Files
jak-project/common/util/FontUtils.cpp
T
ManDude 5b44aece75 random fixes + support clang-cl on visual studio (#1129)
* delete unused shaders

* hide some options in debug menu

* change fullscreen logic a bit

* add "all actors" toggle

* borderless fix and fix alpha in direct renderer untextured (do we need a separate shader for that?)

* fix fuel cell orbit icons in widescreen

* fix `curve` types

* refs

* fix levitator task...

* fix some task stuff

* update font code a bit (temp)

* cmake, third-party and visual studio overhaul

* Update .gitmodules

* update modules

* clone repos

* fix encoding in zydis

* where did these come from

* try again

* add submodule

* Update 11zip

* Update 11zip

* Update 11zip

* delete

* try again

* clang

* update compiler flags

* delete 11zip. go away.

* Create memory-dump-p2s.py

* properly

* fix minimum architecture c++ compiler flags

* fix zydis

* oops

* Update all-types.gc

* fix clang-cl tests

* make "all actors" work better, entity debug qol

* update game-text conversion code to be more modularized

* Create vendor.txt

* fix typos and minor things

* update refs

* clang

* Attempt to add clang-cl support to vs2019 and CI

* vs2022 + clang-cl

* srsly? fix clang build

* Update launch.vs.json

* extend windows CI timer
2022-02-07 19:15:37 -05:00

521 lines
15 KiB
C++
Raw Blame History

This file contains invisible Unicode characters
This file contains invisible Unicode characters that are indistinguishable to humans but may be processed differently by a computer. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
/*!
* @file FontUtils.cpp
*
* Code for handling text and strings in Jak 1's "large font" format.
*
* MAKE SURE THIS FILE IS ENCODED IN UTF-8!!! The various strings here depend on it.
* Always verify the encoding if string detection suddenly goes awry.
*/
#include <algorithm>
#include "FontUtils.h"
#include "third-party/fmt/core.h"
GameTextFontBank::GameTextFontBank(GameTextVersion version,
std::vector<EncodeInfo>* encode_info,
std::vector<ReplaceInfo>* replace_info,
std::unordered_set<char>* passthrus)
: m_version(version),
m_encode_info(encode_info),
m_replace_info(replace_info),
m_passthrus(passthrus) {
std::sort(
m_encode_info->begin(), m_encode_info->end(),
[](const EncodeInfo& a, const EncodeInfo& b) { return a.bytes.size() > b.bytes.size(); });
std::sort(
m_replace_info->begin(), m_replace_info->end(),
[](const ReplaceInfo& a, const ReplaceInfo& b) { return a.from.size() > b.from.size(); });
}
/*!
* Finds a remap info that best matches the byte sequence (is the longest match).
*/
const EncodeInfo* GameTextFontBank::find_encode_to_utf8(const char* in) const {
