X-Git-Url: https://www.chiark.greenend.org.uk/ucgi/~mdw/git/disorder/blobdiff_plain/e5a5a1388e0236c4aa4084493e2f78ff3fbc8c5b..6049fe0e789884ac0589190495d68e470c0fe19c:/lib/test.c diff --git a/lib/test.c b/lib/test.c index f5d94fd..02deb02 100644 --- a/lib/test.c +++ b/lib/test.c @@ -30,6 +30,9 @@ #include #include #include +#include +#include +#include #include "utf8.h" #include "mem.h" @@ -38,17 +41,32 @@ #include "charset.h" #include "mime.h" #include "hex.h" -#include "words.h" #include "heap.h" #include "unicode.h" #include "inputline.h" +#include "wstat.h" +#include "signame.h" +#include "cache.h" +#include "filepart.h" +#include "hash.h" +#include "selection.h" +#include "syscalls.h" +#include "kvp.h" +#include "sink.h" static int tests, errors; +static int fail_first; + +static void count_error() { + ++errors; + if(fail_first) + abort(); +} /** @brief Checks that @p expr is nonzero */ #define insist(expr) do { \ if(!(expr)) { \ - ++errors; \ + count_error(); \ fprintf(stderr, "%s:%d: error checking %s\n", \ __FILE__, __LINE__, #expr); \ } \ @@ -80,31 +98,43 @@ static const char *format_utf32(const uint32_t *s) { dynstr_init(&d); while((c = *s++)) { - if(c >= 32 && c <= 127) - dynstr_append(&d, c); - else { - sprintf(buf, "\\x%04lX", (unsigned long)c); - dynstr_append_string(&d, buf); - } + sprintf(buf, " %04lX", (long)c); + dynstr_append_string(&d, buf); } dynstr_terminate(&d); return d.vec; } -#define check_string(GOT, WANT) do { \ - const char *g = GOT; \ - const char *w = WANT; \ - \ - if(w == 0) { \ - fprintf(stderr, "%s:%d: %s returned 0\n", \ - __FILE__, __LINE__, #GOT); \ - ++errors; \ - } else if(strcmp(w, g)) { \ - fprintf(stderr, "%s:%d: %s returned:\n%s\nexpected:\n%s\n", \ - __FILE__, __LINE__, #GOT, format(g), format(w)); \ - ++errors; \ - } \ - ++tests; \ +#define check_string(GOT, WANT) do { \ + const char *got = GOT; \ + const char *want = WANT; \ + \ + if(want == 0) { \ + fprintf(stderr, "%s:%d: %s returned 0\n", \ + __FILE__, __LINE__, #GOT); \ + count_error(); \ + } else if(strcmp(want, got)) { \ + fprintf(stderr, "%s:%d: %s returned:\n%s\nexpected:\n%s\n", \ + __FILE__, __LINE__, #GOT, format(got), format(want)); \ + count_error(); \ + } \ + ++tests; \ + } while(0) + +#define check_string_prefix(GOT, WANT) do { \ + const char *got = GOT; \ + const char *want = WANT; \ + \ + if(want == 0) { \ + fprintf(stderr, "%s:%d: %s returned 0\n", \ + __FILE__, __LINE__, #GOT); \ + count_error(); \ + } else if(strncmp(want, got, strlen(want))) { \ + fprintf(stderr, "%s:%d: %s returned:\n%s\nexpected:\n%s...\n", \ + __FILE__, __LINE__, #GOT, format(got), format(want)); \ + count_error(); \ + } \ + ++tests; \ } while(0) static uint32_t *ucs4parse(const char *s) { @@ -131,15 +161,16 @@ static void test_utf8(void) { char *u8; \ \ insist(validutf8(CHARS)); \ - ucs = utf82ucs4(CHARS); \ + ucs = utf8_to_utf32(CHARS, strlen(CHARS), 0); \ insist(ucs != 0); \ - insist(!ucs4cmp(w, ucs)); \ - u8 = ucs42utf8(ucs); \ + insist(!utf32_cmp(w, ucs)); \ + u8 = utf32_to_utf8(ucs, utf32_len(ucs), 0); \ insist(u8 != 0); \ insist(!strcmp(u8, CHARS)); \ } while(0) fprintf(stderr, "test_utf8\n"); +#define validutf8(S) utf8_valid((S), strlen(S)) /* empty string */ @@ -210,6 +241,7 @@ static void test_utf8(void) { U8("\xF4\x80\x80\x80", "0x100000"); U8("\xF4\x8F\xBF\xBF", "0x10FFFF"); insist(!validutf8("\xF4\x90\x80\x80")); + insist(!validutf8("\xF4\x80\xFF\x80")); /* miscellaneous non-UTF-8 rubbish */ insist(!validutf8("\x80")); @@ -355,13 +387,17 @@ static void test_hex(void) { static void test_casefold(void) { uint32_t c, l; - const char *input, *folded, *expected; + const char *input, *canon_folded, *compat_folded, *canon_expected, *compat_expected; fprintf(stderr, "test_casefold\n"); + + /* This isn't a very exhaustive test. Unlike for normalization, there don't + * seem to be any public test vectors for these algorithms. */ for(c = 1; c < 256; ++c) { input = utf32_to_utf8(&c, 1, 0); - folded = utf8_casefold_canon(input, strlen(input), 0); + canon_folded = utf8_casefold_canon(input, strlen(input), 0); + compat_folded = utf8_casefold_compat(input, strlen(input), 0); switch(c) { default: if((c >= 'A' && c <= 'Z') @@ -374,23 +410,98 @@ static void test_casefold(void) { l = 0x3BC; /* GREEK SMALL LETTER MU */ break; case 0xDF: /* LATIN SMALL LETTER SHARP S */ - insist(!strcmp(folded, "ss")); + insist(!strcmp(canon_folded, "ss")); + insist(!strcmp(compat_folded, "ss")); l = 0; break; } if(l) { + uint32_t *d; /* Case-folded data is now normalized */ - expected = ucs42utf8(utf32_decompose_canon(&l, 1, 0)); - if(strcmp(folded, expected)) { - fprintf(stderr, "%s:%d: casefolding %#lx got '%s', expected '%s'\n", + d = utf32_decompose_canon(&l, 1, 0); + canon_expected = utf32_to_utf8(d, utf32_len(d), 0); + if(strcmp(canon_folded, canon_expected)) { + fprintf(stderr, "%s:%d: canon-casefolding %#lx got '%s', expected '%s'\n", __FILE__, __LINE__, (unsigned long)c, - format(folded), format(expected)); - ++errors; + format(canon_folded), format(canon_expected)); + count_error(); + } + ++tests; + d = utf32_decompose_compat(&l, 1, 0); + compat_expected = utf32_to_utf8(d, utf32_len(d), 0); + if(strcmp(compat_folded, compat_expected)) { + fprintf(stderr, "%s:%d: compat-casefolding %#lx got '%s', expected '%s'\n", + __FILE__, __LINE__, (unsigned long)c, + format(compat_folded), format(compat_expected)); + count_error(); } ++tests; } } - check_string(casefold(""), ""); + check_string(utf8_casefold_canon("", 0, 0), ""); +} + +struct { + const char *in; + const char *expect[10]; +} wtest[] = { + /* Empty string */ + { "", { 0 } }, + /* Only whitespace and punctuation */ + { " ", { 0 } }, + { " ' ", { 0 } }, + { " ! ", { 0 } }, + { " \"\" ", { 0 } }, + { " @ ", { 0 } }, + /* Basics */ + { "wibble", { "wibble", 0 } }, + { " wibble", { "wibble", 0 } }, + { " wibble ", { "wibble", 0 } }, + { "wibble ", { "wibble", 0 } }, + { "wibble spong", { "wibble", "spong", 0 } }, + { " wibble spong", { "wibble", "spong", 0 } }, + { " wibble spong ", { "wibble", "spong", 0 } }, + { "wibble spong ", { "wibble", "spong", 0 } }, + { "wibble spong splat