Patching: Technical Discussion ...

[rizapn]

@DeadManS:
I also have no getFileSize() entry-point yet. But you can put very big size value in r15 (fileRead), and I think r4 will set to the real size value (you will read that file right? not only to check his size?). If you only want to check size, maybe you can use fileSeek (0xDFB868) or fileStat (0xDFF6AA) ... but I don't know their parameters (but only this, r12=fileHandle)...

ps:
- I can not install ICQ (it is firewalled), but Yahoo!Messanger. Maybe, better if we discuss more using that media?

[DeadManS]

fileOpen (0xDFA73E)
- r15:r14 = pointer to ASCIIZ filename
- r4 = Open result, if success r4=fileHandle, if failed, r4=FFFF

small correction R13:R12 pointer to Filename
another question, when i load the bin file whith this metod (Fileopen, fileread,fileclose) and try to play melody, phone is turned off, where is the trouble (i check, melody is loaded fine)? or file buffer not read ? If i plaice melody manualy (trought debugger), all work fine

[Seklth]

please, write calls:
1) check dictaphone
2) if not playing - send play to dictaphone

need calls to use in java

[lalo.lerry]

@MASTER:
I've a question:
I'm making a patch to show AM-PM hour clock, and I want to add a modified character for AM and PM, but till now I can insert only ASCII character,so...
how to insert different characters form standard ASCII ones (for S45i)?

[rizapn]

Non standard ASCII code

There are two different non-standard ASCII code:
1/ for code below than space (0x20) and greather than 0x7F
2/ using UNICODE (2 bytes).

I don't know exactly which ones is used by your AM/PM modification. Maybe you can write a test patch (to get the correct character and code), and get unusefull single byte character to be modified (not the UNICODE ones). Or maybe, you can also use V_Klay keyboard Layout feature to test them???

rizapn

[coollang]

About output unicode

I write a function like DrawASCIIText, you can use it to output unicode string.

    DrawWString:
      mov [-r0], r9
      mov [-r0], r8
      mov [-r0], r6
      sub r0, #24h ; '$'
      mov [-r0], r15
      mov [-r0], r14
      mov r14, r12
      mov r15, r13
      mov r12, #1Eh
      add r12, r0
      and r12, #3FFFh
      mov r13, DPP1
      calls 0C1h,07CF4h
      add r0, #4
      mov r6, #1
      mov [-r0], r6
      mov r12, [r0+30h]
      mov [-r0], r12
      mov r9, [r0+30h]
      mov r8, [r0+2Eh]
      mov [-r0], r9
      mov [-r0], r8
      mov r13, #0
      mov [-r0], r13
      mov r12, #0Ah
      add r12, r0
      and r12, #3FFFh
      mov r13, DPP1
      mov r14, #24h
      add r14, r0
      and r14, #3FFFh
      mov r15, DPP1
      calls 0C1h, 04A4Ch
      add r0, #0Ah
      mov r12, r0
      and r12, #3FFFh
      mov r13, DPP1
      calls 0C1h, 0400Ah
      mov r12, r0
      and r12, #3FFFh
      mov r13, DPP1
      calls 0C1h, 02F2Ah
      add r0, #24h 
      mov r6, [r0+]
      mov r8, [r0+]
      mov r9, [r0+]
      rets

you can call it by follow:

  mov [-r0], #fontStyle
  mov [-r0], #pag(WString)
  mov [-r0], #pof(WString)
  mov r12, #X  ; relatif position to r14
  mov r13, #Y
  mov r14, #drawAreaWidth
  mov r15, #drawAreaHeight
  calls drawString
  add  r0, #6
  rets
;-------------------------------------------------------------------; 
WString:
    dw pof(String), pag(String), 05h	
String:
    dw 04h,30h,31h,32h,33h

WString is a struct used by fireware, some function servo it.

struct WString{
BUF * pBuf;
int nBufCount;    //sizeof(BUF)/sizeof(WCHAR)
}
struct BUF{
int nSZlen;
WCHAR sz[nBufCount-1];
}

[rizapn]

@coollang: calls 0B3h, 0EFDEh ???

rizapn

[coollang]

I am sorry. it is a mistack.
because i would often switch at SLIKV56 and SLCKV55.
i write this function just for output chinese string.

