* xgettext.c (extract_ty): New field 'literalstring_parser'.
(current_literalstring_parser): New variable.
(extract_from_file): Set 'current_literalstring_parser'.
(savable_comment_convert_encoding): New function.
(arglist_parser_remember_literal): New function.
(arglist_parser_remember): Call 'arglist_parser_remember_literal'.
(arglist_parser_done): Call literalstring_parser on msgctxt,
msgid, and msgid_plural before calling 'remember_a_message';
convert encoding of msgid_comment.
* x-*.h: Register 'literalstring_parser' through SCANNER_*.
* xgettext.h (enum literalstring_escape_type): New enum.
(struct literalstring_parser): New struct.
(struct partial_call): New fields 'msgctxt_escape',
'msgid_escape', and 'msgid_plural_escape'.
(arglist_parser_remember_literal): New function declaration.
(savable_comment_convert_encoding): New function declaration.
+2014-05-09 Daiki Ueno <ueno@gnu.org>
+
+ xgettext: Provide a way to interpret string literals lazily
+ * xgettext.c (extract_ty): New field 'literalstring_parser'.
+ (current_literalstring_parser): New variable.
+ (extract_from_file): Set 'current_literalstring_parser'.
+ (savable_comment_convert_encoding): New function.
+ (arglist_parser_remember_literal): New function.
+ (arglist_parser_remember): Call 'arglist_parser_remember_literal'.
+ (arglist_parser_done): Call literalstring_parser on msgctxt,
+ msgid, and msgid_plural before calling 'remember_a_message';
+ convert encoding of msgid_comment.
+ * x-*.h: Register 'literalstring_parser' through SCANNER_*.
+ * xgettext.h (enum literalstring_escape_type): New enum.
+ (struct literalstring_parser): New struct.
+ (struct partial_call): New fields 'msgctxt_escape',
+ 'msgid_escape', and 'msgid_plural_escape'.
+ (arglist_parser_remember_literal): New function declaration.
+ (savable_comment_convert_encoding): New function declaration.
+
2014-05-03 Daiki Ueno <ueno@gnu.org>
xgettext: Recognize prefixed comment tag
#define SCANNERS_AWK \
{ "awk", extract_awk, \
- &flag_table_awk, &formatstring_awk, NULL }, \
+ &flag_table_awk, &formatstring_awk, NULL, NULL }, \
/* Scan an awk file and add its translatable strings to mdlp. */
extern void extract_awk (FILE *fp, const char *real_filename,
#define SCANNERS_C \
{ "C", extract_c, \
&flag_table_c, \
- &formatstring_c, NULL }, \
+ &formatstring_c, NULL, NULL }, \
{ "C++", extract_c, \
&flag_table_c, \
- &formatstring_c, NULL }, \
+ &formatstring_c, NULL, NULL }, \
{ "ObjectiveC", extract_objc, \
&flag_table_objc, \
- &formatstring_c, &formatstring_objc }, \
+ &formatstring_c, &formatstring_objc, NULL }, \
{ "GCC-source", extract_c, \
&flag_table_gcc_internal, \
- &formatstring_gcc_internal, &formatstring_gfc_internal }, \
+ &formatstring_gcc_internal, &formatstring_gfc_internal, NULL }, \
/* Scan a C/C++ file and add its translatable strings to mdlp. */
extern void extract_c (FILE *fp, const char *real_filename,
#define SCANNERS_CSHARP \
{ "C#", extract_csharp, \
- &flag_table_csharp, &formatstring_csharp, NULL }, \
+ &flag_table_csharp, &formatstring_csharp, NULL, NULL }, \
extern void extract_csharp (FILE *fp, const char *real_filename,
const char *logical_filename,
{ "desktop", "Desktop" }, \
#define SCANNERS_DESKTOP \
