data: one trie for maps and dictionaries, one loader

The hash map served a single exact-match lookup that the trie already answers, at the price of a second container, a second file loader with its own drift, and a Str-to-UTF-8 conversion on every chain probe. The files are small (kanji.dict is 7.5k lines), so the trie holds both. trieopen validates keys and each space-separated value word against Str and reports path:line; trielookup takes the Str every caller holds and treats a nil trie as an unloaded map; trienew/trieput exist for tests. The overflow guards on growth, Trie.root (always 0) and the per-language init loop written twice are gone.
2026-08-16 15:56:32 +09:00
parent eb0f88f764
commit c7718ece52
15 changed files with 206 additions and 609 deletions
--- a/tests/Makefile
+++ b/tests/Makefile
@@ -28,11 +28,11 @@ SMOKE = ibus_live_test ibus_client_smoke gtk_live_test xim_live_test \
 	ipc_live_test
 FAULT = daemon_collision_test daemon_failure_test daemon_restart_test
 LIVE = $(SMOKE) $(FAULT)
-TESTSRC = test_util.c str_test.c hash_test.c trie_test.c \
+TESTSRC = test_util.c str_test.c trie_test.c \
 	ko_test.c vi_test.c engine_test.c dict_test.c ipc_test.c \
 	popup_test.c font_test.c ibus_test.c server_test.c xim_adapter_test.c
 TESTOBJ = $(TESTSRC:.c=.o)
-PARENTSRC = str.c hash.c trie.c dict.c ko.c vi.c ipc.c popup_layout.c \
+PARENTSRC = str.c trie.c dict.c ko.c vi.c ipc.c popup_layout.c \
 	font.c
 PARENTOBJ = $(PARENTSRC:%.c=unit_%.o)
 XIMOBJ = unit_xim_keymap.o unit_ximtext.o
--- a/tests/dict_test.c
+++ b/tests/dict_test.c
@@ -3,21 +3,20 @@
 void
 dictionary_candidates(struct ct *t)
 {
-	static char malformed[] = { 'o', 'k', ' ', (char)0x80 };
 	char many[512], item[8];
 	char *p;
 	Dictreq req;
 	Dictres res;
-	Hmap *saved;
+	Trie *saved;
 	Lang *lang;
 	Str key;
 	int i;

 	lang = getlang(LangJP);
 	saved = lang->dict;
-	lang->dict = hmapalloc(1);
+	lang->dict = trienew();
 	key = mkstr("かな");
-	hmapset(&lang->dict, &key, "  候補1  かな 候補2  ",
+	trieput(lang->dict, "かな", strlen("かな"), "  候補1  かな 候補2  ",
 		strlen("  候補1  かな 候補2  "));
 	memset(&req, 0, sizeof req);
 	req.key = key;
@@ -42,7 +41,7 @@ dictionary_candidates(struct ct *t)
 		p += strlen(item);
 	}
 	*p = '\0';
-	hmapset(&lang->dict, &key, many, strlen(many));
+	trieput(lang->dict, "key", 3, many, strlen(many));
 	req.key = key;
 	req.pre = mkstr("preedit-two");
 	req.seq = 29;
@@ -54,13 +53,8 @@ dictionary_candidates(struct ct *t)
 	CT_EQ_INT(t, 0, scmp(&req.pre, &res.key));
 	checkstr(t, "first capped candidate", "c00", &res.kouho[0]);
 	checkstr(t, "last capped candidate", "c31", &res.kouho[31]);
-	key = mkstr("malformed");
-	hmapset(&lang->dict, &key, malformed, sizeof malformed);
-	req.key = key;
-	dictlookup(&req, &res);
-	CT_EQ_INT(t, 0, res.nkouho);
 cleanup:
-	hmapfree(lang->dict);
+	trieclose(lang->dict);
 	lang->dict = saved;
 }

@@ -77,13 +71,13 @@ dictionary_misses_clear_result(struct ct *t)
 	};
 	Dictreq req;
 	Dictres res;
-	Hmap *saved;
+	Trie *saved;
 	Lang *lang;
 	int i;

 	lang = getlang(LangJP);
 	saved = lang->dict;
-	lang->dict = hmapalloc(1);
+	lang->dict = trienew();
 	for(i = 0; i < nelem(cases); i++){
 		memset(&res, 0xa5, sizeof res);
 		req.key = mkstr(cases[i].key);
@@ -98,7 +92,7 @@ dictionary_misses_clear_result(struct ct *t)
 		CT_EQ_UINT(t, 0xf00d0000U + i, res.seq);
 		checkstr(t, cases[i].name, cases[i].pre, &res.key);
 	}
-	hmapfree(lang->dict);
+	trieclose(lang->dict);
 	lang->dict = saved;
 }

