1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
|
/* $Id$ */
/*
* This file is part of OpenTTD.
* OpenTTD is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, version 2.
* OpenTTD is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
* See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with OpenTTD. If not, see <http://www.gnu.org/licenses/>.
*/
/**
* @file newgrf_text.cpp
* Implementation of Action 04 "universal holder" structure and functions.
* This file implements a linked-lists of strings,
* holding everything that the newgrf action 04 will send over to OpenTTD.
* One of the biggest problems is that Dynamic lang Array uses ISO codes
* as way to identifying current user lang, while newgrf uses bit shift codes
* not related to ISO. So equivalence functionnality had to be set.
*/
#include "stdafx.h"
#include "newgrf.h"
#include "strings_func.h"
#include "newgrf_storage.h"
#include "newgrf_text.h"
#include "newgrf_cargo.h"
#include "string_func.h"
#include "date_type.h"
#include "debug.h"
#include "core/alloc_type.hpp"
#include "core/smallmap_type.hpp"
#include "language.h"
#include "table/strings.h"
#include "table/control_codes.h"
#include "safeguards.h"
#define GRFTAB 28
#define TABSIZE 11
/**
* Explains the newgrf shift bit positioning.
* the grf base will not be used in order to find the string, but rather for
* jumping from standard langID scheme to the new one.
*/
enum GRFBaseLanguages {
GRFLB_AMERICAN = 0x01,
GRFLB_ENGLISH = 0x02,
GRFLB_GERMAN = 0x04,
GRFLB_FRENCH = 0x08,
GRFLB_SPANISH = 0x10,
GRFLB_GENERIC = 0x80,
};
enum GRFExtendedLanguages {
GRFLX_AMERICAN = 0x00,
GRFLX_ENGLISH = 0x01,
GRFLX_GERMAN = 0x02,
GRFLX_FRENCH = 0x03,
GRFLX_SPANISH = 0x04,
GRFLX_UNSPECIFIED = 0x7F,
};
/**
* Element of the linked list.
* Each of those elements represent the string,
* but according to a different lang.
*/
struct GRFText {
public:
/**
* Allocate, and assign a new GRFText with the given text.
* As these strings can have string terminations in them, e.g.
* due to "choice lists" we (sometimes) cannot rely on detecting
* the length by means of strlen. Also, if the length of already
* known not scanning the whole string is more efficient.
* @param langid The language of the text.
* @param text The text to store in the new GRFText.
* @param len The length of the text.
*/
static GRFText *New(byte langid, const char *text, size_t len)
{
return new (len) GRFText(langid, text, len);
}
/**
* Create a copy of this GRFText.
* @param orig the grftext to copy.
* @return an exact copy of the given text.
*/
static GRFText *Copy(GRFText *orig)
{
return GRFText::New(orig->langid, orig->text, orig->len);
}
/**
* Helper allocation function to disallow something.
* Don't allow simple 'news'; they wouldn't have enough memory.
* @param size the amount of space not to allocate.
*/
void *operator new(size_t size)
{
NOT_REACHED();
}
/**
* Free the memory we allocated.
* @param p memory to free.
*/
void operator delete(void *p)
{
free(p);
}
private:
/**
* Actually construct the GRFText.
* @param langid_ The language of the text.
* @param text_ The text to store in this GRFText.
* @param len_ The length of the text to store.
*/
GRFText(byte langid_, const char *text_, size_t len_) : next(NULL), len(len_), langid(langid_)
{
/* We need to use memcpy instead of strcpy due to
* the possibility of "choice lists" and therefore
* intermediate string terminators. */
memcpy(this->text, text_, len);
}
/**
* Allocate memory for this class.
* @param size the size of the instance
* @param extra the extra memory for the text
* @return the requested amount of memory for both the instance and the text
*/
void *operator new(size_t size, size_t extra)
{
return MallocT<byte>(size + extra);
}
public:
GRFText *next; ///< The next GRFText in this chain.
size_t len; ///< The length of the stored string, used for copying.
byte langid; ///< The language associated with this GRFText.
char text[]; ///< The actual (translated) text.
};
/**
* Holder of the above structure.
* Putting both grfid and stringid together allows us to avoid duplicates,
* since it is NOT SUPPOSED to happen.
*/
struct GRFTextEntry {
uint32 grfid;
uint16 stringid;
StringID def_string;
GRFText *textholder;
};
static uint _num_grf_texts = 0;
static GRFTextEntry _grf_text[(1 << TABSIZE) * 3];
static byte _currentLangID = GRFLX_ENGLISH; ///< by default, english is used.