- { "Desktop", extract_desktop, NULL, NULL, NULL }, \
+ { "Desktop", extract_desktop, NULL, NULL, NULL, NULL }, \
/* Scan a Desktop Entry file and add its translatable strings to mdlp. */
extern void extract_desktop (FILE *fp, const char *real_filename,
#define SCANNERS_ELISP \
{ "EmacsLisp", extract_elisp, \
- &flag_table_elisp, &formatstring_elisp, NULL }, \
+ &flag_table_elisp, &formatstring_elisp, NULL, NULL }, \
/* Scan an Emacs Lisp file and add its translatable strings to mdlp. */
extern void extract_elisp (FILE *fp, const char *real_filename,
{ "ui", "glade" }, \
#define SCANNERS_GLADE \
- { "glade", extract_glade, NULL, NULL, NULL }, \
+ { "glade", extract_glade, NULL, NULL, NULL, NULL }, \
/* Scan a glade XML file and add its translatable strings to mdlp. */
extern void extract_glade (FILE *fp, const char *real_filename,
{ "gschema.xml", "gsettings" }, \
#define SCANNERS_GSETTINGS \
- { "gsettings", extract_gsettings, NULL, NULL, NULL }, \
+ { "gsettings", extract_gsettings, NULL, NULL, NULL, NULL }, \
/* Scan a gsettings XML file and add its translatable strings to mdlp. */
extern void extract_gsettings (FILE *fp, const char *real_filename,
#define SCANNERS_JAVA \
{ "Java", extract_java, \
- &flag_table_java, &formatstring_java, NULL }, \
+ &flag_table_java, &formatstring_java, NULL, NULL }, \
extern void extract_java (FILE *fp, const char *real_filename,
const char *logical_filename,
#define SCANNERS_JAVASCRIPT \
{ "JavaScript", extract_javascript, \
- &flag_table_javascript, &formatstring_javascript, NULL }, \
+ &flag_table_javascript, &formatstring_javascript, NULL, NULL }, \
/* Scan a Python file and add its translatable strings to mdlp. */
extern void extract_javascript (FILE *fp, const char *real_filename,
#define SCANNERS_LIBREP \
{ "librep", extract_librep, \
- &flag_table_librep, &formatstring_librep, NULL }, \
+ &flag_table_librep, &formatstring_librep, NULL, NULL }, \
/* Scan a librep file and add its translatable strings to mdlp. */
extern void extract_librep (FILE *fp, const char *real_filename,
#define SCANNERS_LISP \
{ "Lisp", extract_lisp, \
- &flag_table_lisp, &formatstring_lisp, NULL }, \
+ &flag_table_lisp, &formatstring_lisp, NULL, NULL }, \
/* Scan a Lisp file and add its translatable strings to mdlp. */
extern void extract_lisp (FILE *fp, const char *real_filename,
#define SCANNERS_LUA \
{ "Lua", extract_lua, \
- &flag_table_lua, &formatstring_lua, NULL }, \
+ &flag_table_lua, &formatstring_lua, NULL, NULL }, \
/* Scan a Lua file and add its translatable strings to mdlp. */
extern void extract_lua (FILE * fp, const char *real_filename,
#define SCANNERS_PERL \
{ "perl", extract_perl, \
- &flag_table_perl, &formatstring_perl, &formatstring_perl_brace }, \
+ &flag_table_perl, &formatstring_perl, &formatstring_perl_brace, NULL }, \
/* Scan a Perl file and add its translatable strings to mdlp. */
extern void extract_perl (FILE *fp, const char *real_filename,
#define SCANNERS_PHP \
{ "PHP", extract_php, \
- &flag_table_php, &formatstring_php, NULL }, \
+ &flag_table_php, &formatstring_php, NULL, NULL }, \
/* Scan a PHP file and add its translatable strings to mdlp. */
extern void extract_php (FILE *fp, const char *real_filename,
{ "pot", "PO" }, \
#define SCANNERS_PO \