@@ -107,15 +101,15 @@ dictionary_emoji_identity(struct ct *t)
 {
 	Dictreq req;
 	Dictres res;
-	Hmap *saved;
+	Trie *saved;
 	Lang *lang;
 	Str key;

 	lang = getlang(LangEMOJI);
 	saved = lang->dict;
-	lang->dict = hmapalloc(2);
+	lang->dict = trienew();
 	key = mkstr("é");
-	hmapset(&lang->dict, &key, "é", strlen("é"));
+	trieput(lang->dict, "é", strlen("é"), "é", strlen("é"));
 	memset(&req, 0, sizeof req);
 	req.key = key;
 	req.pre = key;
@@ -124,6 +118,6 @@ dictionary_emoji_identity(struct ct *t)
 	dictlookup(&req, &res);
 	if(CT_EQ_INT(t, 1, res.nkouho))
 		checkstr(t, "emoji identity candidate", "é", &res.kouho[0]);
-	hmapfree(lang->dict);
+	trieclose(lang->dict);
 	lang->dict = saved;
 }
--- a/tests/engine_test.c
+++ b/tests/engine_test.c
@@ -311,14 +311,14 @@ engine_rejects_stale_dictionary_results(struct ct *t)
 {
 	Dictreq a, b;
 	Dictres res;
-	Hmap *saved;
+	Trie *saved;
 	Lang *jp;
 	char one, two;
 	int n;

 	jp = getlang(LangJP);
 	saved = jp->dict;
-	jp->dict = hmapalloc(1);
+	jp->dict = trienew();
 	init();
 	im.l = jp;
 	while(channbrecv(dictreqc, &a) > 0)
@@ -362,7 +362,7 @@ engine_rejects_stale_dictionary_results(struct ct *t)
 cleanup:
 	while(channbrecv(dictreqc, &a) > 0)
 		;
-	hmapfree(jp->dict);
+	trieclose(jp->dict);
 	jp->dict = saved;
 }

@@ -720,7 +720,7 @@ struct Searchfix
 	Im	im;
 	Search	search;
 	Lang	*dictlang;
-	Hmap	*dict;
+	Trie	*dict;
 	int	activecap;
 	int	visible;
 	int	candidatechosen;
@@ -939,12 +939,9 @@ engine_candidate_completion(struct ct *t)
 }

 static void
-setdict(Hmap **dict, char *key, char *val)
+setdict(Trie *dict, char *key, char *val)
 {
-	Str s;
-
-	s = mkstr(key);
-	hmapset(dict, &s, val, strlen(val));
+	trieput(dict, key, strlen(key), val, strlen(val));
 }

 static void
@@ -964,22 +961,22 @@ static void
 emojibegin(Searchfix *f, int showpre)
 {
 	searchsave(f, LangEMOJI);
-	f->dictlang->dict = hmapalloc(32);
-	setdict(&f->dictlang->dict, "a", "A B");
-	setdict(&f->dictlang->dict, "ㅁ", "B C");
-	setdict(&f->dictlang->dict, "alpha", "α");
-	setdict(&f->dictlang->dict, "smil", "😀 😄");
-	setdict(&f->dictlang->dict, "smile", "😀 😄");
-	setdict(&f->dictlang->dict, "웃음", "😀 😄 😂");
-	setdict(&f->dictlang->dict, "えがお", "😀 😄 😊");
-	setdict(&f->dictlang->dict, "エガオ", "😀 😄 😊");
-	setdict(&f->dictlang->dict, "heart", "❤️");
-	setdict(&f->dictlang->dict, "coffee", "☕️");
-	setdict(&f->dictlang->dict, "^", "¹ ² ³ ⁴ ⁵ ⁶ ⁷ ⁸ ⁹ ⁽");
-	setdict(&f->dictlang->dict, "^0", "⁰");
-	setdict(&f->dictlang->dict, "_", "₁ ₂ ₃ ₄ ₅ ₆ ₇ ₈ ₉ ₍");
-	setdict(&f->dictlang->dict, "<", "← ≤ ♥ ≠");
-	setdict(&f->dictlang->dict, "many",
+	f->dictlang->dict = trienew();
+	setdict(f->dictlang->dict, "a", "A B");
+	setdict(f->dictlang->dict, "ㅁ", "B C");
+	setdict(f->dictlang->dict, "alpha", "α");
+	setdict(f->dictlang->dict, "smil", "😀 😄");
+	setdict(f->dictlang->dict, "smile", "😀 😄");
+	setdict(f->dictlang->dict, "웃음", "😀 😄 😂");
+	setdict(f->dictlang->dict, "えがお", "😀 😄 😊");
+	setdict(f->dictlang->dict, "エガオ", "😀 😄 😊");
+	setdict(f->dictlang->dict, "heart", "❤️");
+	setdict(f->dictlang->dict, "coffee", "☕️");
+	setdict(f->dictlang->dict, "^", "¹ ² ³ ⁴ ⁵ ⁶ ⁷ ⁸ ⁹ ⁽");
+	setdict(f->dictlang->dict, "^0", "⁰");
+	setdict(f->dictlang->dict, "_", "₁ ₂ ₃ ₄ ₅ ₆ ₇ ₈ ₉ ₍");
+	setdict(f->dictlang->dict, "<", "← ≤ ♥ ≠");
+	setdict(f->dictlang->dict, "many",
 		"c1 c2 c3 c4 c5 c6 c7 c8 c9 c10 c11 c12");
 	init();
 	im.l = getlang(LangKO);
@@ -989,11 +986,11 @@ emojibegin(Searchfix *f, int showpre)
 static void
 searchend(Searchfix *f)
 {
-	Hmap *dict;
+	Trie *dict;

