Mercurial > hg > xemacs-beta
view lib-src/yow.c @ 4604:e0a8715fdb1f
Support new IGNORE-INVALID-SEQUENCESP argument, #'query-coding-region.
lisp/ChangeLog addition:
2009-02-07 Aidan Kehoe <kehoea@parhasard.net>
* coding.el (query-coding-clear-highlights):
Rename the BUFFER argument to BUFFER-OR-STRING, describe it as
possibly being a string in its documentation.
(default-query-coding-region):
Add a new IGNORE-INVALID-SEQUENCESP argument, document that this
function does not support it.
Bind case-fold-search to nil, we don't want this to influence what the
function thinks is encodable or not.
(query-coding-region):
Add a new IGNORE-INVALID-SEQUENCESP argument, document what it
does; reflect this new argument in the associated compiler macro.
(query-coding-string):
Add a new IGNORE-INVALID-SEQUENCESP argument, document what it
does. Support the HIGHLIGHT argument correctly.
* unicode.el (unicode-query-coding-region):
Add a new IGNORE-INVALID-SEQUENCESP argument, document what it
does, implement this. Document a potential problem.
Use #'query-coding-clear-highlights instead of reimplementing it
ourselves.
Remove some debugging messages.
* mule/arabic.el (iso-8859-6):
* mule/cyrillic.el (iso-8859-5):
* mule/greek.el (iso-8859-7):
* mule/hebrew.el (iso-8859-8):
* mule/latin.el (iso-8859-2):
* mule/latin.el (iso-8859-3):
* mule/latin.el (iso-8859-4):
* mule/latin.el (iso-8859-14):
* mule/latin.el (iso-8859-15):
* mule/latin.el (iso-8859-16):
* mule/latin.el (iso-8859-9):
* mule/latin.el (windows-1252):
* mule/mule-coding.el (iso-8859-1):
Avoid the assumption that characters not given an explicit mapping
in these coding systems map to the ISO 8859-1 characters
corresponding to the octets on disk; this makes it much more
reasonable to implement the IGNORE-INVALID-SEQUENCESP argument to
query-coding-region.
* mule/mule-cmds.el (set-language-info):
Correct the docstring.
* mule/mule-cmds.el (finish-set-language-environment):
Treat invalid Unicode sequences produced from
invalid-sequence-coding-system and corresponding to control
characters the same as control characters in redisplay.
* mule/mule-cmds.el:
Document that encode-coding-char is available in coding.el
* mule/mule-coding.el (make-8-bit-generate-helper):
Change to return the both the encode-program generated and the
relevant non-ASCII charset; update the docstring to reflect this.
* mule/mule-coding.el
(make-8-bit-generate-encode-program-and-skip-chars-strings):
Rename this function; have it return skip-chars-strings as well as
the encode program. Have these skip-chars-strings use ranges for
charsets, where possible.
* mule/mule-coding.el (make-8-bit-create-decode-encode-tables):
Revise this to allow people to specify explicitly characters that
should be undefined (= corresponding to keys in
unicode-error-default-translation-table), and treating unspecified
octets above #x7f as undefined by default.
* mule/mule-coding.el (8-bit-fixed-query-coding-region):
Add a new IGNORE-INVALID-SEQUENCESP argument, implement support
for it using the 8-bit-fixed-invalid-sequences-skip-chars coding
system property; remove some debugging messages.
* mule/mule-coding.el (make-8-bit-coding-system):
This function is dumped, autoloading it makes no sense.
Document what happens when characters above #x7f are not
specified, implement this.
* mule/vietnamese.el:
Correct spelling.
tests/ChangeLog addition:
2009-02-07 Aidan Kehoe <kehoea@parhasard.net>
* automated/query-coding-tests.el:
Add FAILING-CASE arguments to the Assert calls, making #'q-c-debug
mostly unnecessary. Remove #'q-c-debug.