- { "PO", extract_po, NULL, NULL, NULL }, \
+ { "PO", extract_po, NULL, NULL, NULL, NULL }, \
/* Scan a PO file and add its translatable strings to mdlp. */
extern void extract_po (FILE *fp, const char *real_filename,
{ "properties", "JavaProperties" }, \
#define SCANNERS_PROPERTIES \
- { "JavaProperties", extract_properties, NULL, NULL, NULL }, \
+ { "JavaProperties", extract_properties, NULL, NULL, NULL, NULL }, \
/* Scan a JavaProperties file and add its translatable strings to mdlp. */
extern void extract_properties (FILE *fp, const char *real_filename,
#define SCANNERS_PYTHON \
{ "Python", extract_python, \
- &flag_table_python, &formatstring_python, &formatstring_python_brace }, \
+ &flag_table_python, &formatstring_python, &formatstring_python_brace, NULL }, \
/* Scan a Python file and add its translatable strings to mdlp. */
extern void extract_python (FILE *fp, const char *real_filename,
#define SCANNERS_RST \
{ "RST", extract_rst, \
- NULL, &formatstring_pascal, NULL }, \
+ NULL, &formatstring_pascal, NULL, NULL }, \
/* Scan an RST file and add its translatable strings to mdlp. */
extern void extract_rst (FILE *fp, const char *real_filename,
#define SCANNERS_SCHEME \
{ "Scheme", extract_scheme, \
- &flag_table_scheme, &formatstring_scheme, NULL }, \
+ &flag_table_scheme, &formatstring_scheme, NULL, NULL }, \
/* Scan a Scheme file and add its translatable strings to mdlp. */
extern void extract_scheme (FILE *fp, const char *real_filename,
#define SCANNERS_SH \
{ "Shell", extract_sh, \
- &flag_table_sh, &formatstring_sh, NULL }, \
+ &flag_table_sh, &formatstring_sh, NULL, NULL }, \
/* Scan a shell script file and add its translatable strings to mdlp. */
extern void extract_sh (FILE *fp, const char *real_filename,
#define SCANNERS_SMALLTALK \
{ "Smalltalk", extract_smalltalk, \
- NULL, &formatstring_smalltalk, NULL }, \
+ NULL, &formatstring_smalltalk, NULL, NULL }, \
/* Scan a Smalltalk file and add its translatable strings to mdlp. */
extern void extract_smalltalk (FILE *fp, const char *real_filename,
{ "strings", "NXStringTable" }, \
#define SCANNERS_STRINGTABLE \
- { "NXStringTable", extract_stringtable, NULL, NULL, NULL }, \
+ { "NXStringTable", extract_stringtable, NULL, NULL, NULL, NULL }, \
/* Scan a JavaProperties file and add its translatable strings to mdlp. */
extern void extract_stringtable (FILE *fp, const char *real_filename,
#define SCANNERS_TCL \
{ "Tcl", extract_tcl, \
- &flag_table_tcl, &formatstring_tcl, NULL }, \
+ &flag_table_tcl, &formatstring_tcl, NULL, NULL }, \
/* Scan a Tcl file and add its translatable strings to mdlp. */
extern void extract_tcl (FILE *fp, const char *real_filename,
#define SCANNERS_VALA \
{ "Vala", extract_vala, \
- &flag_table_vala, &formatstring_c, NULL }, \
+ &flag_table_vala, &formatstring_c, NULL, NULL }, \
/* Scan a Vala file and add its translatable strings to mdlp. */
extern void extract_vala (FILE *fp, const char *real_filename,
#define SCANNERS_YCP \
{ "YCP", extract_ycp, \
- &flag_table_ycp, &formatstring_ycp, NULL }, \
+ &flag_table_ycp, &formatstring_ycp, NULL, NULL }, \
/* Scan an YCP file and add its translatable strings to mdlp. */
extern void extract_ycp (FILE *fp, const char *real_filename,