 	dict = f->dictlang->dict;
 	f->dictlang->dict = f->dict;
-	hmapfree(dict);
+	trieclose(dict);
 	draindraw(nil);
 	im = f->im;
 	search = f->search;
@@ -1079,12 +1076,12 @@ static void
 hanjabegin(Searchfix *f, int lang)
 {
 	searchsave(f, LangHANJA);
-	f->dictlang->dict = hmapalloc(8);
-	setdict(&f->dictlang->dict, "한", "漢 韓");
-	setdict(&f->dictlang->dict, "가", "家");
-	setdict(&f->dictlang->dict, "ㄱ", "假");
-	setdict(&f->dictlang->dict, "한글", "文");
-	setdict(&f->dictlang->dict, "か", "仮");
+	f->dictlang->dict = trienew();
+	setdict(f->dictlang->dict, "한", "漢 韓");
+	setdict(f->dictlang->dict, "가", "家");
+	setdict(f->dictlang->dict, "ㄱ", "假");
+	setdict(f->dictlang->dict, "한글", "文");
+	setdict(f->dictlang->dict, "か", "仮");
 	init();
 	im.l = getlang(lang);
 }
@@ -1156,15 +1153,14 @@ engine_japanese_candidates(struct ct *t)
 {
 	Dictreq req;
 	Dictres res;
-	Hmap *saved;
+	Trie *saved;
 	Lang *jp;
-	Str com, key, shown;
+	Str com, shown;

 	jp = getlang(LangJP);
 	saved = jp->dict;
-	jp->dict = hmapalloc(8);
-	key = mkstr("かんじ");
-	hmapset(&jp->dict, &key, "漢字 幹事", strlen("漢字 幹事"));
+	jp->dict = trienew();
+	setdict(jp->dict, "かんじ", "漢字 幹事");
 	init();
 	im.l = jp;
 	sclear(&com);
@@ -1189,7 +1185,7 @@ engine_japanese_candidates(struct ct *t)
 	checkstr(t, "selected Kanji", "幹事", &com);
 cleanup:
 	newestrequest(&req);
-	hmapfree(jp->dict);
+	trieclose(jp->dict);
 	jp->dict = saved;
 }

