aboutsummaryrefslogtreecommitdiff
path: root/src/hyper/hterror.c
diff options
context:
space:
mode:
Diffstat (limited to 'src/hyper/hterror.c')
-rw-r--r--src/hyper/hterror.c224
1 files changed, 224 insertions, 0 deletions
diff --git a/src/hyper/hterror.c b/src/hyper/hterror.c
new file mode 100644
index 00000000..8cd1e665
--- /dev/null
+++ b/src/hyper/hterror.c
@@ -0,0 +1,224 @@
+/*
+ Copyright (C) 1991-2002, The Numerical Algorithms Group Ltd.
+ All rights reserved.
+ Copyright (C) 2007-2008, Gabriel Dos Reis.
+ All right reserved.
+
+ Redistribution and use in source and binary forms, with or without
+ modification, are permitted provided that the following conditions are
+ met:
+
+ - Redistributions of source code must retain the above copyright
+ notice, this list of conditions and the following disclaimer.
+
+ - Redistributions in binary form must reproduce the above copyright
+ notice, this list of conditions and the following disclaimer in
+ the documentation and/or other materials provided with the
+ distribution.
+
+ - Neither the name of The Numerical Algorithms Group Ltd. nor the
+ names of its contributors may be used to endorse or promote products
+ derived from this software without specific prior written permission.
+
+ THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS
+ IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
+ TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
+ PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
+ OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
+ EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
+ PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
+ PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+ LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+ NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+*/
+
+#define _HTERROR_C
+#define HTERROR
+
+#include "axiom-c-macros.h"
+#include "useproto.h"
+#include "debug.h"
+
+#include "lex.h"
+#include "parse.h"
+
+#include "all_hyper_proto.H1"
+
+char ebuffer[128];
+jmp_buf jmpbuf;
+
+/*
+ * This file is the error handling routine in AXIOM. The main routine is
+ * called htperror(): arguments: msg - like perror it accepts an error
+ * message to be printed errno - the errno which occurred. This is so an
+ * appropriate error message can be printed.
+ *
+ * The prints out the page name, and then the filename in which the error
+ * occurred. If possible it also tries to print out the next ten tokens.
+ */
+
+void
+jump(void)
+{
+ if (gWindow == NULL)
+ exit(-1);
+ longjmp(jmpbuf, 1);
+ fprintf(stderr, "(HyperDoc) Long Jump failed, Exiting\n");
+ exit(-1);
+}
+
+void
+print_page_and_filename(void)
+{
+ char obuff[128];
+
+ if (gPageBeingParsed->type == Normal) {
+
+ /*
+ * Now try to inform the user as close to possible where the error
+ * occurred
+ */
+ sprintf(obuff, "(HyperDoc) While parsing %s on line %d\n\tin the file %s\n",
+ gPageBeingParsed->name, line_number,
+ gPageBeingParsed->filename);
+ }
+ else if (gPageBeingParsed->type == SpadGen) {
+ sprintf(obuff, "While parsing %s from the Spad socket\n",
+ gPageBeingParsed->name);
+ }
+ else if (gPageBeingParsed->type == Unixfd) {
+ sprintf(obuff, "While parsing %s from a Unixpipe\n",
+ gPageBeingParsed->name);
+ }
+ else {
+ /* Unknown page type */
+ sprintf(obuff, "While parsing %s\n", gPageBeingParsed->name);
+ }
+ fprintf(stderr, "%s", obuff);
+}
+
+
+void
+print_next_ten_tokens(void)
+{
+ int i;
+ int v;
+
+ fprintf(stderr, "Trying to print the next ten tokens\n");
+ for (i = 0; i < 10; i++) {
+ v = get_token();
+ if (v == EOF)
+ break;
+ print_token();
+ }
+ fprintf(stderr, "\n");
+}
+
+/* print out a token value */
+void
+print_token(void)
+{
+ if (token.type == Word)
+ printf("%s ", token.id);
+ else {
+ token_name(token.type);
+ printf("\\%s ", ebuffer);
+ }
+ fflush(stdout);
+}
+
+
+void
+token_name(int type)
+{
+ if (type <= NumberUserTokens)
+ strcpy(ebuffer, token_table[type]);
+ else {
+ switch (type) {
+ case Lbrace:
+ strcpy(ebuffer, "{");
+ break;
+ case Rbrace:
+ strcpy(ebuffer, "}");
+ break;
+ case Macro:
+ strcpy(ebuffer, token.id);
+ break;
+ case Group:
+ strcpy(ebuffer, "{");
+ break;
+ case Pound:
+ strcpy(ebuffer, "#");
+ break;
+ case Lsquarebrace:
+ strcpy(ebuffer, "[");
+ break;
+ case Rsquarebrace:
+ strcpy(ebuffer, "]");
+ break;
+ case Punctuation:
+ strcpy(ebuffer, token.id);
+ break;
+ case Dash:
+ strcpy(ebuffer, token.id);
+ break;
+ case Verbatim:
+ strcpy(ebuffer, "\\begin{verbatim}");
+ break;
+ case Scroll:
+ strcpy(ebuffer, "\\begin{scroll}");
+ break;
+ case Dollar:
+ strcpy(ebuffer, "$");
+ break;
+ case Percent:
+ strcpy(ebuffer, "%");
+ break;
+ case Carrot:
+ strcpy(ebuffer, "^");
+ break;
+ case Underscore:
+ strcpy(ebuffer, "_");
+ break;
+ case Tilde:
+ strcpy(ebuffer, "~");
+ break;
+ case Cond:
+ sprintf(ebuffer, "\\%s", token.id);
+ break;
+ case Icorrection:
+ strcpy(ebuffer, "\\/");
+ break;
+ case Paste:
+ strcpy(ebuffer, "\\begin{paste}");
+ break;
+ case Patch:
+ strcpy(ebuffer, "\\begin{patch}");
+ break;
+ default:
+ sprintf(ebuffer, " %d ", type);
+ }
+ /*return 1;*/
+ }
+}
+void
+htperror(char *msg, int errno)
+{
+ char obuff[256];
+
+ /* The first thing I do is create the error message */
+
+ if (errno <= Numerrors) {
+ sprintf(obuff, "%s:%s\n", msg, errmess[errno]);
+ }
+ else {
+ sprintf(obuff, "%s:\n", msg);
+ fprintf(stderr, "Unknown error type %d\n", errno);
+ }
+ fprintf(stderr, "%s", obuff);
+
+ print_page_and_filename();
+
+ print_next_ten_tokens();
+}