struct formatstring_parser *formatstring_parser1;
struct formatstring_parser *formatstring_parser2;
struct formatstring_parser *formatstring_parser3;
+ struct literalstring_parser *literalstring_parser;
};
}
}
+refcounted_string_list_ty *
+savable_comment_convert_encoding (refcounted_string_list_ty *comment,
+ lex_pos_ty *pos)
+{
+ refcounted_string_list_ty *result;
+ size_t i;
+
+ result = XMALLOC (refcounted_string_list_ty);
+ result->refcount = 1;
+ string_list_init (&result->contents);
+
+ for (i = 0; i < comment->contents.nitems; i++)
+ {
+ const char *old_string = comment->contents.item[i];
+ char *string = from_current_source_encoding (old_string,
+ lc_comment,
+ pos->file_name,
+ pos->line_number);
+ string_list_append (&result->contents, string);
+ if (string != old_string)
+ free (string);
+ }
+
+ return result;
+}
+
static FILE *
static struct formatstring_parser *current_formatstring_parser2;
static struct formatstring_parser *current_formatstring_parser3;
+static struct literalstring_parser *current_literalstring_parser;
static void
extract_from_file (const char *file_name, extractor_ty extractor,
current_formatstring_parser1 = extractor.formatstring_parser1;
current_formatstring_parser2 = extractor.formatstring_parser2;
current_formatstring_parser3 = extractor.formatstring_parser3;
+ current_literalstring_parser = extractor.literalstring_parser;
extractor.func (fp, real_file_name, logical_file_name, extractor.flag_table,
mdlp);
ap->alternative[i].argtotal = shapes->shapes[i].argtotal;
ap->alternative[i].xcomments = shapes->shapes[i].xcomments;
ap->alternative[i].msgctxt = NULL;
+ ap->alternative[i].msgctxt_escape = 0;
ap->alternative[i].msgctxt_pos.file_name = NULL;
ap->alternative[i].msgctxt_pos.line_number = (size_t)(-1);
ap->alternative[i].msgid = NULL;
+ ap->alternative[i].msgid_escape = 0;
ap->alternative[i].msgid_context = null_context;
ap->alternative[i].msgid_pos.file_name = NULL;
ap->alternative[i].msgid_pos.line_number = (size_t)(-1);
ap->alternative[i].msgid_comment = NULL;
ap->alternative[i].msgid_plural = NULL;
+ ap->alternative[i].msgid_plural_escape = 0;
ap->alternative[i].msgid_plural_context = null_context;
ap->alternative[i].msgid_plural_pos.file_name = NULL;
ap->alternative[i].msgid_plural_pos.line_number = (size_t)(-1);
ccp->argtotal = cp->argtotal;
ccp->xcomments = cp->xcomments;
ccp->msgctxt = (cp->msgctxt != NULL ? xstrdup (cp->msgctxt) : NULL);
+ ccp->msgctxt_escape = cp->msgctxt_escape;
ccp->msgctxt_pos = cp->msgctxt_pos;
ccp->msgid = (cp->msgid != NULL ? xstrdup (cp->msgid) : NULL);
+ ccp->msgid_escape = cp->msgid_escape;
ccp->msgid_context = cp->msgid_context;
ccp->msgid_pos = cp->msgctxt_pos;
ccp->msgid_comment = add_reference (cp->msgid_comment);
ccp->msgid_plural =
(cp->msgid_plural != NULL ? xstrdup (cp->msgid_plural) : NULL);
+ ccp->msgid_plural_escape = cp->msgid_plural_escape;
ccp->msgid_plural_context = cp->msgid_plural_context;
ccp->msgid_plural_pos = cp->msgid_plural_pos;
}
void
-arglist_parser_remember (struct arglist_parser *ap,
- int argnum, char *string,
- flag_context_ty context,
- char *file_name, size_t line_number,
- refcounted_string_list_ty *comment)
+arglist_parser_remember_literal (struct arglist_parser *ap,
