Mercurial > hg > xemacs-beta
view lib-src/make-po.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 | 576fb035e263 |
children |
line wrap: on
line source
/* Generate .po file from doc-string file. Scan specified doc-string file, creating .po format messages for processing with msgfmt. The results go to standard output or to a file specified with -a or -o (-a to append, -o to start from nothing). Kludge to make up for shortcoming in make-docfile and Snarf-documentation: If arg before input filename is -p, we are scanning an add-on package, which requires slightly different processing. */ #include <stdio.h> #include <stdlib.h> #ifndef EXIT_SUCCESS #define EXIT_SUCCESS 0 #define EXIT_FAILURE 1 #endif /* #define BUFSIZE 8192 */ /* #define BUFSIZE 16384 */ #define BUFSIZE 32768 #define NEWSTRING 31 /* Character signalling start of new doc string */ #define LINEEND "\\n" #define ENDSTRING "\"\n" #define LINEBEGIN " \"" #define LINEBREAK ENDSTRING LINEBEGIN /* some brain-dead headers define this ... */ #undef FALSE #undef TRUE enum boolean { FALSE, TRUE }; /***********************/ /* buffer pseudo-class */ /***********************/ typedef struct _buffer { size_t index; /* current position in buf[] */ size_t size; /* size of buf */ char *buf; } buffer_struct; #define BUF_NULL {0, 0, NULL} int buf_init (buffer_struct *buffer, size_t size); void buf_free (buffer_struct *buffer); void buf_clear (buffer_struct *buffer); int buf_putc (buffer_struct *buffer, int c); int buf_print (buffer_struct *buffer, const char *s); /********************/ /* global variables */ /********************/ FILE *infile = NULL; FILE *outfile = NULL; buffer_struct buf = BUF_NULL; void scan_file (enum boolean package); void initialize (void); void clean_exit (int status); void buf_putc_safe (int c); void buf_print_safe (const char *s); void terminate_string (void); main (int argc, char *argv[]) { register int i; enum boolean package = FALSE; /* TRUE if scanning add-on package */ initialize (); outfile = stdout; /* If first two args are -o FILE, output to FILE. */ i = 1; if (argc > i + 1 && strcmp (argv[i], "-o") == 0) { outfile = fopen (argv[++i], "w"); ++i; } /* ...Or if args are -a FILE, append to FILE. */ if (argc > i + 1 && strcmp (argv[i], "-a") == 0) { outfile = fopen (argv[++i], "a"); ++i; } if (!outfile) { fprintf (stderr, "Unable to open output file %s\n", argv[--i]); return 1; } if (argc > i && !strcmp (argv[i], "-p")) { package = TRUE; ++i; } infile = fopen (argv[i], "r"); if (!infile) { fprintf (stderr, "Unable to open input file %s\n", argv[i]); return 1; } scan_file (package); clean_exit (EXIT_SUCCESS); } void scan_file (enum boolean package) { register int c; /* Character read in */ fprintf (outfile, "###############\n"); fprintf (outfile, "# DOC strings #\n"); fprintf (outfile, "###############\n"); while (c = getc (infile), !feof (infile)) { if (c == NEWSTRING) { /* If a string was being processed, terminate it. */ if (buf.index > 0) terminate_string (); /* Skip function or variable name. */ while (c != '\n') c = getc (infile); c = getc (infile); /* Begin a new string. */ fprintf (outfile, "msgid \""); buf_print_safe ("msgstr \""); } if (c == '\n') { /* Peek at next character. */ c = getc (infile); ungetc (c, infile); /* For add-on (i.e., non-preloaded) documentation, ignore the last carriage return of a string. */ if (!(package && c == NEWSTRING)) { fprintf (outfile, LINEEND); buf_print_safe (LINEEND); } /* If not end of string, continue it on the next line. */ if (c != NEWSTRING) { fprintf (outfile, LINEBREAK); buf_print_safe (LINEBREAK); } } else { /* If character is \ or ", precede it by a backslash. */ if (c == '\\' || c == '\"') { putc ('\\', outfile); buf_putc_safe ('\\'); } putc (c, outfile); buf_putc_safe (c); } } terminate_string (); } /* initialize sets up the global variables. */ void initialize (void) { if (buf_init (&buf, BUFSIZE) != 0) clean_exit (EXIT_FAILURE); } /* clean_exit returns any resources and terminates the program. An error message is printed if status is EXIT_FAILURE. */ void clean_exit (int status) { if (buf.size > 0) buf_free (&buf); if (outfile) fclose (outfile); if (infile) fclose (infile); if (status == EXIT_FAILURE) fprintf (stderr, "make-po abnormally terminated\n"); exit (status); } /* buf_putc_safe writes the character c on the global buffer buf, checking to make sure that the operation was successful. */ void buf_putc_safe (int c) { register int status; status = buf_putc (&buf, c); if (status == EOF) clean_exit (EXIT_FAILURE); } /* buf_putc_safe writes the string s on the global buffer buf, checking to make sure that the operation was successful. */ void buf_print_safe (const char *s) { register int status; status = buf_print (&buf, s); if (status < 0) clean_exit (EXIT_FAILURE); } /* terminate_string terminates the current doc string and outputs the buffer. */ void terminate_string (void) { fprintf (outfile, ENDSTRING); /* Make the "translation" different from the original string. */ buf_print_safe ("_X"); buf_print_safe (ENDSTRING); fprintf (outfile, "%s", buf.buf); buf_clear (&buf); } /*********************************/ /* buffer pseudo-class functions */ /*********************************/ /* buf_init initializes a buffer to the specified size. It returns non-zero if the attempt fails. */ int buf_init (buffer_struct *buffer, size_t size) { buffer->buf = malloc (size); if (buffer->buf == NULL) return 1; buffer->size = size; buf_clear (buffer); return 0; } /* buf_free releases the memory allocated for the buffer. */ void buf_free (buffer_struct *buffer) { free (buffer->buf); buffer->size = 0; } /* buf_clear resets a buffer to an empty string. */ void buf_clear (buffer_struct *buffer) { buffer->index = 0; buffer->buf[0] = '\0'; } /* buf_putc writes the character c on the buffer. It returns the character written, or EOF for error. */ int buf_putc (buffer_struct *buffer, int c) { if (buffer->index >= buffer->size) return EOF; buffer->buf[buffer->index++] = c; return c; } /* buf_print writes the string s on the buffer. It returns the number of characters written, or negative if an error occurred. */ int buf_print (buffer_struct *buffer, const char *s) { register int len; len = strlen (s); if (buffer->index + len >= buffer->size) return -1; sprintf (&(buffer->buf[buffer->index]), s); buffer->index += len; return len; }