const EncodeInfo* best_info = nullptr;
for (auto& info : *m_encode_info) {
if (info.bytes.size() == 0)
continue;
bool found = true;
for (int i = 0; found && i < (int)info.bytes.size(); ++i) {
if (uint8_t(in[i]) != info.bytes.at(i)) {
found = false;
}
}
if (found && (!best_info || info.chars.length() > best_info->chars.length())) {
best_info = &info;
}
}
return best_info;
}
/*!
* Finds a remap info that best matches the character sequence (is the longest match).
*/
const EncodeInfo* GameTextFontBank::find_encode_to_game(const std::string& in, int off) const {
const EncodeInfo* best_info = nullptr;
for (auto& info : *m_encode_info) {
if (info.chars.length() == 0)
continue;
bool found = true;
for (int i = 0; found && i < (int)info.chars.length() && i + off < in.size(); ++i) {
if (in.at(i + off) != info.chars.at(i)) {
found = false;
}
}
if (found && (!best_info || info.chars.length() > best_info->chars.length())) {
best_info = &info;
}
}
return best_info;
}
/*!
* Try to replace specific substrings with better variants.
* These are for hiding confusing text transforms.
*/
std::string GameTextFontBank::replace_to_utf8(std::string& str) const {
for (auto& info : *m_replace_info) {
auto pos = str.find(info.from);
while (pos != std::string::npos) {
str.replace(pos, info.from.size(), info.to);
pos = str.find(info.from, pos + info.to.size());
}
}
return str;
}
std::string GameTextFontBank::replace_to_game(std::string& str) const {
for (auto& info : *m_replace_info) {
auto pos = str.find(info.to);
while (pos != std::string::npos) {
str.replace(pos, info.to.size(), info.from);
pos = str.find(info.to, pos + info.from.size());
}
}
return str;
}
std::string GameTextFontBank::encode_utf8_to_game(std::string& str) const {
std::string new_str;
for (int i = 0; i < str.length();) {
auto remap = find_encode_to_game(str, i);
if (!remap) {
new_str.push_back(str.at(i));
i += 1;
} else {
for (auto b : remap->bytes) {
new_str.push_back(b);
}
i += remap->chars.length();
}
}
str = new_str;
return str;
}
/*!
* Turn a normal readable string into a string readable in the Jak 1 font encoding.
*/
std::string GameTextFontBank::convert_utf8_to_game(std::string str) const {
replace_to_game(str);
encode_utf8_to_game(str);
return str;
}
/*!
* Convert a string from the game-text font encoding to something normal.
* Unprintable characters become escape sequences, including tab and newline.
*/
std::string GameTextFontBank::convert_game_to_utf8(const char* in) const {
std::string result;
while (*in) {
auto remap = find_encode_to_utf8(in);
if (remap != nullptr) {
result.append(remap->chars);
in += remap->bytes.size() - 1;
} else if (((*in >= '0' && *in <= '9') || (*in >= 'A' && *in <= 'Z') ||