+ int argnum, char *string,
+ flag_context_ty context,
+ char *file_name, size_t line_number,
+ refcounted_string_list_ty *comment,
+ enum literalstring_escape_type type)
{
bool stored_string = false;
size_t nalternatives = ap->nalternatives;
if (argnum == cp->argnumc)
{
cp->msgctxt = string;
+ cp->msgctxt_escape = type;
cp->msgctxt_pos.file_name = file_name;
cp->msgctxt_pos.line_number = line_number;
stored_string = true;
if (argnum == cp->argnum1)
{
cp->msgid = string;
+ cp->msgid_escape = type;
cp->msgid_context = context;
cp->msgid_pos.file_name = file_name;
cp->msgid_pos.line_number = line_number;
if (argnum == cp->argnum2)
{
cp->msgid_plural = string;
+ cp->msgid_plural_escape = type;
cp->msgid_plural_context = context;
cp->msgid_plural_pos.file_name = file_name;
cp->msgid_plural_pos.line_number = line_number;
free (string);
}
+void
+arglist_parser_remember (struct arglist_parser *ap,
+ int argnum, char *string,
+ flag_context_ty context,
+ char *file_name, size_t line_number,
+ refcounted_string_list_ty *comment)
+{
+ arglist_parser_remember_literal (ap, argnum, string, context,
+ file_name, line_number,
+ comment, 0);
+}
bool
arglist_parser_decidedp (struct arglist_parser *ap, int argnum)
{
flag_context_ty msgid_context = best_cp->msgid_context;
flag_context_ty msgid_plural_context = best_cp->msgid_plural_context;
+ struct literalstring_parser *parser = current_literalstring_parser;
+ const char *encoding;
/* Special support for the 3-argument tr operator in Qt:
When --qt and --keyword=tr:1,1,2c,3t are specified, add to the
msgid_plural_context.is_format3 = yes_according_to_context;
}
+ if (best_cp->msgctxt != NULL)
+ {
+ if (parser != NULL && best_cp->msgctxt_escape != 0)
+ {
+ char *msgctxt =
+ parser->parse (best_cp->msgctxt,
+ &best_cp->msgctxt_pos,
+ best_cp->msgctxt_escape);
+ free (best_cp->msgctxt);
+ best_cp->msgctxt = msgctxt;
+ }
+ else
+ {
+ lex_pos_ty *pos = &best_cp->msgctxt_pos;
+ CONVERT_STRING (best_cp->msgctxt, lc_string);
+ }
+ }
+
+ if (parser != NULL && best_cp->msgid_escape != 0)
+ {
+ char *msgid = parser->parse (best_cp->msgid,
+ &best_cp->msgid_pos,
+ best_cp->msgid_escape);
+ free (best_cp->msgid);
+ if (best_cp->msgid_plural == best_cp->msgid)
+ best_cp->msgid_plural = msgid;
+ best_cp->msgid = msgid;
+ }
+ else
+ {
+ lex_pos_ty *pos = &best_cp->msgid_pos;
+ CONVERT_STRING (best_cp->msgid, lc_string);
+ }
+
+ if (best_cp->msgid_comment != NULL)
+ {
+ refcounted_string_list_ty *msgid_comment =
+ savable_comment_convert_encoding (best_cp->msgid_comment,
+ &best_cp->msgid_pos);
+ drop_reference (best_cp->msgid_comment);
+ best_cp->msgid_comment = msgid_comment;
+ }
+
+ /* best_cp->msgctxt and best_cp->msgid are already in
+ UTF-8. Prevent further conversion in remember_a_message. */
+ encoding = xgettext_current_source_encoding;
+ xgettext_current_source_encoding = po_charset_utf8;
mp = remember_a_message (ap->mlp, best_cp->msgctxt, best_cp->msgid,
msgid_context,
&best_cp->msgid_pos,
NULL, best_cp->msgid_comment);
+ xgettext_current_source_encoding = encoding;
+
if (mp != NULL && best_cp->msgid_plural != NULL)
- remember_a_message_plural (mp, best_cp->msgid_plural,
- msgid_plural_context,
+ {
+ /* best_cp->msgid_plural may point to best_cp->msgid.