[lalo.lerry]

@MASTER:
My patch just add a character in time string (and switch time to 12 hours format), using built-in show time routine.
I can easily add an ASCII character (yes, I know Siemens letter code isn't stricktly an ASCII one) but I would like not to modify any of them becuase someone could use it, so I'm trying to use a non ASCII character...
Anyway my first release will use two modifed ASCII ones.

@collang:
Thanks for routine but I don't need to draw a new string, just add one character to an existing one.
Morover, I think you are referring to SL45, not S45i...
Thank you anyway!

[lalo.lerry]

@MASTER:
May I know:
What is the real Siemens code?
I've seen it's neither strictly ASCII nor UNICODE...
and how can we add "UNICODE" support for call logging and similar patches?
Thanks

Lalo

[ntcn]

Originally Posted by lalo.lerry

@MASTER:
What is the real Siemens code?
I've seen it's neither strictly ASCII nor UNICODE...

In S45i at the address 0F218Ah (during an incoming call) we found name from addressbook. If the line consist only of latin symbols - line in a standard format, but if the line consist cyrillic symbols - in the beginning of a line added "81 nn 08" (nn - length of string in symbols) and cyrillic symbols are coded as a low byte of corresponding symbols in UTF8. That means "81 nn 08": the following nn symbols codes will be in UTF8 (low part) with a base code "08" ?

P.S. The unicode string can be taken from address 0F2742h, but there the line contains incoming call icons (home/work/etc), but the "pure" line for write to a log (in UTF8 format) is required.

p.s. sorry for my english

[rizapn]

UNICODE

*) Yes, it happened also in SL45iv56 (address 0008:2C56 for incoming, and 0008:2B9A for outgoing).
*) For call logging, maybe the easiest way to support UNICODE is saving all data in the UNICODE format. So, the file will be UNICODE file. I don't know how to write a UNICODE character in a normal (single byte) text file.

In Siemens string, the combination of ASCII and UNICODE is supported. So, not all character need to be converted to UNICODE format (saving much space).

rizapn

[ntcn]

Originally Posted by rizapn

UNICODE
*) For call logging, maybe the easiest way to support UNICODE is saving all data in the UNICODE format.

At address 0F2742h stored the string which is displayed on the screen when incoming call. The string in structure begin from 0F216Eh (type of a call, number length in byte, number, name from a addressbook, 91h if number begins with "+", etc) is stored in the mixed format, not unicode. For example, unicode string "0401h 0451h 20h" in this structure is stored as "81 03 08 81 D1 20", but unicode string "0031h 0032h 0033h" is stored as "31 32 33". Is not known that mean the first "81 nn 08" (next 81h and D1h - low part from coded in UTF8 unicode cyrillic symbols 0401h and 0451h) ? thanx.

[rizapn]

@ntcn:
Yes, I knew that. That's why I wrote: "In Siemens string, the combination of ASCII and UNICODE is supported. So, not all character need to be converted to UNICODE format (saving much space)." ...
What I was trying to say is : "I don't know how to create a text file (call logging is about to create a file right?) which has ASCII and UNICODE chars mixed ... ", so, creating a UNICODE file is easier ...

rizapn

[ntcn]

Originally Posted by rizapn

Yes, I knew that. ... What I was trying to say is : "I don't know how to create a text file which has ASCII and UNICODE chars mixed ... "

Algorithm of convert from unicode to UTF8 (UTF8 means use 1 byte for symbols with codes 0..80h and 2 bytes for symbols with codes more 80h, UTF8 read any programs which read unicode) is known. Help to define algorithm, if know, how from a format in which phone stores number in calls structure (for example, from " 81 02 08 81 90 D1") to convert in unicode (in " 0401h 0410h 0451h ")? It is impossible to simply delete the first 3 bytes and convert other - this example for Russian symbols, for other languages instead of " 81 nn 08 " will be other bytes - how to understand that "81 nn 08" should be converted to "04xxh" (unicode, russian symbols) ?

Sorry for such quantity of questions successively..