConsole()
  CALL   PB_OpenConsole
; For a = 0 To 20
  MOV    dword [esp+0],0
_For1
  MOV    eax,20
  CMP    eax,dword [esp+0]
  JL      NEAR _Next2
; PrintN("Hello World from Console")
  MOV    eax,dword _S3
  CALL   PB_PrintN
; Next a
  ADD    dword [esp+0],1
  JMP    NEAR _For1
_Next2
; Input()
  PUSH   dword [PB_StringBase]
  CALL   PB_Input
  POP    dword [PB_StringBase]
; CloseConsole()
  CALL   PB_CloseConsole
; EndProcedure
  XOR    eax,eax
_EndProcedure3
  ADD    esp,4
  POP    edi
  POP    esi
  POP    ebp
  POP    ecx
  POP    ebx
  RET
_EndProcedure2
; 
_PB_EOP
  CALL  _PB_EndFunctions
; 
; End Of Program
; 
  PUSH   dword [PB_MemoryBase]
  CALL  _HeapDestroy@4
  POP    ebx
  PUSH   dword 0
  CALL  _ExitProcess@4
_PB_EndFunctions
  RET
; 
; 
PB_AllocateString
  PUSH   ecx
  PUSH   edx
  MOV    eax,[ecx]
  CMP    eax,0
  JE     PB_AllocateStringSkip
  PUSH   eax
  PUSH   dword 0
  PUSH   dword [PB_MemoryBase]
  CALL  _HeapFree@12
PB_AllocateStringSkip
  MOV    eax,[PB_StringBase]
  POP    edx
  SUB    eax,edx
  PUSH   edx
  INC    eax
  PUSH   eax
  PUSH   dword 0
  PUSH   dword [PB_MemoryBase]
  CALL  _HeapAlloc@12
  POP    edx
  POP    ecx
  MOV    [ecx],eax
  MOV    [PB_StringBase],edx
PB_AllocateStringLoop
  MOV    cl,[edx]
  INC    edx
  MOV    [eax],cl
  INC    eax
  CMP    cl,0
  JNE    PB_AllocateStringLoop
  RET
; 
PB_FreeString
  CMP    edx,0
  JE     PB_FreeStringEnd
  PUSH   eax
  PUSH   edx
  PUSH   dword 0
  PUSH   dword [PB_MemoryBase]
  CALL  _HeapFree@12
  POP    eax
PB_FreeStringEnd
  RET
; 
; 
PB_CopyString
  MOV    ecx,[PB_StringBase]
  CMP    edx,0
  JE     PB_CopyStringEnd_Null
PB_CopyStringLoop
  MOV    al,[edx]
  INC    edx
  MOV    [ecx],al
  INC    ecx
  CMP    al,0
  JNE    PB_CopyStringLoop
  DEC    ecx
  MOV    [PB_StringBase],ecx
  RET
PB_CopyStringEnd_Null
  MOV    [ecx],edx
  RET
; 
SEGMENT .data CLASS=DATA
; 
_PBV_ExecutableType: dd 0
_S1: db "INFO",0
_S2: db "Hello World!",0
_S3: db "Hello World from Console",0
PB_NullString: db 0
; 
SEGMENT .bss CLASS=DATA2
; 
I_BSSStart
_PB_MemoryBase:
PB_MemoryBase: RESD 1
_PB_StringBase:
PB_StringBase: RESD 1
_PB_Instance:
PB_Instance: RESD 1
_PB_DDrawBase: RESD 1
_PB_DirectInput: RESD 1
_PB_DirectSound: RESD 1
PB_FPUControl: RESW 1
_PB_FullPath: RESB 500
; 
  ALIGNB 4
PB_DataPointer RESD 1
  ALIGNB 4
I_BSSEnd
SEGMENT .data CLASS=DATA:
PB_DataSectionStart:
Booooaaahhhh!!

All procedures are still included in the output,
even if you dont need it.
You cant include a file with all your standard-procedures
in a PureBasic project.

The 1st example is 4,03k .EXE and the
2nd is 6,03k.
Both examples are only a little MessageBox
with "Hello World!" - no memory needed, no string
operations needed, no Path thing needed and the
2 procedures are not used.