+ In that case, it is already interpreted and converted. */
+ if (best_cp->msgid_plural != best_cp->msgid)
+ {
+ if (parser != NULL
+ && best_cp->msgid_plural_escape != 0)
+ {
+ char *msgid_plural =
+ parser->parse (best_cp->msgid_plural,
&best_cp->msgid_plural_pos,
- NULL);
+ best_cp->msgid_plural_escape);
+ free (best_cp->msgid_plural);
+ best_cp->msgid_plural = msgid_plural;
+ }
+ else
+ {
+ lex_pos_ty *pos = &best_cp->msgid_plural_pos;
+ CONVERT_STRING (best_cp->msgid_plural, lc_string);
+ }
+ }
+
+ encoding = xgettext_current_source_encoding;
+ xgettext_current_source_encoding = po_charset_utf8;
+ remember_a_message_plural (mp, best_cp->msgid_plural,
+ msgid_plural_context,
+ &best_cp->msgid_plural_pos,
+ NULL);
+ xgettext_current_source_encoding = encoding;
+ }
}
if (best_cp->xcomments.nitems > 0)
flag_context_list_table_ty *flag_table;
struct formatstring_parser *formatstring_parser1;
struct formatstring_parser *formatstring_parser2;
+ struct literalstring_parser *literalstring_parser;
};
typedef struct table_ty table_ty;
result.formatstring_parser1 = tp->formatstring_parser1;
result.formatstring_parser2 = tp->formatstring_parser2;
result.formatstring_parser3 = NULL;
+ result.literalstring_parser = tp->literalstring_parser;
/* Handle --qt. It's preferrable to handle this facility here rather
than through an option --language=C++/Qt because the latter would
extern void savable_comment_add (const char *str);
extern void savable_comment_reset (void);
+/* Convert character encoding of COMMENT according to the current
+ source encoding. Returns a new refcounted_string_list_ty. */
+extern refcounted_string_list_ty *
+ savable_comment_convert_encoding (refcounted_string_list_ty *comment,
+ lex_pos_ty *pos);
+
+
+enum literalstring_escape_type
+{
+ LET_ANSI_C = 1 << 0,
+ LET_UNICODE = 1 << 1
+};
+
+struct literalstring_parser
+{
+ char * (*parse) (const char *string, lex_pos_ty *pos,
+ enum literalstring_escape_type type);
+};
/* Add a message to the list of extracted messages.
msgctxt must be either NULL or a malloc()ed string; its ownership is passed
lex_pos_ty *pos,
refcounted_string_list_ty *comment);
-
/* Represents the progressive parsing of an argument list w.r.t. a single
'struct callshape'. */
struct partial_call
int argtotal; /* total number of arguments, 0 if unspecified */
string_list_ty xcomments; /* auto-extracted comments */
char *msgctxt; /* context - owned string, or NULL */
+ enum literalstring_escape_type msgctxt_escape;
lex_pos_ty msgctxt_pos;
char *msgid; /* msgid - owned string, or NULL */
+ enum literalstring_escape_type msgid_escape;
flag_context_ty msgid_context;
lex_pos_ty msgid_pos;
refcounted_string_list_ty *msgid_comment;
char *msgid_plural; /* msgid_plural - owned string, or NULL */
+ enum literalstring_escape_type msgid_plural_escape;
flag_context_ty msgid_plural_context;
lex_pos_ty msgid_plural_pos;
};
flag_context_ty context,
char *file_name, size_t line_number,
refcounted_string_list_ty *comment);
+/* Adds an uninterpreted string argument to an arglist_parser. ARGNUM
+ must be > 0.
+ STRING is must be malloc()ed string; its ownership is passed to the callee.
+ FILE_NAME must be allocated with indefinite extent.
+ COMMENT may be savable_comment, or it may be a saved copy of savable_comment
+ (then add_reference must be used when saving it, and drop_reference while
+ dropping it). Clear savable_comment. */
+extern void arglist_parser_remember_literal (struct arglist_parser *ap,
+ int argnum, char *string,
+ flag_context_ty context,
+ char *file_name, size_t line_number,
+ refcounted_string_list_ty *comment,
+ enum literalstring_escape_type type);
/* Tests whether an arglist_parser has is not waiting for more arguments after
argument ARGNUM. */
extern bool arglist_parser_decidedp (struct arglist_parser *ap, int argnum);
projects / thirdparty / gettext.git / commitdiff
