diff options
Diffstat (limited to 'core/string')
| -rw-r--r-- | core/string/fuzzy_search.cpp | 349 | ||||
| -rw-r--r-- | core/string/fuzzy_search.h | 101 | ||||
| -rw-r--r-- | core/string/node_path.cpp | 2 | ||||
| -rw-r--r-- | core/string/translation_domain.cpp | 5 | ||||
| -rw-r--r-- | core/string/translation_po.cpp | 10 | ||||
| -rw-r--r-- | core/string/translation_server.cpp | 139 | ||||
| -rw-r--r-- | core/string/translation_server.h | 20 | ||||
| -rw-r--r-- | core/string/ustring.cpp | 43 | ||||
| -rw-r--r-- | core/string/ustring.h | 5 |
9 files changed, 594 insertions, 80 deletions
diff --git a/core/string/fuzzy_search.cpp b/core/string/fuzzy_search.cpp new file mode 100644 index 0000000000..2fd0d3995e --- /dev/null +++ b/core/string/fuzzy_search.cpp @@ -0,0 +1,349 @@ +/**************************************************************************/ +/* fuzzy_search.cpp */ +/**************************************************************************/ +/* This file is part of: */ +/* GODOT ENGINE */ +/* https://godotengine.org */ +/**************************************************************************/ +/* Copyright (c) 2014-present Godot Engine contributors (see AUTHORS.md). */ +/* Copyright (c) 2007-2014 Juan Linietsky, Ariel Manzur. */ +/* */ +/* Permission is hereby granted, free of charge, to any person obtaining */ +/* a copy of this software and associated documentation files (the */ +/* "Software"), to deal in the Software without restriction, including */ +/* without limitation the rights to use, copy, modify, merge, publish, */ +/* distribute, sublicense, and/or sell copies of the Software, and to */ +/* permit persons to whom the Software is furnished to do so, subject to */ +/* the following conditions: */ +/* */ +/* The above copyright notice and this permission notice shall be */ +/* included in all copies or substantial portions of the Software. */ +/* */ +/* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, */ +/* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF */ +/* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. */ +/* IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY */ +/* CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, */ +/* TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE */ +/* SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. */ +/**************************************************************************/ + +#include "fuzzy_search.h" + +constexpr float cull_factor = 0.1f; +constexpr float cull_cutoff = 30.0f; +const String boundary_chars = "/\\-_."; + +static bool _is_valid_interval(const Vector2i &p_interval) { + // Empty intervals are represented as (-1, -1). + return p_interval.x >= 0 && p_interval.y >= p_interval.x; +} + +static Vector2i _extend_interval(const Vector2i &p_a, const Vector2i &p_b) { + if (!_is_valid_interval(p_a)) { + return p_b; + } + if (!_is_valid_interval(p_b)) { + return p_a; + } + return Vector2i(MIN(p_a.x, p_b.x), MAX(p_a.y, p_b.y)); +} + +static bool _is_word_boundary(const String &p_str, int p_index) { + if (p_index == -1 || p_index == p_str.size()) { + return true; + } + return boundary_chars.find_char(p_str[p_index]) != -1; +} + +bool FuzzySearchToken::try_exact_match(FuzzyTokenMatch &p_match, const String &p_target, int p_offset) const { + p_match.token_idx = idx; + p_match.token_length = string.length(); + int match_idx = p_target.find(string, p_offset); + if (match_idx == -1) { + return false; + } + p_match.add_substring(match_idx, string.length()); + return true; +} + +bool FuzzySearchToken::try_fuzzy_match(FuzzyTokenMatch &p_match, const String &p_target, int p_offset, int p_miss_budget) const { + p_match.token_idx = idx; + p_match.token_length = string.length(); + int run_start = -1; + int run_len = 0; + + // Search for the subsequence p_token in p_target starting from p_offset, recording each substring for + // later scoring and display. + for (int i = 0; i < string.length(); i++) { + int new_offset = p_target.find_char(string[i], p_offset); + if (new_offset < 0) { + p_miss_budget--; + if (p_miss_budget < 0) { + return false; + } + } else { + if (run_start == -1 || p_offset != new_offset) { + if (run_start != -1) { + p_match.add_substring(run_start, run_len); + } + run_start = new_offset; + run_len = 1; + } else { + run_len += 1; + } + p_offset = new_offset + 1; + } + } + + if (run_start != -1) { + p_match.add_substring(run_start, run_len); + } + + return true; +} + +void FuzzyTokenMatch::add_substring(int p_substring_start, int p_substring_length) { + substrings.append(Vector2i(p_substring_start, p_substring_length)); + matched_length += p_substring_length; + Vector2i substring_interval = { p_substring_start, p_substring_start + p_substring_length - 1 }; + interval = _extend_interval(interval, substring_interval); +} + +bool FuzzyTokenMatch::intersects(const Vector2i &p_other_interval) const { + if (!