m_passthrus->find(*in) != m_passthrus->end()) &&
*in != '\\') {
result.push_back(*in);
} else if (*in == '\n') {
result += "\\n";
} else if (*in == '\t') {
result += "\\t";
} else if (*in == '\\') {
result += "\\\\";
} else {
result += fmt::format("\\c{:02x}", uint8_t(*in));
}
in++;
}
return replace_to_utf8(result);
}
/*!
* ===========================
* GAME TEXT FONT BANK - JAK 1
* ===========================
* This font is used in:
* - Jak & Daxter: The Precursor Legacy (Black Label)
*/
static std::unordered_set<char> passthrus = {'~', ' ', ',', '.', '-', '+', '(', ')', '!', ':', '?',
'=', '%', '*', '/', '#', ';', '<', '>', '@', '[', '_'};
static std::vector<EncodeInfo> g_encode_info_jak1 = {
// random
{"ˇ", {0x10}}, // caron
{"`", {0x11}}, // grave accent
{"'", {0x12}}, // apostrophe
{"^", {0x13}}, // circumflex
{"<TIL>", {0x14}}, // tilde
{"¨", {0x15}}, // umlaut
{"º", {0x16}}, // numero/overring
{"¡", {0x17}}, // inverted exclamation mark
{"¿", {0x18}}, // inverted question mark
{"", {0x1a}}, // umi
{"Æ", {0x1b}}, // aesc
{"", {0x1c}}, // kai
{"Ç", {0x1d}}, // c-cedilla
{"", {0x1e}}, // gaku
{"ß", {0x1f}}, // eszett
{"\"", {0x22}}, // double-quotes
{"", {0x24}}, // wa
{"", {0x26}}, // wo
{"", {0x27}}, // -n
{"", {0x5c}}, // iwa
{"", {0x5d}}, // kyuu
{"", {0x5e}}, // sora
//{"掘", {0x5f}}, // horu
{"", {0x60}}, // -wa
{"", {0x61}}, // utsu
{"", {0x62}}, // kashikoi
{"", {0x63}}, // mizuumi
{"", {0x64}}, // kuchi
{"", {0x65}}, // iku
{"", {0x66}}, // ai
{"", {0x67}}, // shi
{"", {0x68}}, // tera
{"", {0x69}}, // yama
{"", {0x6a}}, // mono
{"", {0x6b}}, // tokoro
{"", {0x6c}}, // kaku
{"", {0x6d}}, // shou
{"", {0x6e}}, // numa
{"", {0x6f}}, // ue
{"", {0x70}}, // shiro
{"", {0x71}}, // ba
{"", {0x72}}, // shutsu
{"", {0x73}}, // yami
{"", {0x74}}, // nokosu
{"", {0x75}}, // ki
{"", {0x76}}, // ya
{"", {0x77}}, // shita
{"", {0x78}}, // ie
{"", {0x79}}, // hi
{"", {0x7a}}, // hana
{"", {0x7b}}, // re
{"Œ", {0x7c}}, // oe
{"", {0x7d}}, // ro
{"", {0x7f}}, // ao
{"", {0x90}}, // nakaguro
{"", {0x91}}, // dakuten
{"", {0x92}}, // handakuten
{"", {0x93}}, // chouompu
{"", {0x94}}, // nijuukagikakko left
{"", {0x95}}, // nijuukagikakko right
// hiragana
{"", {0x96}}, // -a
{"", {0x97}}, // a
{"", {0x98}}, // -i
{"", {0x99}}, // i
{"", {0x9a}}, // -u
{"", {0x9b}}, // u
{"", {0x9c}}, // -e
{"", {0x9d}}, // e
{"", {0x9e}}, // -o
{"", {0x9f}}, // o
{"", {0xa0}}, // ka
{"", {0xa1}}, // ki
{"", {0xa2}}, // ku
{"", {0xa3}}, // ke
{"", {0xa4}}, // ko
{"", {0xa5}}, // sa
{"", {0xa6}}, // shi
{"", {0xa7}}, // su
{"", {0xa8}}, // se
{"", {0xa9}}, // so
{"", {0xaa}}, // ta
{"", {0xab}}, // chi
{"", {0xac}}, // sokuon
{"", {0xad}}, // tsu
{"", {0xae}}, // te
