~drizzle-trunk/drizzle/development

Speciella användbara nya string-rutiner:

  bcmp(s1, s2, len) returns 0 if the "len" bytes starting at "s1" are

  identical to the "len" bytes starting at "s2", non-zero if they are

  different.

  bfill(dst, len, fill) moves "len" fill characters to "dst".

  Thus to set a buffer to 80 spaces, do bfill(buff, 80, ' ').

  bmove(dst, src, len) moves exactly "len" bytes from the source "src"

  to the destination "dst".  It does not check for NUL characters as

  strncpy() and stpncpy() do.

  bmove_upp(dst, src, len) moves exactly "len" bytes from the source

 "src-len" to the destination "dst-len" counting downwards.

  bzero(dst, len) moves "len" 0 bytes to "dst".

  Thus to clear a disc buffer to 0s do bzero(buffer, BUFSIZ).

  int2str(dst, radix, val)

  converts the (long) integer "val" to character form and moves it to

  the destination string "dst" followed by a terminating NUL.  The

  result is normally a pointer to this NUL character, but if the radix

  is dud the result will be NullS and nothing will be changed.

  If radix is -2..-36, val is taken to be SIGNED.

  If radix is  2.. 36, val is taken to be UNSIGNED.

  That is, val is signed if and only if radix is.  You will normally

  use radix -10 only through itoa and ltoa, for radix 2, 8, or 16

  unsigned is what you generally want.

  m_ctype.h

  A better inplementation of the UNIX ctype(3) library.

  Notes:   global.h should be included before ctype.h

    - Se efter i filen \c\local\include\m_ctype.h

    - Används istället för ctype.h för att klara internationella karakterer.

  m_string.h

  Använd instället för string.h för att supporta snabbare strängfunktioner.

  strintstr(src, from, pat) looks for an instance of pat in src

  backwards from pos from.  pat is not a regex(3) pattern, it is a literal

  string which must be matched exactly.

  The result 0 if the pattern was not found else it is the start char of

  the pattern counted from the begining of the string.

  strappend(dest, len, fill) appends fill-characters to a string so that

  the result length == len. If the string is longer than len it's

  trunked. The des+len character is allways set to NULL.

  strcat(s, t) concatenates t on the end of s.  There  had  better  be

  enough  room  in  the  space s points to; strcat has no way to tell.

  Note that strcat has to search for the end of s, so if you are doing

  a lot of concatenating it may be better to use strmov, e.g.

	stpcpy(stpcpy(stpcpy(stpcpy(s,a),b),c),d)

    rather than

	strcat(strcat(strcat(strcpy(s,a),b),c),d).

    strcat returns the old value of s.

	- Använd inte strcat, använd strmov (se ovan).

  strcend(s, c) returns a pointer to the  first  place  in  s	where  c

  occurs,  or a pointer to the end-null of s if c does not occur in s.

  strend(s) returns a character pointer to the NUL which ends s.  That

  is,  strend(s)-s  ==  strlen(s). This is useful for adding things at

  the end of strings.  It is redundant, because  strchr(s,'\0')  could

  strfill(dest, len, fill) makes a string of fill-characters. The result

  string is of length == len. The des+len character is allways set to NULL.

  strfill() returns pointer to dest+len;

  strfind(src, pat) looks for an instance of pat in src.  pat is not a

  regex(3) pattern, it is a literal string which must be matched exactly.

  The result is a pointer to the first character of the located instance,

  or NullS if pat does not occur in src.

  strmake(dst,src,length) moves length characters, or until end, of src to

  dst and appends a closing NUL to dst.

  strmake() returns pointer to closing null;

  stpcpy(dst, src) moves all the  characters  of  src  (including  the

  closing NUL) to dst, and returns a pointer to the new closing NUL in

  dst.	 The similar UNIX routine strcpy returns the old value of dst,

  which I have never found useful.  stpcpy(stpcpy(dst,a),b) moves a//b

  into dst, which seems useful.

  stpncpy(dst,src,length) moves length characters, or until end, of src to

  dst and appends a closing NUL to dst if src is shorter than length.

  The result is a pointer to the first NUL in dst, or is dst+n if dst was

  truncated.

 strrchr(s, c) returns a pointer to the  last  place  in  s	where  c

 occurs,  or  NullS if c does not occur in s. This function is called

 rindex in V7 and 4.?bsd systems.

 strrchr  looks  for single characters, not for sets or strings.

 strxmov(dst, src1, ..., srcn, NullS)

 moves the concatenation of src1,...,srcn to dst, terminates it

 with a NUL character, and returns a pointer to the terminating NUL.

 It is just like strmov except that it concatenates multiple sources.

 Beware: the last argument should be the null character pointer.

 Take VERY great care not to omit it!  Also be careful to use NullS

 and NOT to use 0, as on some machines 0 is not the same size as a

 character pointer, or not the same bit pattern as NullS.

 strxnmov(dst, len, src1, ..., srcn, NullS)

 moves the first len characters of the concatenation of src1,...,srcn

 to dst.  If there aren't that many characters, a NUL character will

 be added to the end of dst to terminate it properly.  This gives the

 same effect as calling strxcpy(buff, src1, ..., srcn, NullS) with a

 large enough buffer, and then calling stpncpy(dst, buff, len).

 It is just like stpncpy except that it concatenates multiple sources.

 Beware: the last argument should be the null character pointer.

 Take VERY great care not to omit it!  Also be careful to use NullS

 and NOT to use 0, as on some machines 0 is not the same size as a

 character pointer, or not the same bit pattern as NullS.

 Note: strxnmov is like stpncpy in that it always moves EXACTLY len

 characters; dst will be padded on the right with NUL characters as

 needed.  strxncpy does the same.  strxncat, like strncat, does NOT.

I mysys:

 stripp_sp(string str)

 Strips end-space from string and returns new length.

 strlength(const string str)

 Return length of string with end-space:s not counted.

 void caseup _A((string str,uint32_t length));

 void casedn _A((string str,uint32_t length));

 void caseup_str _A((string str));

 void casedn_str _A((string str));

 Converts strings or part of string to upper or lower-case.

 void case_sort _A((string str,uint32_t length));

 Converts string to a string with can be compared with strcmp() to

 get strings in rigth order.

 string strcfind(str,search)

 find string in another with no case_sensivity

 my_strcasecmp(s,t)

 Compare strings without regarding to case

 - For many strings it quicker to forst use case_sort on all strings and

   then compare them with strcmp().