_is_valid_interval(interval) || !_is_valid_interval(p_other_interval)) { + return false; + } + return interval.y >= p_other_interval.x && interval.x <= p_other_interval.y; +} + +bool FuzzySearchResult::can_add_token_match(const FuzzyTokenMatch &p_match) const { + if (p_match.get_miss_count() > miss_budget) { + return false; + } + + if (p_match.intersects(match_interval)) { + if (token_matches.size() == 1) { + return false; + } + for (const FuzzyTokenMatch &existing_match : token_matches) { + if (existing_match.intersects(p_match.interval)) { + return false; + } + } + } + + return true; +} + +bool FuzzyTokenMatch::is_case_insensitive(const String &p_original, const String &p_adjusted) const { + for (const Vector2i &substr : substrings) { + const int end = substr.x + substr.y; + for (int i = substr.x; i < end; i++) { + if (p_original[i] != p_adjusted[i]) { + return true; + } + } + } + return false; +} + +void FuzzySearchResult::score_token_match(FuzzyTokenMatch &p_match, bool p_case_insensitive) const { + // This can always be tweaked more. The intuition is that exact matches should almost always + // be prioritized over broken up matches, and other criteria more or less act as tie breakers. + + p_match.score = -20 * p_match.get_miss_count() - (p_case_insensitive ? 3 : 0); + + for (const Vector2i &substring : p_match.substrings) { + // Score longer substrings higher than short substrings. + int substring_score = substring.y * substring.y; + // Score matches deeper in path higher than shallower matches + if (substring.x > dir_index) { + substring_score *= 2; + } + // Score matches on a word boundary higher than matches within a word + if (_is_word_boundary(target, substring.x - 1) || _is_word_boundary(target, substring.x + substring.y)) { + substring_score += 4; + } + // Score exact query matches higher than non-compact subsequence matches + if (substring.y == p_match.token_length) { + substring_score += 100; + } + p_match.score += substring_score; + } +} + +void FuzzySearchResult::maybe_apply_score_bonus() { + // This adds a small bonus to results which match tokens in the same order they appear in the query. + int *token_range_starts = (int *)alloca(sizeof(int) * token_matches.size()); + + for (const FuzzyTokenMatch &match : token_matches) { + token_range_starts[match.token_idx] = match.interval.x; + } + + int last = token_range_starts[0]; + for (int i = 1; i < token_matches.size(); i++) { + if (last > token_range_starts[i]) { + return; + } + last = token_range_starts[i]; + } + + score += 1; +} + +void FuzzySearchResult::add_token_match(const FuzzyTokenMatch &p_match) { + score += p_match.score; + match_interval = _extend_interval(match_interval, p_match.interval); + miss_budget -= p_match.get_miss_count(); + token_matches.append(p_match); +} + +void remove_low_scores(Vector<FuzzySearchResult> &p_results, float p_cull_score) { + // Removes all results with score < p_cull_score in-place. + int i = 0; + int j = p_results.size() - 1; + FuzzySearchResult *results = p_results.ptrw(); + + while (true) { + // Advances i to an element to remove and j to an element to keep. + while (j >= i && results[j].score < p_cull_score) { + j--; + } + while (i < j && results[i].score >= p_cull_score) { + i++; + } + if (i >= j) { + break; + } + results[i++] = results[j--]; + } + + p_results.resize(j + 1); +} + +void FuzzySearch::sort_and_filter(Vector<FuzzySearchResult> &p_results) const { + if (p_results.is_empty()) { + return; + } + + float avg_score = 0; + float max_score = 0; + + for (const FuzzySearchResult &result : p_results) { + avg_score += result.score; + max_score = MAX(max_score, result.score); + } + + // TODO: Tune scoring and culling here to display fewer subsequence soup matches when good matches + // are available. + avg_score /= p_results.size(); + float cull_score = MIN(cull_cutoff, Math::lerp(avg_score, max_score, cull_factor)); + remove_low_scores(p_results, cull_score); + + struct