{"", {0xaf}}, // to
{"", {0xb0}}, // na
{"", {0xb1}}, // ni
{"", {0xb2}}, // nu
{"", {0xb3}}, // ne
{"", {0xb4}}, // no
{"", {0xb5}}, // ha
{"", {0xb6}}, // hi
{"", {0xb7}}, // hu
{"", {0xb8}}, // he
{"", {0xb9}}, // ho
{"", {0xba}}, // ma
{"", {0xbb}}, // mi
{"", {0xbc}}, // mu
{"", {0xbd}}, // me
{"", {0xbe}}, // mo
{"", {0xbf}}, // youon ya
{"", {0xc0}}, // ya
{"", {0xc1}}, // youon yu
{"", {0xc2}}, // yu
{"", {0xc3}}, // youon yo
{"", {0xc4}}, // yo
{"", {0xc5}}, // ra
{"", {0xc6}}, // ri
{"", {0xc7}}, // ru
{"", {0xc8}}, // re
{"", {0xc9}}, // ro
{"", {0xca}}, // -wa
{"", {0xcb}}, // wa
{"", {0xcc}}, // wo
{"", {0xcd}}, // -n
// katakana
{"", {0xce}}, // -a
{"", {0xcf}}, // a
{"", {0xd0}}, // -i
{"", {0xd1}}, // i
{"", {0xd2}}, // -u
{"", {0xd3}}, // u
{"", {0xd4}}, // -e
{"", {0xd5}}, // e
{"", {0xd6}}, // -o
{"", {0xd7}}, // o
{"", {0xd8}}, // ka
{"", {0xd9}}, // ki
{"", {0xda}}, // ku
{"", {0xdb}}, // ke
{"", {0xdc}}, // ko
{"", {0xdd}}, // sa
{"", {0xde}}, // shi
{"", {0xdf}}, // su
{"", {0xe0}}, // se
{"", {0xe1}}, // so
{"", {0xe2}}, // ta
{"", {0xe3}}, // chi
{"", {0xe4}}, // sokuon
{"", {0xe5}}, // tsu
{"", {0xe6}}, // te
{"", {0xe7}}, // to
{"", {0xe8}}, // na
{"", {0xe9}}, // ni
{"", {0xea}}, // nu
{"", {0xeb}}, // ne
{"", {0xec}}, // no
{"", {0xed}}, // ha
{"", {0xee}}, // hi
{"", {0xef}}, // hu
{"", {0xf0}}, // he
{"", {0xf1}}, // ho
{"", {0xf2}}, // ma
{"", {0xf3}}, // mi
{"", {0xf4}}, // mu
{"", {0xf5}}, // me
{"", {0xf6}}, // mo
{"", {0xf7}}, // youon ya
{"", {0xf8}}, // ya
{"", {0xf9}}, // youon yu
{"", {0xfa}}, // yu
{"", {0xfb}}, // youon yo
{"", {0xfc}}, // yo
{"", {0xfd}}, // ra
{"", {0xfe}}, // ri
{"", {0xff}}, // ru
// kanji 2
{"", {1, 0x01}}, // takara
{"", {1, 0x10}}, // ishi
{"", {1, 0x11}}, // aka
{"", {1, 0x12}}, // ato
{"", {1, 0x13}}, // kawa
{"", {1, 0x14}}, // ikusa
{"", {1, 0x15}}, // mura
{"", {1, 0x16}}, // tai
{"", {1, 0x17}}, // utena
{"", {1, 0x18}}, // osa
{"", {1, 0x19}}, // tori
{"", {1, 0x1a}}, // tei
{"", {1, 0x1b}}, // hora
{"", {1, 0x1c}}, // michi
{"", {1, 0x1d}}, // hatsu
{"", {1, 0x1e}}, // tobu
{"", {1, 0x1f}}, // fuku
{"", {1, 0xa0}}, // ike
{"", {1, 0xa1}}, // naka
{"", {1, 0xa2}}, // tou
{"", {1, 0xa3}}, // shima
{"", {1, 0xa4}}, // bu
{"", {1, 0xa5}}, // hou
{"", {1, 0xa6}}, // san
{"", {1, 0xa7}}, // kaerimiru
{"", {1, 0xa8}}, // chikara
{"", {1, 0xa9}}, // midori
{"", {1, 0xaa}}, // kishi
{"", {1, 0xab}}, // zou
{"", {1, 0xac}}, // tani
{"", {1, 0xad}}, // kokoro
{"", {1, 0xae}}, // mori
{"", {1, 0xaf}}, // mizu
{"", {1, 0xb0}}, // fune
{"", {1, 0xb1}}, // trademark
};
static std::vector<ReplaceInfo> g_replace_info_jak1 = {
// \" -> " (confusing)
{"\\\"", "\""},
// other
