663 |
*************************************************/ |
*************************************************/ |
664 |
|
|
665 |
/* This is used both at compile and run-time to check for <xxx> escapes, where |
/* This is used both at compile and run-time to check for <xxx> escapes, where |
666 |
xxx is LF, CR, CRLF, or ANY. Print a message and return 0 if there is no match. |
xxx is LF, CR, CRLF, ANYCRLF, or ANY. Print a message and return 0 if there is |
667 |
|
no match. |
668 |
|
|
669 |
Arguments: |
Arguments: |
670 |
p points after the leading '<' |
p points after the leading '<' |
679 |
if (strncmp((char *)p, "cr>", 3) == 0) return PCRE_NEWLINE_CR; |
if (strncmp((char *)p, "cr>", 3) == 0) return PCRE_NEWLINE_CR; |
680 |
if (strncmp((char *)p, "lf>", 3) == 0) return PCRE_NEWLINE_LF; |
if (strncmp((char *)p, "lf>", 3) == 0) return PCRE_NEWLINE_LF; |
681 |
if (strncmp((char *)p, "crlf>", 5) == 0) return PCRE_NEWLINE_CRLF; |
if (strncmp((char *)p, "crlf>", 5) == 0) return PCRE_NEWLINE_CRLF; |
682 |
|
if (strncmp((char *)p, "anycrlf>", 8) == 0) return PCRE_NEWLINE_ANYCRLF; |
683 |
if (strncmp((char *)p, "any>", 4) == 0) return PCRE_NEWLINE_ANY; |
if (strncmp((char *)p, "any>", 4) == 0) return PCRE_NEWLINE_ANY; |
684 |
fprintf(f, "Unknown newline type at: <%s\n", p); |
fprintf(f, "Unknown newline type at: <%s\n", p); |
685 |
return 0; |
return 0; |
852 |
(void)pcre_config(PCRE_CONFIG_NEWLINE, &rc); |
(void)pcre_config(PCRE_CONFIG_NEWLINE, &rc); |
853 |
printf(" Newline sequence is %s\n", (rc == '\r')? "CR" : |
printf(" Newline sequence is %s\n", (rc == '\r')? "CR" : |
854 |
(rc == '\n')? "LF" : (rc == ('\r'<<8 | '\n'))? "CRLF" : |
(rc == '\n')? "LF" : (rc == ('\r'<<8 | '\n'))? "CRLF" : |
855 |
|
(rc == -2)? "ANYCRLF" : |
856 |
(rc == -1)? "ANY" : "???"); |
(rc == -1)? "ANY" : "???"); |
857 |
(void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc); |
(void)pcre_config(PCRE_CONFIG_LINK_SIZE, &rc); |
858 |
printf(" Internal link size = %d\n", rc); |
printf(" Internal link size = %d\n", rc); |
1448 |
fprintf(outfile, "Forced newline sequence: CRLF\n"); |
fprintf(outfile, "Forced newline sequence: CRLF\n"); |
1449 |
break; |
break; |
1450 |
|
|
1451 |
|
case PCRE_NEWLINE_ANYCRLF: |
1452 |
|
fprintf(outfile, "Forced newline sequence: ANYCRLF\n"); |
1453 |
|
break; |
1454 |
|
|
1455 |
case PCRE_NEWLINE_ANY: |
case PCRE_NEWLINE_ANY: |
1456 |
fprintf(outfile, "Forced newline sequence: ANY\n"); |
fprintf(outfile, "Forced newline sequence: ANY\n"); |
1457 |
break; |
break; |
2225 |
to advance the start offset, and continue. We won't be at the end of the |
to advance the start offset, and continue. We won't be at the end of the |
2226 |
string - that was checked before setting g_notempty. |
string - that was checked before setting g_notempty. |
2227 |
|
|
2228 |
Complication arises in the case when the newline option is "any". |
Complication arises in the case when the newline option is "any" or |
2229 |
If the previous match was at the end of a line terminated by CRLF, an |
"anycrlf". If the previous match was at the end of a line terminated by |
2230 |
advance of one character just passes the \r, whereas we should prefer the |
CRLF, an advance of one character just passes the \r, whereas we should |
2231 |
longer newline sequence, as does the code in pcre_exec(). Fudge the |
prefer the longer newline sequence, as does the code in pcre_exec(). |
2232 |
offset value to achieve this. |
Fudge the offset value to achieve this. |
2233 |
|
|
2234 |
Otherwise, in the case of UTF-8 matching, the advance must be one |
Otherwise, in the case of UTF-8 matching, the advance must be one |
2235 |
character, not one byte. */ |
character, not one byte. */ |
2248 |
obits = (d == '\r')? PCRE_NEWLINE_CR : |
obits = (d == '\r')? PCRE_NEWLINE_CR : |
2249 |
(d == '\n')? PCRE_NEWLINE_LF : |
(d == '\n')? PCRE_NEWLINE_LF : |
2250 |
(d == ('\r'<<8 | '\n'))? PCRE_NEWLINE_CRLF : |
(d == ('\r'<<8 | '\n'))? PCRE_NEWLINE_CRLF : |
2251 |
|
(d == -2)? PCRE_NEWLINE_ANYCRLF : |
2252 |
(d == -1)? PCRE_NEWLINE_ANY : 0; |
(d == -1)? PCRE_NEWLINE_ANY : 0; |
2253 |
} |
} |
2254 |
if ((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY && |
if (((obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANY || |
2255 |
|
(obits & PCRE_NEWLINE_BITS) == PCRE_NEWLINE_ANYCRLF) |
2256 |
|
&& |
2257 |
start_offset < len - 1 && |
start_offset < len - 1 && |
2258 |
bptr[start_offset] == '\r' && |
bptr[start_offset] == '\r' && |
2259 |
bptr[start_offset+1] == '\n') |
bptr[start_offset+1] == '\n') |