FuzzySearchResultComparator { + bool operator()(const FuzzySearchResult &p_lhs, const FuzzySearchResult &p_rhs) const { + // Sort on (score, length, alphanumeric) to ensure consistent ordering. + if (p_lhs.score == p_rhs.score) { + if (p_lhs.target.length() == p_rhs.target.length()) { + return p_lhs.target < p_rhs.target; + } + return p_lhs.target.length() < p_rhs.target.length(); + } + return p_lhs.score > p_rhs.score; + } + }; + + SortArray<FuzzySearchResult, FuzzySearchResultComparator> sorter; + + if (p_results.size() > max_results) { + sorter.partial_sort(0, p_results.size(), max_results, p_results.ptrw()); + p_results.resize(max_results); + } else { + sorter.sort(p_results.ptrw(), p_results.size()); + } +} + +void FuzzySearch::set_query(const String &p_query) { + tokens.clear(); + for (const String &string : p_query.split(" ", false)) { + tokens.append({ static_cast<int>(tokens.size()), string }); + } + + case_sensitive = !p_query.is_lowercase(); + + struct TokenComparator { + bool operator()(const FuzzySearchToken &A, const FuzzySearchToken &B) const { + if (A.string.length() == B.string.length()) { + return A.idx < B.idx; + } + return A.string.length() > B.string.length(); + } + }; + + // Prioritize matching longer tokens before shorter ones since match overlaps are not accepted. + tokens.sort_custom<TokenComparator>(); +} + +bool FuzzySearch::search(const String &p_target, FuzzySearchResult &p_result) const { + p_result.target = p_target; + p_result.dir_index = p_target.rfind_char('/'); + p_result.miss_budget = max_misses; + + String adjusted_target = case_sensitive ? p_target : p_target.to_lower(); + + // For each token, eagerly generate subsequences starting from index 0 and keep the best scoring one + // which does not conflict with prior token matches. This is not ensured to find the highest scoring + // combination of matches, or necessarily the highest scoring single subsequence, as it only considers + // eager subsequences for a given index, and likewise eagerly finds matches for each token in sequence. + for (const FuzzySearchToken &token : tokens) { + FuzzyTokenMatch best_match; + int offset = start_offset; + + while (true) { + FuzzyTokenMatch match; + if (allow_subsequences) { + if (!token.try_fuzzy_match(match, adjusted_target, offset, p_result.miss_budget)) { + break; + } + } else { + if (!token.try_exact_match(match, adjusted_target, offset)) { + break; + } + } + if (p_result.can_add_token_match(match)) { + p_result.score_token_match(match, match.is_case_insensitive(p_target, adjusted_target)); + if (best_match.token_idx == -1 || best_match.score < match.score) { + best_match = match; + } + } + if (_is_valid_interval(match.interval)) { + offset = match.interval.x + 1; + } else { + break; + } + } + + if (best_match.token_idx == -1) { + return false; + } + + p_result.add_token_match(best_match); + } + + p_result.maybe_apply_score_bonus(); + return true; +} + +void FuzzySearch::search_all(const PackedStringArray &p_targets, Vector<FuzzySearchResult> &p_results) const { + p_results.clear(); + + for (const String &target : p_targets) { + FuzzySearchResult result; + if (search(target, result)) { + p_results.append(result); + } + } + + sort_and_filter(p_results); +} diff --git a/core/string/fuzzy_search.h b/core/string/fuzzy_search.h new file mode 100644 index 0000000000..5d8ed813c7 --- /dev/null +++ b/core/string/fuzzy_search.h @@ -0,0 +1,101 @@ +/**************************************************************************/ +/* fuzzy_search.h */ +/**************************************************************************/ +/* This file is part of: */ +/* GODOT ENGINE */ +/* https://godotengine.org */ +/**************************************************************************/ +/* Copyright (c) 2014-present Godot Engine contributors (see AUTHORS.md). */ +/* Copyright (c) 2007-2014 Juan Linietsky, Ariel Manzur. */ +/* */ +/* Permission is hereby granted, free of charge, to any person obtaining */ +/* a copy of this software and associated documentation files (the */ +/* "Software"), to deal in the Software without restriction, including */ +/* without limitation the rights to use, copy, modify, merge, publish, */ +/* distribute, sublicense, and/or sell copies of the Software, and to */ +/* permit persons to whom the Software is furnished to do so, subject to */ +/* the following conditions: */ +/* */ +/* The above copyright notice and this permission notice shall be */ +/* included in all copies or substantial portions of the Software. */ +/* */ +/* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, */ +/* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF */ +/* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. */ +/* IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY */ +/* CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, */ +/* TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE */ +/* SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. */ +/**************************************************************************/ + +#ifndef FUZZY_SEARCH_H +#define FUZZY_SEARCH_H + +#include "core/variant/variant.h" + +class FuzzyTokenMatch; + +struct FuzzySearchToken { + int idx = -1; + String string; + + bool try_exact_match(FuzzyTokenMatch &p_match, const String &p_target, int p_offset) const; + bool try_fuzzy_match(FuzzyTokenMatch &p_match, const String &p_target, int p_offset, int p_miss_budget) const; +}; + +class FuzzyTokenMatch { + friend struct FuzzySearchToken; + friend class FuzzySearchResult; + friend class FuzzySearch; + + int matched_length = 0; + int token_length = 0; + int token_idx = -1; + Vector2i interval = Vector2i(-1, -1); // x and y are both inclusive indices. + + void add_substring(int p_substring_start, int p_substring_length); + bool intersects(const Vector2i &p_other_interval) const; + bool is_case_insensitive(const String &p_original, const String &p_adjusted) const; + int get_miss_count() const { return token_length - matched_length; } + +public: + int score = 0; + Vector<Vector2i> substrings; // x is start index, y is length. +}; + +class FuzzySearchResult { + friend class FuzzySearch; + + int miss_budget = 0; + Vector2i match_interval = Vector2i(-1, -1); + + bool can_add_token_match(const FuzzyTokenMatch &p_match) const; + void score_token_match(FuzzyTokenMatch &p_match, bool p_case_insensitive) const; + void add_token_match(const FuzzyTokenMatch &p_match); + void maybe_apply_score_bonus(); + +public: + String target; + int score = 0; + int dir_index = -1; + Vector<FuzzyTokenMatch> token_matches; +}; + +class FuzzySearch { + Vector<FuzzySearchToken> tokens; + + void sort_and_filter(Vector<FuzzySearchResult> &p_results) const; + +public: + int start_offset = 0; + bool case_sensitive = false; + int max_results = 100; + int max_misses = 2; + bool allow_subsequences = true; + + void set_query(const String &p_query); + bool search(const String &p_target, FuzzySearchResult &p_result) const; + void search_all(const PackedStringArray &p_targets, Vector<FuzzySearchResult> &p_results) const; +}; + +#endif // FUZZY_SEARCH_H diff --git a/core/string/node_path.cpp b/core/string/node_path.cpp index fdc72bc8dc..3faf3bb0c5 100644 --- a/core/string/node_path.cpp +++ b/core/string/node_path.cpp @@ -420,7 +420,7 @@ NodePath::NodePath(const String &p_path) { continue; // Allow end-of-path : } - ERR_FAIL_MSG("Invalid NodePath '" + p_path + "'."); + ERR_FAIL_MSG(vformat("Invalid NodePath '%s'.", p_path)); } subpath.push_back(str); diff --git a/core/string/translation_domain.cpp b/core/string/translation_domain.cpp index cf6689efff..1ff8dcd752 100644 --- a/core/string/translation_domain.cpp +++ b/core/string/translation_domain.cpp @@ -247,7 +247,10 @@ PackedStringArray TranslationDomain::get_loaded_locales() const { PackedStringArray locales; for (const Ref<Translation> &E : translations) { ERR_CONTINUE(E.is_null()); - locales.push_back(E->get_locale()); + const String &locale = E->get_locale(); + if (!locales.has(locale)) { + locales.push_back(locale); + } } return locales; } diff --git a/core/string/translation_po.cpp b/core/string/translation_po.cpp index 8e275505b0..da79e472e7 100644 --- a/core/string/translation_po.cpp +++ b/core/string/translation_po.cpp @@ -246,7 +246,7 @@ void TranslationPO::add_message(const StringName &p_src_text, const StringName & HashMap<StringName, Vector<StringName>> &map_id_str = translation_map[p_context]; if (map_id_str.has(p_src_text)) { - WARN_PRINT("Double translations for \"" + String(p_src_text) + "\" under the same context \"" + String(p_context) + "\" for locale \"" + get_locale() + "\".\nThere should only be one unique translation for a given string under the same context."); + WARN_PRINT(vformat("Double translations for \"%s\" under the same context \"%s\" for locale \"%s\".