foo zot ", { "wibble", "spong", "splat", "foo", "zot", 0 } }, + /* Apostrophes */ + { "wibble 'spong", { "wibble", "spong", 0 } }, + { " wibble's", { "wibble's", 0 } }, + { " wibblespong' ", { "wibblespong", 0 } }, + { "wibble sp''ong ", { "wibble", "sp", "ong", 0 } }, +}; +#define NWTEST (sizeof wtest / sizeof *wtest) + +static void test_words(void) { + size_t t, nexpect, ngot, i; + int right; + + fprintf(stderr, "test_words\n"); + for(t = 0; t < NWTEST; ++t) { + char **got = utf8_word_split(wtest[t].in, strlen(wtest[t].in), &ngot, 0); + + for(nexpect = 0; wtest[t].expect[nexpect]; ++nexpect) + ; + if(nexpect == ngot) { + for(i = 0; i < ngot; ++i) + if(strcmp(wtest[t].expect[i], got[i])) + break; + right = i == ngot; + } else + right = 0; + if(!right) { + fprintf(stderr, "word split %zu failed\n", t); + fprintf(stderr, "input: %s\n", wtest[t].in); + fprintf(stderr, " | %-30s | %-30s\n", + "expected", "got"); + for(i = 0; i < nexpect || i < ngot; ++i) { + const char *e = i < nexpect ? wtest[t].expect[i] : ""; + const char *g = i < ngot ? got[i] : ""; + fprintf(stderr, " %2zu | %-30s | %-30s\n", i, e, g); + } + count_error(); + } + ++tests; + } } /** @brief Less-than comparison function for integer heap */ @@ -422,26 +533,97 @@ static void test_heap(void) { putchar('\n'); } +/** @brief Open a Unicode test file */ +static FILE *open_unicode_test(const char *path) { + const char *base; + FILE *fp; + char buffer[1024]; + int w; + + if((base = strrchr(path, '/'))) + ++base; + else + base = path; + if(!(fp = fopen(base, "r"))) { + snprintf(buffer, sizeof buffer, + "wget http://www.unicode.org/Public/5.0.0/ucd/%s", path); + if((w = system(buffer))) + fatal(0, "%s: %s", buffer, wstat(w)); + if(chmod(base, 0444) < 0) + fatal(errno, "chmod %s", base); + if(!(fp = fopen(base, "r"))) + fatal(errno, "%s", base); + } + return fp; +} + +/** @brief Run breaking tests for utf32_grapheme_boundary() etc */ +static void breaktest(const char *path, + int (*breakfn)(const uint32_t *, size_t, size_t)) { + FILE *fp = open_unicode_test(path); + int lineno = 0; + char *l, *lp; + size_t bn, n; + char break_allowed[1024]; + uint32_t buffer[1024]; + + while(!inputline(path, fp, &l, '\n')) { + ++lineno; + if(l[0] == '#') continue; + bn = 0; + lp = l; + while(*lp) { + if(*lp == ' ' || *lp == '\t') { + ++lp; + continue; + } + if(*lp == '#') + break; + if((unsigned char)*lp == 0xC3 && (unsigned char)lp[1] == 0xB7) { + /* 00F7 DIVISION SIGN */ + break_allowed[bn] = 1; + lp += 2; + continue; + } + if((unsigned char)*lp == 0xC3 && (unsigned char)lp[1] == 0x97) { + /* 00D7 MULTIPLICATION SIGN */ + break_allowed[bn] = 0; + lp += 2; + continue; + } + if(isxdigit((unsigned char)*lp)) { + buffer[bn++] = strtoul(lp, &lp, 16); + continue; + } + fatal(0, "%s:%d: evil line: %s", path, lineno, l); + } + for(n = 0; n <= bn; ++n) { + if(breakfn(buffer, bn, n) != break_allowed[n]) { + fprintf(stderr, + "%s:%d: offset %zu: mismatch\n" + "%s\n" + "\n", + path, lineno, n, l); + count_error(); + } + ++tests; + } + xfree(l); + } + fclose(fp); +} + /** @brief Tests for @ref lib/unicode.h */ static void test_unicode(void) { FILE *fp; int lineno = 0; char *l, *lp; uint32_t buffer[1024]; - uint32_t *c[6], *NFD_c[6], *NFKD_c[6]; /* 1-indexed */ + uint32_t *c[6], *NFD_c[6], *NFKD_c[6], *NFC_c[6], *NFKC_c[6]; /* 1-indexed */ int cn, bn; fprintf(stderr, "test_unicode\n"); - if(!