{"A~Y~-21H~-5Vº~Z", "Å"},
{"N~Y~-6Hº~Z~+10H", ""},
// tildes
{"N~Y~-22H~-4V<TIL>~Z", "Ñ"},
{"A~Y~-21H~-5V<TIL>~Z", "Ã"}, // custom
{"O~Y~-22H~-4V<TIL>~Z", "Õ"}, // custom
// acute accents
{"A~Y~-21H~-5V'~Z", "Á"},
{"E~Y~-22H~-5V'~Z", "É"},
{"I~Y~-19H~-5V'~Z", "Í"},
{"O~Y~-22H~-4V'~Z", "Ó"},
{"U~Y~-24H~-3V'~Z", "Ú"},
// circumflex
{"A~Y~-20H~-4V^~Z", "Â"}, // custom
{"E~Y~-20H~-5V^~Z", "Ê"},
{"I~Y~-19H~-5V^~Z", "Î"},
{"O~Y~-20H~-4V^~Z", "Ô"}, // custom
{"U~Y~-24H~-3V^~Z", "Û"},
// grave accents
{"A~Y~-21H~-5V`~Z", "À"},
{"E~Y~-22H~-5V`~Z", "È"},
{"I~Y~-19H~-5V`~Z", "Ì"},
{"O~Y~-22H~-4V`~Z", "Ò"}, // custom
{"U~Y~-24H~-3V`~Z", "Ù"},
// umlaut
{"A~Y~-21H~-5V¨~Z", "Ä"},
{"E~Y~-20H~-5V¨~Z", "Ë"},
{"I~Y~-19H~-5V¨~Z", "Ï"}, // custom
{"O~Y~-22H~-4V¨~Z", "Ö"},
{"O~Y~-22H~-3V¨~Z", "ö"}, // dumb
{"U~Y~-22H~-3V¨~Z", "Ü"},
// dakuten katakana
{"~Yウ~Z゛", ""},
{"~Yカ~Z゛", ""},
{"~Yキ~Z゛", ""},
{"~Yク~Z゛", ""},
{"~Yケ~Z゛", ""},
{"~Yコ~Z゛", ""},
{"~Yサ~Z゛", ""},
{"~Yシ~Z゛", ""},
{"~Yス~Z゛", ""},
{"~Yセ~Z゛", ""},
{"~Yソ~Z゛", ""},
{"~Yタ~Z゛", ""},
{"~Yチ~Z゛", ""},
{"~Yツ~Z゛", ""},
{"~Yテ~Z゛", ""},
{"~Yト~Z゛", ""},
{"~Yハ~Z゛", ""},
{"~Yヒ~Z゛", ""},
{"~Yフ~Z゛", ""},
{"~Yヘ~Z゛", ""},
{"~Yホ~Z゛", ""},
// handakuten katakana
{"~Yハ~Z゜", ""},
{"~Yヒ~Z゜", ""},
{"~Yフ~Z゜", ""},
{"~Yヘ~Z゜", ""},
{"~Yホ~Z゜", ""},
// dakuten hiragana
{"~Yか~Z゛", ""},
{"~Yき~Z゛", ""},
{"~Yく~Z゛", ""},
{"~Yけ~Z゛", ""},
{"~Yこ~Z゛", ""},
{"~Yさ~Z゛", ""},
{"~Yし~Z゛", ""},
{"~Yす~Z゛", ""},
{"~Yせ~Z゛", ""},
{"~Yそ~Z゛", ""},
{"~Yた~Z゛", ""},
{"~Yち~Z゛", ""},
{"~Yつ~Z゛", ""},
{"~Yて~Z゛", ""},
{"~Yと~Z゛", ""},
{"~Yは~Z゛", ""},
{"~Yひ~Z゛", ""},
{"~Yふ~Z゛", ""},
{"~Yへ~Z゛", ""},
{"~Yほ~Z゛", ""},
// handakuten hiragana
{"~Yは~Z゜", ""},
{"~Yひ~Z゜", ""},
{"~Yふ~Z゜", ""},
{"~Yへ~Z゜", ""},
{"~Yほ~Z゜", ""},
// japanese punctuation
{",~+8H", ""},
{"~+8H ", " "},
// (hack) special case kanji
{"~~", ""},
// playstation buttons
{"~Y~22L<~Z~Y~27L*~Z~Y~1L>~Z~Y~23L[~Z~+26H", "<PAD_X>"},
{"~Y~22L<~Z~Y~26L;~Z~Y~1L>~Z~Y~23L[~Z~+26H", "<PAD_TRIANGLE>"},
{"~Y~22L<~Z~Y~25L@~Z~Y~1L>~Z~Y~23L[~Z~+26H", "<PAD_CIRCLE>"},
{"~Y~22L<~Z~Y~24L#~Z~Y~1L>~Z~Y~23L[~Z~+26H", "<PAD_SQUARE>"}, // custom
};
GameTextFontBank g_font_bank_jak1(GameTextVersion::JAK1_V1,
&g_encode_info_jak1,
&g_replace_info_jak1,
&passthrus);
/*!
* ========================
* GAME TEXT FONT BANK LIST
* ========================
* The list of available font banks and a couple of helper functions.
*/
std::map<GameTextVersion, GameTextFontBank*> g_font_banks = {
{GameTextVersion::JAK1_V1, &g_font_bank_jak1}};
const GameTextFontBank* get_font_bank(GameTextVersion version) {
return g_font_banks.at(version);
}
bool font_bank_exists(GameTextVersion version) {
return g_font_banks.find(version) != g_font_banks.cend();
}