/**
* Get the mapping from the NewGRF supplied ID to OpenTTD's internal ID.
* @param newgrf_id The NewGRF ID to map.
* @param gender Whether to map genders or cases.
* @return The, to OpenTTD's internal ID, mapped index, or -1 if there is no mapping.
*/
int LanguageMap::GetMapping(int newgrf_id, bool gender) const
{
const SmallVector<Mapping, 1> &map = gender ? this->gender_map : this->case_map;
for (const Mapping *m = map.Begin(); m != map.End(); m++) {
if (m->newgrf_id == newgrf_id) return m->openttd_id;
}
return -1;
}
/**
* Get the mapping from OpenTTD's internal ID to the NewGRF supplied ID.
* @param openttd_id The OpenTTD ID to map.
* @param gender Whether to map genders or cases.
* @return The, to the NewGRF supplied ID, mapped index, or -1 if there is no mapping.
*/
int LanguageMap::GetReverseMapping(int openttd_id, bool gender) const
{
const SmallVector<Mapping, 1> &map = gender ? this->gender_map : this->case_map;
for (const Mapping *m = map.Begin(); m != map.End(); m++) {
if (m->openttd_id == openttd_id) return m->newgrf_id;
}
return -1;
}
/** Helper structure for mapping choice lists. */
struct UnmappedChoiceList : ZeroedMemoryAllocator {
/** Clean everything up. */
~UnmappedChoiceList()
{
for (SmallPair<byte, char *> *p = this->strings.Begin(); p < this->strings.End(); p++) {
free(p->second);
}
}
/**
* Initialise the mapping.
* @param type The type of mapping.
* @param old_d The old begin of the string, i.e. from where to start writing again.
* @param offset The offset to get the plural/gender from.
*/
UnmappedChoiceList(StringControlCode type, char *old_d, int offset) :
type(type), old_d(old_d), offset(offset)
{
}
StringControlCode type; ///< The type of choice list.
char *old_d; ///< The old/original location of the "d" local variable.
int offset; ///< The offset for the plural/gender form.
/** Mapping of NewGRF supplied ID to the different strings in the choice list. */
SmallMap<byte, char *> strings;
/**
* Flush this choice list into the old d variable.
* @param lm The current language mapping.
* @return The new location of the output string.
*/
char *Flush(const LanguageMap *lm)
{
if (!this->strings.Contains(0)) {
/* In case of a (broken) NewGRF without a default,
* assume an empty string. */
grfmsg(1, "choice list misses default value");
this->strings[0] = stredup("");
}
char *d = old_d;
if (lm == NULL) {
/* In case there is no mapping, just ignore everything but the default.
* A probable cause for this happening is when the language file has
* been removed by the user and as such no mapping could be made. */
size_t len = strlen(this->strings[0]);
memcpy(d, this->strings[0], len);
return d + len;
}
d += Utf8Encode(d, this->type);
if (this->type == SCC_SWITCH_CASE) {
/*
* Format for case switch:
* <NUM CASES> <CASE1> <LEN1> <STRING1> <CASE2> <LEN2> <STRING2> <CASE3> <LEN3> <STRING3> <STRINGDEFAULT>
* Each LEN is printed using 2 bytes in big endian order.