Add new tests that use the IGNORE-INVALID-SEQUENCESP argument to
#'query-coding-region; rework the existing ones to respect it.
author | Aidan Kehoe <kehoea@parhasard.net> |
---|---|
date | Sat, 07 Feb 2009 17:13:37 +0000 |
parents | 943eaba38521 |
children |
line wrap: on
line source
/* * yow.c * * Print a quotation from Zippy the Pinhead. * Qux <Kaufman-David@Yale> March 6, 1986 * * With dynamic memory allocation. */ /* Synched up with: FSF 19.28. */ #include <config.h> #include <stdio.h> #include <ctype.h> #include <../src/paths.h> /* For PATH_DATA. */ #if __STDC__ || defined(STDC_HEADERS) #include <stdlib.h> #include <unistd.h> #include <string.h> #include <time.h> /* for time() */ #endif #define BUFSIZE 80 #define SEP '\0' #ifndef YOW_FILE #define YOW_FILE "yow.lines" #endif void yow (FILE *fp); void setup_yow (FILE *fp); #ifdef WIN32_NATIVE #define rootrelativepath(rel) \ ({\ static char res[BUFSIZE], *p;\ strcpy (res, argv[0]);\ p = res + strlen (res);\ while (p != res && *p != '/' && *p != '\\' && *p != ':') p--;\ strcpy (p + 1, "../");\ strcpy (p + 4, rel);\ &res;}) #endif int main (int argc, char *argv[]) { FILE *fp; char file[BUFSIZ]; if (argc > 2 && !strcmp (argv[1], "-f")) strcpy (file, argv[2]); else #ifdef PATH_DATA #ifdef vms sprintf (file, "%s%s", PATH_DATA, YOW_FILE); #else sprintf (file, "%s/%s", PATH_DATA, YOW_FILE); #endif #else /* !PATH_DATA */ { fprintf (stderr, "%s: the location of the \"%s\" file was not supplied at compile-time.\n\ You must supply it with the -f command-line option.\n", argv[0], YOW_FILE); exit (1); } #endif if ((fp = fopen(file, "r")) == NULL) { perror(file); exit(1); } /* initialize random seed */ srand((int) (getpid() + time((time_t *) 0))); setup_yow(fp); yow(fp); fclose(fp); return 0; } static long len = -1; static long header_len; #define AVG_LEN 40 /* average length of a quotation */ /* Sets len and header_len */ void setup_yow (FILE *fp) { int c; /* Get length of file */ /* Because the header (stuff before the first SEP) can be very long, * thus biasing our search in favor of the first quotation in the file, * we explicitly skip that. */ while ((c = getc(fp)) != SEP) { if (c == EOF) { fprintf(stderr, "File contains no separators.\n"); exit(2); } } header_len = ftell(fp); if (header_len > AVG_LEN) header_len -= AVG_LEN; /* allow the first quotation to appear */ if (fseek(fp, 0L, 2) == -1) { perror("fseek 1"); exit(1); } len = ftell(fp) - header_len; } /* go to a random place in the file and print the quotation there */ void yow (FILE *fp) { long offset; int c, i = 0; char *buf; int bufsize; offset = rand() % len + header_len; if (fseek(fp, offset, 0) == -1) { perror("fseek 2"); exit(1); } /* Read until SEP, read next line, print it. (Note that we will never print anything before the first separator.) If we hit EOF looking for the first SEP, just recurse. */ while ((c = getc(fp)) != SEP) if (c == EOF) { yow(fp); return; } /* Skip leading whitespace, then read in a quotation. If we hit EOF before we find a non-whitespace char, recurse. */ while (isspace(c = getc(fp))) ; if (c == EOF) { yow(fp); return; } bufsize = BUFSIZE; buf = (char *) malloc(bufsize); if (buf == (char *)0) { fprintf(stderr, "can't allocate any memory\n"); exit (3); } buf[i++] = c; while ((c = getc(fp)) != SEP && c != EOF) { buf[i++] = c; if (i == bufsize-1) { /* Yow! Is this quotation too long yet? */ bufsize *= 2; buf = (char *)realloc(buf, bufsize); if (buf == (char *)0) { fprintf(stderr, "can't allocate more memory\n"); exit (3); } } } buf[i++] = 0; printf("%s\n", buf); }