houdini_escape_xml tests

main.c

#include <houdini.h>
#include <buffer.h>
#include <string.h>
#include <stdio.h>

static unsigned int failures = 0;

static void test_input(const char *input, const char *expected_output)
{
  struct buf *encoded = bufnew(128);

  houdini_escape_xml(encoded, (uint8_t *)input, strlen(input));
  if (strcmp(expected_output, (char *)encoded->data)) {
    failures++;
    fprintf(stderr, "encoded: input '%s' expected '%s' got '%s'\n",
            input, expected_output, encoded->data);
  }
}

int main(int argc, char **argv)
{
  static const struct {
    const char *input;
    const char *expected_output;
  } tests[] = {
    { "\x1d", "?" },
    { "\t\r\n ", "\t\r\n " },
    { "\"&'/<=>", "&quot;&amp;&apos;/&lt;=&gt;" },
    { "~\x7f\xc2\x80", "~\x7f\xc2\x80" },
    { "~\x7f\xc2\x80", "~\x7f\xc2\x80" },
    { "\x80\xbf", "??" }, /* start with intermediate bytes 10xxxxxx */
    { "\xc0\xa0\xc0\xbc\xc0\xbe", "???" }, /* non-short form */
    { "\xc2\xa2", "\xc2\xa2" }, /* cent */
    { "\xe2\x82\xac", "\xe2\x82\xac" }, /* euro */
    { "\xf0\xa4\xad\xa2", "\xf0\xa4\xad\xa2" }, /* 4 bytes char */
    { "\xed\x9f\xbf", "\xed\x9f\xbf" }, /* 0xd7ff */
    { "\xed\xa0\x80", "?" }, /* 0xd800 */
    { "\xed\xbf\xbf", "?" }, /* 0xdfff */
    { "\xee\x80\x80", "\xee\x80\x80" }, /* 0xe000 */
    { "\xef\xbf\xbd", "\xef\xbf\xbd" }, /* 0xfffd */
    { "\xef\xbf\xbe", "?" }, /* 0xfffe */

    { "\xf0\x90\x80\x80", "\xf0\x90\x80\x80" }, /* 0x10000 */
    { "\xf4\x8f\xbf\xbf", "\xf4\x8f\xbf\xbf" }, /* 0x10ffff */
    { "\xf4\x90\x80\x80", "?" }, /* 0x110000 */
    { "\xf8\xfc\xff", "???" }, /* invalid UTF-8 bytes 11111xxx */
  };
  int i;

  for (i = 0; i < sizeof(tests)/sizeof(tests[0]); i++)
    test_input(tests[i].input, tests[i].expected_output);
  return failures < 255 ? failures : 255;
}

Makefile

MODULE = main
SOURCES = $(wildcard *.c)
OBJS = $(SOURCES:%.c=%.o)
CFLAGS = -Wall -I. -ggdb -O0

all: $(MODULE)

$(MODULE): $(OBJS)

clean:
	rm -f $(OBJS) $(MODULE)

.PHONY:
	all clean