diplodocus.org Git - nmh/blob - h/fmt_scan.h

   1
   2 /*
   3  * fmt_scan.h -- definitions for fmt_scan()
   4  */
   5
   6 /*
   7  * This structure describes an "interesting" component.  It holds
   8  * the name & text from the component (if found) and one piece of
   9  * auxilary info.  The structure for a particular component is located
  10  * by (open) hashing the name and using it as an index into the ptr array
  11  * "wantcomp".  All format entries that reference a particular component
  12  * point to its comp struct (so we only have to do component specific
  13  * processing once.  e.g., parse an address.).
  14  *
  15  * In previous implementations "wantcomp" was made available to other
  16  * functions, but now it's private and is accessed via functions.
  17  */
  18 struct comp {
  19     char        *c_name;        /* component name (in lower case) */
  20     char        *c_text;        /* component text (if found)      */
  21     struct comp *c_next;        /* hash chain linkage             */
  22     short        c_flags;       /* misc. flags (from fmt_scan)    */
  23     short        c_type;        /* type info   (from fmt_compile) */
  24     union {
  25         struct tws *c_u_tws;
  26         struct mailname *c_u_mn;
  27     } c_un;
  28     int          c_refcount;    /* Reference count                */
  29 };
  30
  31 #define c_tws c_un.c_u_tws
  32 #define c_mn  c_un.c_u_mn
  33
  34 /*
  35  * c_type bits
  36  */
  37 #define CT_ADDR       (1<<0)    /* referenced as address    */
  38 #define CT_DATE       (1<<1)    /* referenced as date       */
  39
  40 #define CT_BITS       "\020\01ADDR\02DATE"
  41
  42 /*
  43  * c_flags bits
  44  */
  45 #define CF_TRUE       (1<<0)    /* usually means component is present */
  46 #define CF_PARSED     (1<<1)    /* address/date has been parsed */
  47 #define CF_DATEFAB    (1<<2)    /* datefield fabricated */
  48 #define CF_TRIMMED    (1<<3)    /* Component has been trimmed */
  49
  50 #define CF_BITS       "\020\01TRUE\02PARSED\03CF_DATEFAB\04TRIMMED"
  51
  52 /*
  53  * This structure defines one formatting instruction.
  54  */
  55 struct format {
  56     unsigned char f_type;
  57     char          f_fill;
  58     short         f_width;      /* output field width   */
  59     union {
  60         struct comp *f_u_comp;  /* associated component */
  61         char        *f_u_text;  /* literal text         */
  62         char         f_u_char;  /* literal character    */
  63         int          f_u_value; /* literal value        */
  64     } f_un;
  65     short         f_flags;      /* misc. flags          */
  66 };
  67
  68 #define f_skip f_width          /* instr to skip (false "if") */
  69
  70 #define f_comp  f_un.f_u_comp
  71 #define f_text  f_un.f_u_text
  72 #define f_char  f_un.f_u_char
  73 #define f_value f_un.f_u_value
  74
  75 /*
  76  * f_flags bits
  77  */
  78
  79 #define FF_STRALLOC     (1<<0)  /* String has been allocated */
  80 #define FF_COMPREF      (1<<1)  /* Component reference */
  81
  82 /*
  83  * prototypes used by the format engine
  84  */
  85
  86 /*
  87  * These are the definitions used by the callbacks for fmt_scan()
  88  */
  89
  90 typedef char * (*formataddr_cb)(char *, char *);
  91 typedef char * (*concataddr_cb)(char *, char *);
  92 typedef void (*trace_cb)(void *, struct format *, int, char *, char *);
  93
  94 struct fmt_callbacks {
  95     formataddr_cb       formataddr;
  96     concataddr_cb       concataddr;
  97     trace_cb            trace_func;
  98     void *              trace_context;
  99 };
 100
 101 /*
 102  * Create a new format string.  Arguments are:
 103  *
 104  * form         - Name of format file.  Will be searched by etcpath(), see that
 105  *                function for details.
 106  * format       - The format string to be used if no format file is given
 107  * default_fs   - The default format string to be used if neither form nor
 108  *                format is given
 109  *
 110  * This function also takes care of processing \ escapes like \n, \t, etc.
 111  *
 112  * Returns an allocated format string.
 113  */
 114
 115 char *new_fs (char *form, char *format, char *default_fs);
 116
 117 /*
 118  * Compile a format string into a set of format instructions.  Arguments are:
 119  *
 120  * fstring      - The format string (the "source code").
 121  * fmt          - Returns an allocated array of "struct fmt" elements.  Each
 122  *                struct fmt is one format instruction interpreted by the
 123  *                format engine.
 124  * reset        - If set to true, the format compiler will reset the
 125  *                component hash table.  The component hash table contains
 126  *                all of the references to message components refered to in
 127  *                the format instructions.  If you have multiple format
 128  *                strings that you want to compile and operate on the
 129  *                same message, this should be set to false.
 130  *
 131  * Returns the total number of components referenced by all format instructions
 132  * since the last reset of the hash table.
 133  */
 134
 135 int fmt_compile (char *fstring, struct format **fmt, int reset);
 136
 137 /*
 138  * Interpret a sequence of compiled format instructions.  Arguments are:
 139  *
 140  * format       - Array of format instructions generated by fmt_compile()
 141  * scanl        - Passed-in character array that will contain the output
 142  *                of the format instructions.  Is always terminated with
 143  *                a newline (\n).
 144  * max          - Maximum number of bytes to be written to "scanl" (in other
 145  *                words, the buffer size).  Includes the trailing NUL.