*/
/* "<NUM CASES>" */
int count = 0;
for (uint8 i = 0; i < _current_language->num_cases; i++) {
/* Count the ones we have a mapped string for. */
if (this->strings.Contains(lm->GetReverseMapping(i, false))) count++;
}
*d++ = count;
for (uint8 i = 0; i < _current_language->num_cases; i++) {
/* Resolve the string we're looking for. */
int idx = lm->GetReverseMapping(i, false);
if (!this->strings.Contains(idx)) continue;
char *str = this->strings[idx];
/* "<CASEn>" */
*d++ = i + 1;
/* "<LENn>" */
size_t len = strlen(str) + 1;
*d++ = GB(len, 8, 8);
*d++ = GB(len, 0, 8);
/* "<STRINGn>" */
memcpy(d, str, len);
d += len;
}
/* "<STRINGDEFAULT>" */
size_t len = strlen(this->strings[0]) + 1;
memcpy(d, this->strings[0], len);
d += len;
} else {
if (this->type == SCC_PLURAL_LIST) {
*d++ = lm->plural_form;
}
/*
* Format for choice list:
* <OFFSET> <NUM CHOICES> <LENs> <STRINGs>
*/
/* "<OFFSET>" */
*d++ = this->offset - 0x80;
/* "<NUM CHOICES>" */
int count = (this->type == SCC_GENDER_LIST ? _current_language->num_genders : LANGUAGE_MAX_PLURAL_FORMS);
*d++ = count;
/* "<LENs>" */
for (int i = 0; i < count; i++) {
int idx = (this->type == SCC_GENDER_LIST ? lm->GetReverseMapping(i, true) : i + 1);
const char *str = this->strings[this->strings.Contains(idx) ? idx : 0];
size_t len = strlen(str) + 1;
if (len > 0xFF) grfmsg(1, "choice list string is too long");
*d++ = GB(len, 0, 8);
}
/* "<STRINGs>" */
for (int i = 0; i < count; i++) {
int idx = (this->type == SCC_GENDER_LIST ? lm->GetReverseMapping(i, true) : i + 1);
const char *str = this->strings[this->strings.Contains(idx) ? idx : 0];
/* Limit the length of the string we copy to 0xFE. The length is written above
* as a byte and we need room for the final '\0'. */
size_t len = min<size_t>(0xFE, strlen(str));
memcpy(d, str, len);
d += len;
*d++ = '\0';
}
}
return d;
}
};
/**
* Translate TTDPatch string codes into something OpenTTD can handle (better).
* @param grfid The (NewGRF) ID associated with this string
* @param language_id The (NewGRF) language ID associated with this string.
* @param allow_newlines Whether newlines are allowed in the string or not.
* @param str The string to translate.
* @param [out] olen The length of the final string.
* @param byte80 The control code to use as replacement for the 0x80-value.
* @return The translated string.
*/
char *TranslateTTDPatchCodes(uint32 grfid, uint8 language_id, bool allow_newlines, const char *str, int *olen, StringControlCode byte80)
{
char *tmp = MallocT<char>(strlen(str) * 10 + 1); // Allocate space to allow for expansion
char *d = tmp;
bool unicode = false;
WChar c;
size_t len = Utf8Decode(&c, str);
/* Helper variable for a possible (string) mapping. */
UnmappedChoiceList *mapping = NULL;
if (c == NFO_UTF8_IDENTIFIER) {
unicode = true;
str += len;
}
for (;;) {
if (unicode && Utf8EncodedCharLen(*str) != 0) {
c = Utf8Consume(&str);
/* 'Magic' range of control codes. */
if (GB(c, 8, 8) == 0xE0) {
c = GB(c, 0, 8);
} else if (c >= 0x20) {
if (!IsValidChar(c, CS_ALPHANUMERAL)) c = '?';
d += Utf8Encode(d, c);
continue;
}
} else {
c = (byte)*str++;
}
if (c == '\0') break;
switch (c) {
case 0x01:
if (str[0] == '\0') goto string_end;
d += Utf8Encode(d, ' ');
str++;
break;
case 0x0A: break;
case 0x0D:
if (allow_newlines) {
*d++ = 0x0A;
} else {
grfmsg(1, "Detected newline in string that does not allow one");
}
break;
case 0x0E: d += Utf8Encode(d, SCC_TINYFONT); break;
case 0x0F: d += Utf8Encode(d, SCC_BIGFONT); break;
case 0x1F:
if (str[0] == '\0' || str[1] == '\0') goto string_end;
d += Utf8Encode(d, ' ');
str += 2;
break;
case 0x7B:
case 0x7C:
case 0x7D:
case 0x7E:
case 0x7F: d += Utf8Encode(d, SCC_NEWGRF_PRINT_DWORD_SIGNED + c - 0x7B); break;
case 0x80: d += Utf8Encode(d, byte80); break;
case 0x81: {
if (str[0] == '\0' || str[1] == '\0') goto string_end;