--- a/tests/hash_test.c
+++ b/tests/hash_test.c
@@ -1,112 +0,0 @@
-#include "test.h"
-
-void
-hmap_set_replace_and_grow(struct ct *t)
-{
-	char keybuf[16], valbuf[16], source[] = "copied";
-	Hmap *h;
-	Hnode *n;
-	Str key;
-	int i;
-
-	h = hmapalloc(1);
-	for(i = 0; i < 4; i++){
-		snprint(keybuf, sizeof keybuf, "key%d", i);
-		snprint(valbuf, sizeof valbuf, "value%d", i);
-		key = mkstr(keybuf);
-		hmapset(&h, &key, valbuf, strlen(valbuf));
-	}
-	for(i = 0; i < 4; i++){
-		snprint(keybuf, sizeof keybuf, "key%d", i);
-		snprint(valbuf, sizeof valbuf, "value%d", i);
-		key = mkstr(keybuf);
-		n = hmapget(h, &key);
-		if(n == nil || strcmp(n->val, valbuf) != 0)
-			CT_ERRORF(t, "%s: collision chain lost value", keybuf);
-	}
-	key = mkstr("key1");
-	hmapset(&h, &key, source, strlen(source));
-	source[0] = 'X';
-	n = hmapget(h, &key);
-	if(!CT_CHECK(t, n != nil))
-		goto cleanup;
-	CT_EQ_STR(t, "copied", n->val);
-	key = mkstr("key2");
-	n = hmapget(h, &key);
-	if(!CT_CHECK(t, n != nil))
-		goto cleanup;
-	CT_EQ_STR(t, "value2", n->val);
-	key = mkstr("key1");
-	hmapset(&h, &key, nil, 0);
-	n = hmapget(h, &key);
-	if(!CT_CHECK(t, n != nil))
-		goto cleanup;
-	CT_EQ_INT(t, 0, n->vlen);
-	CT_EQ_PTR(t, nil, n->val);
-	key = mkstr("missing");
-	CT_EQ_PTR(t, nil, hmapget(h, &key));
-cleanup:
-	hmapfree(h);
-}
-
-void
-hmap_long_utf8_keys(struct ct *t)
-{
-	Hmap *h;
-	Hnode *n;
-	Str a, b;
-	int i;
-
-	for(i = 0; i < Maxrunes-1; i++)
-		a.r[i] = b.r[i] = 0x1f600;
-	a.r[Maxrunes-1] = 0x1f601;
-	b.r[Maxrunes-1] = 0x1f602;
-	a.n = b.n = Maxrunes;
-	h = hmapalloc(1);
-	hmapset(&h, &a, "first", 5);
-	hmapset(&h, &b, "second", 6);
-	n = hmapget(h, &a);
-	if(!CT_CHECK(t, n != nil))
-		goto cleanup;
-	CT_EQ_STR(t, "first", n != nil ? n->val : nil);
-	n = hmapget(h, &b);
-	if(!CT_CHECK(t, n != nil))
-		goto cleanup;
-	CT_EQ_STR(t, "second", n != nil ? n->val : nil);
-cleanup:
-	hmapfree(h);
-}
-
-void
-hmap_binary_keys_and_invalid_lengths(struct ct *t)
-{
-	Hmap *h;
-	Hnode *n;
-	Str key, other;
-
-	h = hmapalloc(1);
-	CT_CHECK(t, h != nil);
-	key.n = 3;
-	key.r[0] = 'a';
-	key.r[1] = 0;
-	key.r[2] = 'b';
-	other = key;
-	other.r[2] = 'c';
-	hmapset(&h, &key, "one", 3);
-	hmapset(&h, &other, "two", 3);
-	n = hmapget(h, &key);
-	if(CT_CHECK(t, n != nil)){
-		CT_EQ_INT(t, 3, n->klen);
-		CT_EQ_MEM(t, "one", n->val, n->vlen);
-	}
-	n = hmapget(h, &other);
-	if(CT_CHECK(t, n != nil))
-		CT_EQ_MEM(t, "two", n->val, n->vlen);
-	hmapset(&h, &key, "bad", -1);
-	hmapset(&h, &key, nil, 1);
-	n = hmapget(h, &key);
-	if(CT_CHECK(t, n != nil))
-		CT_EQ_MEM(t, "one", n->val, n->vlen);
-	CT_EQ_PTR(t, nil, hmapalloc(0));
-	hmapfree(h);
-}
--- a/tests/test.h
+++ b/tests/test.h
@@ -12,11 +12,8 @@ void	str_init_utf8(struct ct*);
 void	str_edit_and_alias(struct ct*);
 void	str_utf8_capacity(struct ct*);
 void	str_invalid_and_full_appends(struct ct*);
-void	hmap_set_replace_and_grow(struct ct*);
-void	hmap_long_utf8_keys(struct ct*);
-void	hmap_binary_keys_and_invalid_lengths(struct ct*);
 void	trie_exact_prefix_and_duplicate(struct ct*);
-void	trie_optional_outputs_and_invalid_lengths(struct ct*);
+void	trie_put_and_unloaded(struct ct*);
 void	popup_layout(struct ct*);
 void	font_render(struct ct*);
 void	production_maps_load(struct ct*);
--- a/tests/trie_test.c
+++ b/tests/trie_test.c
@@ -5,26 +5,26 @@ trie_exact_prefix_and_duplicate(struct ct *t)
 {
 	static const struct {
 		char *key;
-		int klen;
 		int match;
 		char *want;
 	} cases[] = {
-		{ "a", 1, TrieExact, "alpha" },
-		{ "dupli", 5, TriePrefix, nil },
-		{ "duplicate", 9, TrieExact, "second" },
-		{ "missing", 7, TrieMiss, nil },
-		{ nil, 1, TrieMiss, nil },
-		{ "a", -1, TrieMiss, nil },
+		{ "a", TrieExact, "alpha" },
+		{ "ab", TrieExact, "beta" },
+		{ "한", TrieExact, "값" },
+		{ "dupli", TriePrefix, nil },
+		{ "duplicate", TrieExact, "second" },
+		{ "missing", TrieMiss, nil },
+		{ "", TriePrefix, nil },
 	};
 	char *v;
 	Trie *trie;
+	Str key;
 	int i, match, n;