(fp = fopen("NormalizationTest.txt", "r"))) { - system("wget http://www.unicode.org/Public/5.0.0/ucd/NormalizationTest.txt"); - chmod("NormalizationTest.txt", 0444); - if(!(fp = fopen("NormalizationTest.txt", "r"))) { - perror("NormalizationTest.txt"); - ++tests; /* don't know how many... */ - ++errors; - return; - } - } + fp = open_unicode_test("NormalizationTest.txt"); while(!inputline("NormalizationTest.txt", fp, &l, '\n')) { ++lineno; if(*l == '#' || *l == '@') @@ -470,16 +652,22 @@ static void test_unicode(void) { for(cn = 1; cn <= 5; ++cn) { NFD_c[cn] = utf32_decompose_canon(c[cn], utf32_len(c[cn]), 0); NFKD_c[cn] = utf32_decompose_compat(c[cn], utf32_len(c[cn]), 0); + NFC_c[cn] = utf32_compose_canon(c[cn], utf32_len(c[cn]), 0); + NFKC_c[cn] = utf32_compose_compat(c[cn], utf32_len(c[cn]), 0); } #define unt_check(T, A, B) do { \ ++tests; \ if(utf32_cmp(c[A], T##_c[B])) { \ - fprintf(stderr, "L%d: c%d != "#T"(c%d)\n", lineno, A, B); \ - fprintf(stderr, " c%d: %s\n", \ + fprintf(stderr, \ + "NormalizationTest.txt:%d: c%d != "#T"(c%d)\n", \ + lineno, A, B); \ + fprintf(stderr, " c%d:%s\n", \ A, format_utf32(c[A])); \ - fprintf(stderr, "%4s(c%d): %s\n", \ + fprintf(stderr, " c%d:%s\n", \ + B, format_utf32(c[B])); \ + fprintf(stderr, "%4s(c%d):%s\n", \ #T, B, format_utf32(T##_c[B])); \ - ++errors; \ + count_error(); \ } \ } while(0) unt_check(NFD, 3, 1); @@ -492,15 +680,212 @@ static void test_unicode(void) { unt_check(NFKD, 5, 3); unt_check(NFKD, 5, 4); unt_check(NFKD, 5, 5); + unt_check(NFC, 2, 1); + unt_check(NFC, 2, 2); + unt_check(NFC, 2, 3); + unt_check(NFC, 4, 4); + unt_check(NFC, 4, 5); + unt_check(NFKC, 4, 1); + unt_check(NFKC, 4, 2); + unt_check(NFKC, 4, 3); + unt_check(NFKC, 4, 4); + unt_check(NFKC, 4, 5); for(cn = 1; cn <= 5; ++cn) { xfree(NFD_c[cn]); xfree(NFKD_c[cn]); } xfree(l); } + fclose(fp); + breaktest("auxiliary/GraphemeBreakTest.txt", utf32_is_grapheme_boundary); + breaktest("auxiliary/WordBreakTest.txt", utf32_is_word_boundary); + insist(utf32_combining_class(0x40000) == 0); + insist(utf32_combining_class(0xE0000) == 0); +} + +static void test_signame(void) { + fprintf(stderr, "test_signame\n"); + insist(find_signal("SIGTERM") == SIGTERM); + insist(find_signal("SIGHUP") == SIGHUP); + insist(find_signal("SIGINT") == SIGINT); + insist(find_signal("SIGQUIT") == SIGQUIT); + insist(find_signal("SIGKILL") == SIGKILL); + insist(find_signal("SIGYOURMUM") == -1); +} + +static void test_cache(void) { + const struct cache_type t1 = { 1 }, t2 = { 10 }; + const char v11[] = "spong", v12[] = "wibble", v2[] = "blat"; + fprintf(stderr, "test_cache\n"); + cache_put(&t1, "1_1", v11); + cache_put(&t1, "1_2", v12); + cache_put(&t2, "2", v2); + insist(cache_count() == 