 146  * width        - Maximum number of displayed characters.  Does not include
 147  *                characters marked as nonprinting or (depending on the
 148  *                encoding) bytes in a multibyte encoding that exceed the
 149  *                character's column width.
 150  * dat          - An integer array that contains data used by certain format
 151  *                functions.  Currently the following instructions use
 152  *                dat[]:
 153  *
 154  *              dat[0] - %(msg), %(dat)
 155  *              dat[1] - %(cur)
 156  *              dat[2] - %(size)
 157  *              dat[3] - %(width)
 158  *              dat[4] - %(unseen)
 159  *
 160  * callbacks    - A set of a callback functions used by the format engine.
 161  *                Can be NULL.  If structure elements are NULL, a default
 162  *                function will be used.  Callback structure elements are:
 163  *
 164  *              formataddr      - A callback for the %(formataddr) instruction
 165  *              concataddr      - A callback for the %(concataddr) instruction
 166  *              trace           - Called for every format instruction executed
 167  *
 168  * The return value is a pointer to the next format instruction to
 169  * execute, which is currently always NULL.
 170  */
 171
 172 struct format *fmt_scan (struct format *format, char *scanl, size_t max,
 173                          int width, int *dat, struct fmt_callbacks *callbacks);
 174
 175 /*
 176  * Free a format structure and/or component hash table.  Arguments are:
 177  *
 178  * format       - An array of format structures allocated by fmt_compile,
 179  *                or NULL.
 180  * reset        - If true, reset and remove all references in the component
 181  *                hash table.
 182  */
 183
 184 void fmt_free (struct format *fmt, int reset);
 185
 186 /*
 187  * Free all of the component text structures in the component hash table
 188  */
 189
 190 void fmt_freecomptext(void);
 191
 192 /*
 193  * Search for a component structure in the component hash table.  Arguments are:
 194  *
 195  * component    - The name of the component to search for.  By convention
 196  *                all component names used in format strings are lower case,
 197  *                but for backwards compatibility this search is done in
 198  *                a case-SENSITIVE manner.
 199  *
 200  * This function returns a "struct comp" corresponding to the named component,
 201  * or NULL if the component is not found in the hash table.
 202  */
 203
 204 struct comp *fmt_findcomp(char *component);
 205
 206 /*
 207  * Search for a component structure in the component hash table.
 208  *
 209  * Identical to fmd_findcomp(), but is case-INSENSITIVE.
 210  */
 211
 212 struct comp *fmt_findcasecomp(char *component);
 213
 214 /*
 215  * Add a component entry to the component hash table
 216  *
 217  * component    - The name of the component to add to the hash table.
 218  *
 219  * If the component is already in the hash table, this function will do
 220  * nothing.  Returns 1 if a component was added, 0 if it already existed.
 221  */
 222
 223 int fmt_addcompentry(char *component);
 224
 225 /*
 226  * Add a string to a component hash table entry.  Arguments are:
 227  *
 228  * component    - The name of the component to add text to.  The component
 229  *                is searched for in a case-INSENSITIVE manner (note that
 230  *                this is different than fmt_findcomp()).  If the component
 231  *                is not found in the hash table, this function will silently
 232  *                return.
 233  * text         - The text to add to a component hash table entry.  Note that
 234  *                if the last character of the existing component
 235  *                text is a newline AND it is marked as an address
 236  *                component (the the CT_ADDR flag is set) existing
 237  *                component buffer is a newline, it will be separated
 238  *                from previous text by ",\n\t"; otherwise if the last
 239  *                character of the previous text is a newline it will
 240  *                simply be seperated by a "\t".  This unusual processing
 241  *                is designed to handle the case where you have multiple
 242  *                headers with the same name (e.g.: multiple "cc:" headers,
 243  *                even though that isn't technically allowed in the RFCs).
 244  *
 245  * This function is designed to be called when you start processing a new
 246  * component.  The function returns the integer value of the hash table
 247  * bucket corresponding to this component.  If there was no entry found
 248  * in the component hash table, this function will return -1.
 249  */
 250
 251 int fmt_addcomptext(char *component, char *text);
 252
 253 /*
 254  * Append to an existing component.  Arguments are:
 255  *
 256  * bucket       - The hash table bucket corresponding to this component,
 257  *                as returned by fmt_addcomp().  If -1, this function will
 258  *                return with no actions performed.
 259  * component    - The component to append text to.  Like fmt_addcomp, the
 260  *                component is searched case-INSENSITIVELY.
 261  * text         - The text to append to the component.  No special processing
 262  *                is done.
 263  *
 264  * This function is designed to be called when you are processing continuation
 265  * lines on the same header (state == FLDPLUS).
 266  */
 267
 268 void fmt_appendcomp(int bucket, char *component, char *text);
 269
 270 /*
 271  * The implementation of the %(formataddr) function.  This is available for
 272  * programs to provide their own local implementation if they wish to do
 273  * special processing (see uip/replsbr.c for an example).  Arguments are:
 274  *
 275  * orig         - Existing list of addresses
 276  * str          - New address(es) to append to list.
 277  *
 278  * This function returns an allocated string containing the new list of
 279  * addresses.
 280  */
 281
 282 char *formataddr(char *orig, char *str);
 283
 284 /*
 285  * The implementation of the %(concataddr) function.  Arguments and behavior
 286  * are the same as %(formataddr).  Again, see uip/replsbr.c to see how you
 287  * can override this behavior.
 288  */
 289
 290 char *concataddr(char *orig, char *str);