\nThere should only be one unique translation for a given string under the same context.", String(p_src_text), String(p_context), get_locale())); map_id_str[p_src_text].set(0, p_xlated_text); } else { map_id_str[p_src_text].push_back(p_xlated_text); @@ -254,12 +254,12 @@ void TranslationPO::add_message(const StringName &p_src_text, const StringName & } void TranslationPO::add_plural_message(const StringName &p_src_text, const Vector<String> &p_plural_xlated_texts, const StringName &p_context) { - ERR_FAIL_COND_MSG(p_plural_xlated_texts.size() != plural_forms, "Trying to add plural texts that don't match the required number of plural forms for locale \"" + get_locale() + "\""); + ERR_FAIL_COND_MSG(p_plural_xlated_texts.size() != plural_forms, vformat("Trying to add plural texts that don't match the required number of plural forms for locale \"%s\".", get_locale())); HashMap<StringName, Vector<StringName>> &map_id_str = translation_map[p_context]; if (map_id_str.has(p_src_text)) { - WARN_PRINT("Double translations for \"" + p_src_text + "\" under the same context \"" + p_context + "\" for locale " + get_locale() + ".\nThere should only be one unique translation for a given string under the same context."); + WARN_PRINT(vformat("Double translations for \"%s\" under the same context \"%s\" for locale %s.\nThere should only be one unique translation for a given string under the same context.", p_src_text, p_context, get_locale())); map_id_str[p_src_text].clear(); } @@ -280,7 +280,7 @@ StringName TranslationPO::get_message(const StringName &p_src_text, const String if (!translation_map.has(p_context) || !translation_map[p_context].has(p_src_text)) { return StringName(); } - ERR_FAIL_COND_V_MSG(translation_map[p_context][p_src_text].is_empty(), StringName(), "Source text \"" + String(p_src_text) + "\" is registered but doesn't have a translation. Please report this bug."); + ERR_FAIL_COND_V_MSG(translation_map[p_context][p_src_text].is_empty(), StringName(), vformat("Source text \"%s\" is registered but doesn't have a translation. Please report this bug.", String(p_src_text))); return translation_map[p_context][p_src_text][0]; } @@ -296,7 +296,7 @@ StringName TranslationPO::get_plural_message(const StringName &p_src_text, const if (!translation_map.has(p_context) || !translation_map[p_context].has(p_src_text)) { return StringName(); } - ERR_FAIL_COND_V_MSG(translation_map[p_context][p_src_text].is_empty(), StringName(), "Source text \"" + String(p_src_text) + "\" is registered but doesn't have a translation. Please report this bug."); + ERR_FAIL_COND_V_MSG(translation_map[p_context][p_src_text].is_empty(), StringName(), vformat("Source text \"%s\" is registered but doesn't have a translation. Please report this bug.", String(p_src_text))); int plural_index = _get_plural_index(p_n); ERR_FAIL_COND_V_MSG(plural_index < 0 || translation_map[p_context][p_src_text].size() < plural_index + 1, StringName(), "Plural index returned or number of plural translations is not valid. Please report this bug."); diff --git a/core/string/translation_server.cpp b/core/string/translation_server.cpp index 92b473b61f..31c221dad7 100644 --- a/core/string/translation_server.cpp +++ b/core/string/translation_server.cpp @@ -118,36 +118,45 @@ void TranslationServer::init_locale_info() { } } -String TranslationServer::standardize_locale(const String &p_locale) const { - return _standardize_locale(p_locale, false); +TranslationServer::Locale::operator String() const { + String out = language; + if (!script.is_empty()) { + out = out + "_" + script; + } + if (!country.is_empty()) { + out = out + "_" + country; + } + if (!variant.is_empty()) { + out = out + "_" + variant; + } + return out; } -String TranslationServer::_standardize_locale(const String &p_locale, bool p_add_defaults) const { +TranslationServer::Locale::Locale(const TranslationServer &p_server, const String &p_locale, bool p_add_defaults) { // Replaces '-' with '_' for macOS style locales. String univ_locale = p_locale.replace("-", "_"); // Extract locale elements. - String lang_name, script_name, country_name, variant_name; Vector<String> locale_elements = univ_locale.get_slice("@", 0).split("_"); - lang_name = locale_elements[0]; + language = locale_elements[0]; if (locale_elements.size() >= 2) { if (locale_elements[1].length() == 4 && is_ascii_upper_case(locale_elements[1][0]) && is_ascii_lower_case(locale_elements[1][1]) && is_ascii_lower_case(locale_elements[1][2]) && is_ascii_lower_case(locale_elements[1][3])) { - script_name = locale_elements[1]; + script = locale_elements[1]; } if (locale_elements[1].length() == 2 && is_ascii_upper_case(locale_elements[1][0]) && is_ascii_upper_case(locale_elements[1][1])) { - country_name = locale_elements[1]; + country = locale_elements[1]; } } if (locale_elements.size() >= 3) { if (locale_elements[2].length() == 2 && is_ascii_upper_case(locale_elements[2][0]) && is_ascii_upper_case(locale_elements[2][1])) { - country_name = locale_elements[2]; - } else if (variant_map.has(locale_elements[2].to_lower()) && variant_map[locale_elements[2].to_lower()] == lang_name) { - variant_name = locale_elements[2].to_lower(); + country = locale_elements[2]; + } else if (p_server.variant_map.has(locale_elements[2].to_lower()) && p_server.variant_map[locale_elements[2].to_lower()] == language) { + variant = locale_elements[2].to_lower(); } } if (locale_elements.size() >= 4) { - if (variant_map.has(locale_elements[3].to_lower()) && variant_map[locale_elements[3].to_lower()] == lang_name) { - variant_name = locale_elements[3].to_lower(); + if (p_server.variant_map.has(locale_elements[3].to_lower()) && p_server.variant_map[locale_elements[3].to_lower()] == language) { + variant = locale_elements[3].to_lower(); } } @@ -155,71 +164,62 @@ String TranslationServer::_standardize_locale(const String &p_locale, bool p_add Vector<String> script_extra = univ_locale.get_slice("@", 1).split(";"); for (int i = 0; i < script_extra.size(); i++) { if (script_extra[i].to_lower() == "cyrillic") { - script_name = "Cyrl"; + script = "Cyrl"; break; } else if (script_extra[i].to_lower() == "latin") { - script_name = "Latn"; + script = "Latn"; break; } else if (script_extra[i].to_lower() == "devanagari") { - script_name = "Deva"; + script = "Deva"; break; - } else if (variant_map.has(script_extra[i].to_lower()) && variant_map[script_extra[i].to_lower()] == lang_name) { - variant_name = script_extra[i].to_lower(); + } else if (p_server.variant_map.has(script_extra[i].to_lower()) && p_server.variant_map[script_extra[i].to_lower()] == language) { + variant = script_extra[i].to_lower(); } } // Handles known non-ISO language names used e.g. on Windows. - if (locale_rename_map.has(lang_name)) { - lang_name = locale_rename_map[lang_name]; + if (p_server.locale_rename_map.has(language)) { + language = p_server.locale_rename_map[language]; } // Handle country renames. - if (country_rename_map.has(country_name)) { - country_name = country_rename_map[country_name]; + if (p_server.country_rename_map.has(country)) { + country = p_server.country_rename_map[country]; } // Remove unsupported script codes. - if (!script_map.has(script_name)) { - script_name = ""; + if (!p_server.script_map.has(script)) { + script = ""; } // Add script code base on language and country codes for some ambiguous cases. if (p_add_defaults) { - if (script_name.is_empty()) { - for (int i = 0; i < locale_script_info.size(); i++) { - const LocaleScriptInfo &info = locale_script_info[i]; - if (info.name == lang_name) { - if (country_name.is_empty() || info.supported_countries.has(country_name)) { - script_name = info.script; + if (script.is_empty()) { + for (int i = 0; i < p_server.locale_script_info.size(); i++) { + const LocaleScriptInfo &info = p_server.locale_script_info[i]; + if (info.name == language) { + if (country.is_empty() || info.supported_countries.has(country)) { + script = info.script; break; } } } } - if (!script_name.is_empty() && country_name.is_empty()) { + if (!script.is_empty() && country.is_empty()) { // Add conntry code based on script for some ambiguous cases. - for (int i = 0; i < locale_script_info.size(); i++) { - const LocaleScriptInfo &info = locale_script_info[i]; - if (info.name == lang_name && info.script == script_name) { - country_name = info.default_country; + for (int i = 0; i < p_server.locale_script_info.size(); i++) { + const LocaleScriptInfo &info = p_server.locale_script_info[i]; + if (info.name == language && info.script == script) { + country = info.default_country; break; } } } } +} - // Combine results. - String out = lang_name; - if (!script_name.is_empty()) { - out = out + "_" + script_name; - } - if (!country_name.is_empty()) { - out = out + "_" + country_name; - } - if (!variant_name.is_empty()) { - out = out + "_" + variant_name; - } - return out; +String TranslationServer::standardize_locale(const String &p_locale) const { + return Locale(*this, p_locale, false).operator String(); } int TranslationServer::compare_locales(const String &p_locale_a, const String &p_locale_b) const { @@ -234,8 +234,8 @@ int TranslationServer::compare_locales(const String &p_locale_a, const String &p return *cached_result; } - String locale_a = _standardize_locale(p_locale_a, true); - String locale_b = _standardize_locale(p_locale_b, true); + Locale locale_a = Locale(*this, p_locale_a, true); + Locale locale_b = Locale(*this, p_locale_b, true); if (locale_a == locale_b) { // Exact match. @@ -243,26 +243,41 @@ int TranslationServer::compare_locales(const String &p_locale_a, const String &p return 10; } - Vector<String> locale_a_elements = locale_a.split("_"); - Vector<String> locale_b_elements = locale_b.split("_"); - if (locale_a_elements[0] != locale_b_elements[0]) { + if (locale_a.language != locale_b.language) { // No match. locale_compare_cache.insert(cache_key, 0); return 0; } - // Matching language, both locales have extra parts. - // Return number of matching elements. - int matching_elements = 1; - for (int i = 1; i < locale_a_elements.size(); i++) { - for (int j = 1; j < locale_b_elements.size(); j++) { - if (locale_a_elements[i] == locale_b_elements[j]) { - matching_elements++; - } + // Matching language, both locales have extra parts. Compare the + // remaining elements. If both elements are non-empty, check the + // match to increase or decrease the score. If either element or + // both are empty, leave the score as is. + int score = 5; + if (!locale_a.script.is_empty() && !locale_b.script.is_empty()) { + if (locale_a.script == locale_b.script) { + score++; + } else { + score--; } } - locale_compare_cache.insert(cache_key, matching_elements); - return matching_elements; + if (!locale_a.country.is_empty() && !locale_b.country.is_empty()) { + if (locale_a.country == locale_b.country) { + score++; + } else { + score--; + } + } + if (!locale_a.variant.is_empty() && !locale_b.variant.is_empty()) { + if (locale_a.variant == locale_b.variant) { + score++; + } else { + score--; + } + } + + locale_compare_cache.insert(cache_key, score); + return score; } String TranslationServer::get_locale_name(const String &p_locale) const { @@ -396,8 +411,6 @@ StringName TranslationServer::translate_plural(const StringName &p_message, cons return main_domain->translate_plural(p_message, p_message_plural, p_n, p_context); } -TranslationServer *TranslationServer::singleton = nullptr; - bool TranslationServer::_load_translations(const String &p_from) { if (ProjectSettings::get_singleton()->has_setting(p_from)) { const Vector<String> &translation_names = GLOBAL_GET(p_from); diff --git a/core/string/translation_server.h b/core/string/translation_server.h index 2438349a69..bc59c34a38 100644 --- a/core/string/translation_server.h +++ b/core/string/translation_server.h @@ -50,7 +50,7 @@ class TranslationServer : public Object { bool enabled = true; - static TranslationServer *singleton; + static inline TranslationServer *singleton = nullptr; bool _load_translations(const String &p_from); String _standardize_locale(const String &p_locale, bool p_add_defaults) const; @@ -64,6 +64,24 @@ class TranslationServer : public Object { }; static Vector<LocaleScriptInfo> locale_script_info; + struct Locale { + String language; + String script; + String country; + String variant; + + bool operator==(const Locale &p_locale) const { + return (p_locale.language == language) && + (p_locale.script == script) && + (p_locale.country == country) && + (p_locale.variant == variant); + } + + operator String() const; + + Locale(const TranslationServer &p_server, const String &p_locale, bool p_add_defaults); + }; + static HashMap<String, String> language_map; static HashMap<String, String> script_map; static HashMap<String, String> locale_rename_map; diff --git a/core/string/ustring.cpp b/core/string/ustring.cpp index 28319fc643..521dfe0b8c 100644 --- a/core/string/ustring.cpp +++ b/core/string/ustring.cpp @@ -1819,7 +1819,7 @@ String String::num(double p_num, int p_decimals) { #endif buf[324] = 0; - //destroy trailing zeroes + // Destroy trailing zeroes, except one after period. { bool period = false; int z = 0; @@ -1836,7 +1836,7 @@ String String::num(double p_num, int p_decimals) { if (buf[z] == '0') { buf[z] = 0; } else if (buf[z] == '.') { - buf[z] = 0; + buf[z + 1] = '0'; break; } else { break; @@ -1929,14 +1929,28 @@ String String::num_real(double p_num, bool p_trailing) { return num_int64((int64_t)p_num); } } -#ifdef REAL_T_IS_DOUBLE int decimals = 14; -#else + // We want to align the digits to the above sane default, so we only need + // to subtract log10 for numbers with a positive power of ten magnitude. + const double abs_num = Math::abs(p_num); + if (abs_num > 10) { + decimals -= (int)floor(log10(abs_num)); + } + return num(p_num, decimals); +} + +String String::num_real(float p_num, bool p_trailing) { + if (p_num == (float)(int64_t)p_num) { + if (p_trailing) { + return num_int64((int64_t)p_num) + ".0"; + } else { + return num_int64((int64_t)p_num); + } + } int decimals = 6; -#endif // We want to align the digits to the above sane default, so we only need // to subtract log10 for numbers with a positive power of ten magnitude. - double abs_num = Math::abs(p_num); + const float abs_num = Math::abs(p_num); if (abs_num > 10) { decimals -= (int)floor(log10(abs_num)); } @@ -3373,7 +3387,7 @@ int String::find(const char *p_str, int p_from) const { return -1; } -int String::find_char(const char32_t &p_char, int p_from) const { +int String::find_char(char32_t p_char, int p_from) const { return _cowdata.find(p_char, p_from); } @@ -3610,6 +3624,10 @@ int String::rfind(const char *p_str, int p_from) const { return -1; } +int String::rfind_char(char32_t p_char, int p_from) const { + return _cowdata.rfind(p_char, p_from); +} + int String::rfindn(const String &p_str, int p_from) const { // establish a limit int limit = length() - p_str.length(); @@ -3823,6 +3841,15 @@ bool String::is_quoted() const { return is_enclosed_in("\"") || is_enclosed_in("'"); } +bool String::is_lowercase() const { + for (const char32_t *str = &operator[](0); *str; str++) { + if (is_unicode_upper_case(*str)) { + return false; + } + } + return true; +} + int String::_count(const String &p_string, int p_from, int p_to, bool p_case_insensitive) const { if (p_string.is_empty()) { return 0; @@ -4616,7 +4643,7 @@ String String::humanize_size(uint64_t p_size) { } if (magnitude == 0) { - return String::num(p_size) + " " + RTR("B"); + return String::num_uint64(p_size) + " " + RTR("B"); } else { String suffix; switch (magnitude) { diff --git a/core/string/ustring.h b/core/string/ustring.h index 11c0f74062..d6e563223a 100644 --- a/core/string/ustring.h +++ b/core/string/ustring.h @@ -287,11 +287,12 @@ public: String substr(int p_from, int p_chars = -1) const; int find(const String &p_str, int p_from = 0) const; ///< return <0 if failed int find(const char *p_str, int p_from = 0) const; ///< return <0 if failed - int find_char(const char32_t &p_char, int p_from = 0) const; ///< return <0 if failed + int find_char(char32_t p_char, int p_from = 0) const; ///< return <0 if failed int findn(const String &p_str, int p_from = 0) const; ///< return <0 if failed, case insensitive int findn(const char *p_str, int p_from = 0) const; ///< return <0 if failed int rfind(const String &p_str, int p_from = -1) const; ///< return <0 if failed int rfind(const char *p_str, int p_from = -1) const; ///< return <0 if failed + int rfind_char(char32_t p_char, int p_from = -1) const; ///< return <0 if failed int rfindn(const String &p_str, int p_from = -1) const; ///< return <0 if failed, case insensitive int rfindn(const char *p_str, int p_from = -1) const; ///< return <0 if failed int findmk(const Vector<String> &p_keys, int p_from = 0, int *r_key = nullptr) const; ///< return <0 if failed @@ -305,6 +306,7 @@ public: bool is_subsequence_of(const String &p_string) const; bool is_subsequence_ofn(const String &p_string) const; bool is_quoted() const; + bool is_lowercase() const; Vector<String> bigrams() const; float similarity(const String &p_string) const; String format(const Variant &values, const String &placeholder = "{_}") const; @@ -332,6 +334,7 @@ public: static String num(double p_num, int p_decimals = -1); static String num_scientific(double p_num); static String num_real(double p_num, bool p_trailing = true); + static String num_real(float p_num, bool p_trailing = true); static String num_int64(int64_t p_num, int base = 10, bool capitalize_hex = false); static String num_uint64(uint64_t p_num, int base = 10, bool capitalize_hex = false); static String chr(char32_t p_char); |