3); + insist(cache_get(&t2, "2") == v2); + insist(cache_get(&t1, "1_1") == v11); + insist(cache_get(&t1, "1_2") == v12); + insist(cache_get(&t1, "2") == 0); + insist(cache_get(&t2, "1_1") == 0); + insist(cache_get(&t2, "1_2") == 0); + insist(cache_get(&t1, "2") == 0); + insist(cache_get(&t2, "1_1") == 0); + insist(cache_get(&t2, "1_2") == 0); + sleep(2); + cache_expire(); + insist(cache_count() == 1); + insist(cache_get(&t1, "1_1") == 0); + insist(cache_get(&t1, "1_2") == 0); + insist(cache_get(&t2, "2") == v2); + cache_clean(0); + insist(cache_count() == 0); + insist(cache_get(&t2, "2") == 0); +} + +static void test_filepart(void) { + fprintf(stderr, "test_filepart\n"); + check_string(d_dirname("/"), "/"); + check_string(d_dirname("/spong"), "/"); + check_string(d_dirname("/foo/bar"), "/foo"); + check_string(d_dirname("./bar"), "."); + check_string(d_dirname("."), "."); + check_string(d_dirname(".."), "."); + check_string(d_dirname("../blat"), ".."); + check_string(d_dirname("wibble"), "."); + check_string(extension("foo.c"), ".c"); + check_string(extension(".c"), ".c"); + check_string(extension("."), "."); + check_string(extension("foo"), ""); + check_string(extension("./foo"), ""); + check_string(extension("./foo.c"), ".c"); +} + +static void test_selection(void) { + hash *h; + fprintf(stderr, "test_selection\n"); + insist((h = selection_new()) != 0); + selection_set(h, "one", 1); + selection_set(h, "two", 1); + selection_set(h, "three", 0); + selection_set(h, "four", 1); + insist(selection_selected(h, "one") == 1); + insist(selection_selected(h, "two") == 1); + insist(selection_selected(h, "three") == 0); + insist(selection_selected(h, "four") == 1); + insist(selection_selected(h, "five") == 0); + insist(hash_count(h) == 3); + selection_flip(h, "one"); + selection_flip(h, "three"); + insist(selection_selected(h, "one") == 0); + insist(selection_selected(h, "three") == 1); + insist(hash_count(h) == 3); + selection_live(h, "one"); + selection_live(h, "two"); + selection_live(h, "three"); + selection_cleanup(h); + insist(selection_selected(h, "one") == 0); + insist(selection_selected(h, "two") == 1); + insist(selection_selected(h, "three") == 1); + insist(selection_selected(h, "four") == 0); + insist(selection_selected(h, "five") == 0); + insist(hash_count(h) == 2); + selection_empty(h); + insist(selection_selected(h, "one") == 0); + insist(selection_selected(h, "two") == 0); + insist(selection_selected(h, "three") == 0); + insist(selection_selected(h, "four") == 0); + insist(selection_selected(h, "five") == 0); + insist(hash_count(h) == 0); +} + +static void test_wstat(void) { + pid_t pid; + int w; + + fprintf(stderr, "test_wstat\n"); + if(!(pid = xfork())) { + _exit(1); + } + while(waitpid(pid, &w, 0) < 0 && errno == EINTR) + ; + check_string(wstat(w), "exited with status 1"); + if(!