 	trie = trieopen("data/trie.map");
 	for(i = 0; i < nelem(cases); i++){
-		v = "unchanged";
-		n = 77;
-		match = trielookup(trie, cases[i].key, cases[i].klen, &v, &n);
+		key = mkstr(cases[i].key);
+		match = trielookup(trie, &key, &v, &n);
 		if(match != cases[i].match){
 			CT_ERRORF(t, "case %d: want match %d, got %d",
 				i, cases[i].match, match);
@@ -42,22 +42,28 @@ trie_exact_prefix_and_duplicate(struct ct *t)
 }

 void
-trie_optional_outputs_and_invalid_lengths(struct ct *t)
+trie_put_and_unloaded(struct ct *t)
 {
 	char *v;
 	Trie *trie;
+	Str key;
 	int n;

-	trie = trieopen("data/trie.map");
-	CT_EQ_INT(t, TrieExact, trielookup(trie, "a", 1, &v, nil));
-	CT_CHECK(t, v != nil);
-	CT_EQ_INT(t, TrieExact, trielookup(trie, "a", 1, nil, &n));
-	CT_EQ_INT(t, 5, n);
-	CT_EQ_INT(t, TriePrefix,
-		trielookup(trie, "dupli", 5, nil, nil));
-	CT_EQ_INT(t, TriePrefix, trielookup(trie, nil, 0, &v, &n));
-	CT_EQ_PTR(t, nil, v);
-	CT_EQ_INT(t, 0, n);
+	key = mkstr("k");
+	CT_EQ_INT(t, TrieMiss, trielookup(nil, &key, &v, &n));
+	trie = trienew();
+	CT_EQ_INT(t, TrieMiss, trielookup(trie, &key, &v, &n));
+	trieput(trie, "k", 1, "one two", 7);
+	trieput(trie, "ka", 2, "", 0);
+	if(CT_EQ_INT(t, TrieExact, trielookup(trie, &key, &v, &n)))
+		CT_EQ_MEM(t, "one two", v, 7);
+	key = mkstr("ka");
+	if(CT_EQ_INT(t, TrieExact, trielookup(trie, &key, &v, &n)))
+		CT_EQ_INT(t, 0, n);
+	trieput(trie, "k", 1, "three", 5);
+	key = mkstr("k");
+	if(CT_EQ_INT(t, TrieExact, trielookup(trie, &key, &v, &n)))
+		CT_EQ_MEM(t, "three", v, 5);
 	trieclose(trie);
 }

--- a/tests/unit_test.c
+++ b/tests/unit_test.c
@@ -67,11 +67,8 @@ static const struct ct_test tests[] = {
 	{ "str/edit-and-alias", str_edit_and_alias },
 	{ "str/utf8-capacity", str_utf8_capacity },
 	{ "str/invalid-full-appends", str_invalid_and_full_appends },
-	{ "hmap/set-replace-grow", hmap_set_replace_and_grow },
-	{ "hmap/long-utf8-keys", hmap_long_utf8_keys },
-	{ "hmap/binary-invalid-lengths", hmap_binary_keys_and_invalid_lengths },
 	{ "trie/exact-prefix-duplicate", trie_exact_prefix_and_duplicate },
-	{ "trie/optional-invalid-lengths", trie_optional_outputs_and_invalid_lengths },
+	{ "trie/put-and-unloaded", trie_put_and_unloaded },
 	{ "popup/layout", popup_layout },
 	{ "font/render", font_render },
 	{ "map/production-lifecycle", production_maps_load },