Fred (no offence!), you know that this is NOT
optimized output. Everybody knows that.
Thats why i said "The PB compiler is very simple".

If you want to make a very good compiler
that can compete with C and other languages
you should make the output optimized.
The init-stuff isnt the big thing, but the
thing with the procedures is really bad.
The compiler can check easily if a procedure
is used or not, so it could include only the
needed stuff.

PureBasic is GOOD -> and most important: ITS FUN.

With some little optimizations you could make it VERY GOOD -
better than most other compilers.

>May be I should put a useless 'disable optimizations'
>flags to let the user see them...

Well... thats your idea - just do it.
I think you understand what i´m talking about.

Beginners and people that dont know much internals
say something like
>"I trust you, without knowing what's going on in the compiler"

The programmer who knows how all the stuff works
can only laught about such comments.


As i said:
PureBasic IS good, but it could be better.

I wouldnt stay here when i think that PB is bull.
I DONT want to fight here (hi PB ), but i want
to discuss some things.
If we cant talk about some things here, we dont need
this forum - right ??

Communication is very important.

That brings me to another topic: DOUBLES.
Frederic: Many, many people asked about that,
but every time somebody is asking for doubles
you go away and hide (in the chat).
Why cant you answer this ??
I dont see a reason, so i dont understand
why you cant answer that.
Its also annoying for the user to ask again
and again - and its not good for your nerves.

Cant you simply say "PureBasic will never support
doubles" or "I´ll try to implement doubles for
PureBasic 3.5" or something like that ??

To get no answer is very bad. Million peoples
are waiting for it, and nobody knows the answer.


Remember: No offence, no fighting - only simple communication.

PEACE!

cya,
...Danilo

(registered PureBasic user)

Posted: Thu Jul 04, 2002 6:49 am
by BackupUser
Restored from previous forum. Originally posted by PB.

> Remember: No offence, no fighting - only simple communication.

I agree!

And I'm sorry that some of my suggestions generate such controversy!


PB - Registered PureBasic Coder

Edited by - PB on 04 July 2002 07:50:40

Posted: Mon Jul 08, 2002 5:36 pm
by BackupUser
Restored from previous forum. Originally posted by Franco.
Beginners and people that dont know much internals say something like
>"I trust you, without knowing what's going on in the compiler"

Well, actually I DON'T WANT TO KNOW THE INTERNALS.
It's too much time consuming...

I want to have a LIFE besides programming... (without a computer - there are other things to learn and to enjoy...)



Have a nice day...
Franco

Sometimes you have to go a lonely way to accomplish genius things.

Posted: Mon Jul 08, 2002 6:06 pm
by BackupUser
Restored from previous forum. Originally posted by fred.

Hello Danilo :)

All you say about the compiler output is true (of course), I just don't see optimizations like you do.. You considering a little space overhead (mainly String functions) and I was considering speed. Anyway, I wont go deeply on 'How pb internally optimize the code' and I will take in count your remarks to remove the unused procedure and init stuff of the produced code. Just think than writing a compiler for a general purpose high level langange isn't the same at all than writing the program directly in asm. You have to be flexible enough and generic which produce some overhead. Anyway PB still has a respectable size and speed generation .

Fred - AlphaSND

Posted: Mon Jul 08, 2002 6:11 pm
by BackupUser
Restored from previous forum. Originally posted by fred.

Ooops, just forget to answer about doubles: it's planned but not for now as it need a complete expression processor rewrite. Probably with v4.0 (Hi Paul ! ).

Fred - AlphaSND

Posted: Mon Jul 08, 2002 6:12 pm
by BackupUser
Restored from previous forum. Originally posted by Franco.

Let's discuss one little question:
"What is PureBasic - Compiler or Translator?"

A Compiler takes code written in a particular syntax and convert it to machine code instructions.
PureBasic is not able to generate machine code instructions by itself.
It uses NASM to generate executable files.
NASM is a compiler (in this special case also called an assembler) because it creates machine code from assembler.

So PureBasic is a translator...


Have a nice day...
Franco

Sometimes you have to go a lonely way to accomplish genius things.
All times are UTC+01:00
Page 2 of 3

Powered by phpBB® Forum Software © phpBB Limited