(pid = xfork())) { + kill(getpid(), SIGTERM); + _exit(-1); + } + while(waitpid(pid, &w, 0) < 0 && errno == EINTR) + ; + check_string_prefix(wstat(w), "terminated by signal 15"); +} + +static void test_kvp(void) { + struct kvp *k; + size_t n; + + fprintf(stderr, "test_kvp\n"); + /* decoding */ +#define KVP_URLDECODE(S) kvp_urldecode((S), strlen(S)) + insist(KVP_URLDECODE("=%zz") == 0); + insist(KVP_URLDECODE("=%0") == 0); + insist(KVP_URLDECODE("=%0z") == 0); + insist(KVP_URLDECODE("=%%") == 0); + insist(KVP_URLDECODE("==%") == 0); + insist(KVP_URLDECODE("wibble") == 0); + insist(KVP_URLDECODE("") == 0); + insist(KVP_URLDECODE("wibble&") == 0); + insist((k = KVP_URLDECODE("one=bl%61t+foo")) != 0); + check_string(kvp_get(k, "one"), "blat foo"); + insist(kvp_get(k, "ONE") == 0); + insist(k->next == 0); + insist((k = KVP_URLDECODE("wibble=splat&bar=spong")) != 0); + check_string(kvp_get(k, "wibble"), "splat"); + check_string(kvp_get(k, "bar"), "spong"); + insist(kvp_get(k, "ONE") == 0); + insist(k->next->next == 0); + /* encoding */ + insist(kvp_set(&k, "bar", "spong") == 0); + insist(kvp_set(&k, "bar", "foo") == 1); + insist(kvp_set(&k, "zog", "%") == 1); + insist(kvp_set(&k, "wibble", 0) == 1); + insist(kvp_set(&k, "wibble", 0) == 0); + check_string(kvp_urlencode(k, 0), + "bar=foo&zog=%25"); + check_string(kvp_urlencode(k, &n), + "bar=foo&zog=%25"); + insist(n == strlen("bar=foo&zog=%25")); + check_string(urlencodestring("abc% +\n"), + "abc%25%20%2b%0a"); +} + +static void test_sink(void) { + struct sink *s; + struct dynstr d[1]; + FILE *fp; + char *l; + + fprintf(stderr, "test_sink\n"); + + fp = tmpfile(); + assert(fp != 0); + s = sink_stdio("tmpfile", fp); + insist(sink_printf(s, "test: %d\n", 999) == 10); + insist(sink_printf(s, "wibble: %s\n", "foobar") == 15); + rewind(fp); + insist(inputline("tmpfile", fp, &l, '\n') == 0); + check_string(l, "test: 999"); + insist(inputline("tmpfile", fp, &l, '\n') == 0); + check_string(l, "wibble: foobar"); + insist(inputline("tmpfile", fp, &l, '\n') == -1); + + dynstr_init(d); + s = sink_dynstr(d); + insist(sink_printf(s, "test: %d\n", 999) == 10); + insist(sink_printf(s, "wibble: %s\n", "foobar") == 15); + dynstr_terminate(d); + check_string(d->vec, "test: 999\nwibble: foobar\n"); } int main(void) { + fail_first = !!getenv("FAIL_FIRST"); insist('\n' == 0x0A); insist('\r' == 0x0D); insist(' ' == 0x20); @@ -518,6 +903,8 @@ int main(void) { /* client.c */ /* configuration.c */ /* event.c */ + /* filepart.c */ + test_filepart(); /* fprintf.c */ /* heap.c */ test_heap(); @@ -525,6 +912,7 @@ int main(void) { test_hex(); /* inputline.c */ /* kvp.c */ + test_kvp(); /* log.c */ /* mem.c */ /* mime.c */ @@ -534,6 +922,7 @@ int main(void) { /* printf.c */ /* queue.c */ /* sink.c */ + test_sink(); /* snprintf.c */ /* split.c */ /* syscalls.c */ @@ -545,8 +934,15 @@ int main(void) { /* vector.c */ /* words.c */ test_casefold(); - /* XXX words() */ + test_words(); /* wstat.c */ + test_wstat(); + /* signame.c */ + test_signame(); + /* cache.c */ + test_cache(); + /* selection.c */ + test_selection(); fprintf(stderr, "%d errors out of %d tests\n", errors, tests); return !!errors; } @@ -555,6 +951,7 @@ int main(void) { Local Variables: c-basic-offset:2 comment-column:40 +fill-column:79 +indent-tabs-mode:nil End: */ -