X-Git-Url: https://git.distorted.org.uk/~mdw/sgt/halibut/blobdiff_plain/f15300499bce37cd28ea2ace0f2bd1c364fc835e..8333e399e0540063d20dd6b344f2cd09458975ef:/main.c

diff --git a/main.c b/main.c
index f09b33e..e291d06 100644
--- a/main.c
+++ b/main.c
@@ -43,13 +43,23 @@ int main(int argc, char **argv) {
     int nogo;
     int errs;
     int reportcols;
+    int input_charset;
     int debug;
     int backendbits, prebackbits;
     int k, b;
     paragraph *cfg, *cfg_tail;
     void *pre_backend_data[16];
 
+    /*
+     * Use the specified locale everywhere. It'll be used for
+     * output of error messages, and as the default character set
+     * for input files if one is not explicitly specified.
+     * 
+     * However, we need to use standard numeric formatting for
+     * output of things like PDF.
+     */
     setlocale(LC_ALL, "");
+    setlocale(LC_NUMERIC, "C");
 
     /*
      * Set up initial (default) parameters.
@@ -58,6 +68,7 @@ int main(int argc, char **argv) {
     nfiles = 0;
     nogo = errs = FALSE;
     reportcols = 0;
+    input_charset = CS_ASCII;
     debug = 0;
     backendbits = 0;
     cfg = cfg_tail = NULL;
@@ -113,6 +124,17 @@ int main(int argc, char **argv) {
 			    }
 			if (k < (int)lenof(backends)) {
 			    /* do nothing */;
+			} else if (!strcmp(opt, "-input-charset")) {
+			    if (!val) {
+				errs = TRUE, error(err_optnoarg, opt);
+			    } else {
+				int charset = charset_from_localenc(val);
+				if (charset == CS_NONE) {
+				    errs = TRUE, error(err_cmdcharset, val);
+				} else {
+				    input_charset = charset;
+				}
+			    }
 			} else if (!strcmp(opt, "-help")) {
 			    help();
 			    nogo = TRUE;
@@ -123,6 +145,9 @@ int main(int argc, char **argv) {
 				   !strcmp(opt, "-license")) {
 			    licence();
 			    nogo = TRUE;
+			} else if (!strcmp(opt, "-list-charsets")) {
+			    listcharsets();
+			    nogo = TRUE;
 			} else if (!strcmp(opt, "-precise")) {
 			    reportcols = 1;
 			} else {
@@ -193,6 +218,10 @@ int main(int argc, char **argv) {
 			    while (*q) {
 				if (*q == ':') {
 				    *r = '\0';
+				    /* XXX ad-hoc diagnostic */
+				    if (!strcmp(s, "input-charset"))
+					error(err_futileopt, "Cinput-charset",
+					      "; use --input-charset");
 				    cmdline_cfg_add(para, s);
 				    r = s;
 				} else {
@@ -263,7 +292,7 @@ int main(int argc, char **argv) {
 	in.pushback = NULL;
 	in.reportcols = reportcols;
 	in.stack = NULL;
-	in.defcharset = charset_from_locale();
+	in.defcharset = input_charset;
 
 	idx = make_index();
 
@@ -430,6 +459,8 @@ static void dbg_prtwordlist(int level, word *w) {
 	    printf("\"");
 	} else
 	    printf("(no text)");
+	if (w->breaks)
+	    printf(" [breaks]");
 	if (w->alt) {
 	    printf(" alt = {\n");
 	    dbg_prtwordlist(level+1, w->alt);