StringID string;
string = ((uint8)*str++);
string |= ((uint8)*str++) << 8;
d += Utf8Encode(d, SCC_NEWGRF_STRINL);
d += Utf8Encode(d, MapGRFStringID(grfid, string));
break;
}
case 0x82:
case 0x83:
case 0x84: d += Utf8Encode(d, SCC_NEWGRF_PRINT_WORD_DATE_LONG + c - 0x82); break;
case 0x85: d += Utf8Encode(d, SCC_NEWGRF_DISCARD_WORD); break;
case 0x86: d += Utf8Encode(d, SCC_NEWGRF_ROTATE_TOP_4_WORDS); break;
case 0x87: d += Utf8Encode(d, SCC_NEWGRF_PRINT_WORD_VOLUME_LONG); break;
case 0x88: d += Utf8Encode(d, SCC_BLUE); break;
case 0x89: d += Utf8Encode(d, SCC_SILVER); break;
case 0x8A: d += Utf8Encode(d, SCC_GOLD); break;
case 0x8B: d += Utf8Encode(d, SCC_RED); break;
case 0x8C: d += Utf8Encode(d, SCC_PURPLE); break;
case 0x8D: d += Utf8Encode(d, SCC_LTBROWN); break;
case 0x8E: d += Utf8Encode(d, SCC_ORANGE); break;
case 0x8F: d += Utf8Encode(d, SCC_GREEN); break;
case 0x90: d += Utf8Encode(d, SCC_YELLOW); break;
case 0x91: d += Utf8Encode(d, SCC_DKGREEN); break;
case 0x92: d += Utf8Encode(d, SCC_CREAM); break;
case 0x93: d += Utf8Encode(d, SCC_BROWN); break;
case 0x94: d += Utf8Encode(d, SCC_WHITE); break;
case 0x95: d += Utf8Encode(d, SCC_LTBLUE); break;
case 0x96: d += Utf8Encode(d, SCC_GRAY); break;
case 0x97: d += Utf8Encode(d, SCC_DKBLUE); break;
case 0x98: d += Utf8Encode(d, SCC_BLACK); break;
case 0x9A: {
int code = *str++;
switch (code) {
case 0x00: goto string_end;
case 0x01: d += Utf8Encode(d, SCC_NEWGRF_PRINT_QWORD_CURRENCY); break;
/* 0x02: ignore next colour byte is not supported. It works on the final
* string and as such hooks into the string drawing routine. At that
* point many things already happened, such as splitting up of strings
* when drawn over multiple lines or right-to-left translations, which
* make the behaviour peculiar, e.g. only happening at specific width
* of windows. Or we need to add another pass over the string to just
* support this. As such it is not implemented in OpenTTD. */
case 0x03: {
if (str[0] == '\0' || str[1] == '\0') goto string_end;
uint16 tmp = ((uint8)*str++);
tmp |= ((uint8)*str++) << 8;
d += Utf8Encode(d, SCC_NEWGRF_PUSH_WORD);
d += Utf8Encode(d, tmp);
break;
}
case 0x04:
if (str[0] == '\0') goto string_end;
d += Utf8Encode(d, SCC_NEWGRF_UNPRINT);
d += Utf8Encode(d, *str++);
break;
case 0x06: d += Utf8Encode(d, SCC_NEWGRF_PRINT_BYTE_HEX); break;
case 0x07: d += Utf8Encode(d, SCC_NEWGRF_PRINT_WORD_HEX); break;
case 0x08: d += Utf8Encode(d, SCC_NEWGRF_PRINT_DWORD_HEX); break;
/* 0x09, 0x0A are TTDPatch internal use only string codes. */
case 0x0B: d += Utf8Encode(d, SCC_NEWGRF_PRINT_QWORD_HEX); break;
case 0x0C: d += Utf8Encode(d, SCC_NEWGRF_PRINT_WORD_STATION_NAME); break;
case 0x0D: d += Utf8Encode(d, SCC_NEWGRF_PRINT_WORD_WEIGHT_LONG); break;
case 0x0E:
case 0x0F: {
if (str[0] == '\0') goto string_end;
const LanguageMap *lm = LanguageMap::GetLanguageMap(grfid, language_id);
int index = *str++;
int mapped = lm != NULL ? lm->GetMapping(index, code == 0x0E) : -1;
if (mapped >= 0) {
d += Utf8Encode(d, code == 0x0E ? SCC_GENDER_INDEX : SCC_SET_CASE);
d += Utf8Encode(d, code == 0x0E ? mapped : mapped + 1);
}
break;
}
case 0x10:
case 0x11:
if (str[0] == '\0') goto string_end;
if (mapping == NULL) {
if (code == 0x10) str++; // Skip the index
grfmsg(1, "choice list %s marker found when not expected", code == 0x10 ? "next" : "default");
break;
} else {
/* Terminate the previous string. */
*d = '\0';
int index = (code == 0x10 ? *str++ : 0);
if (mapping->strings.Contains(index)) {
grfmsg(1, "duplicate choice list string, ignoring");
d++;
} else {
d = mapping->strings[index] = MallocT<char>(strlen(str) * 10 + 1);
}
}
break;
case 0x12:
if (mapping == NULL) {
grfmsg(1, "choice list end marker found when not expected");
} else {
/* Terminate the previous string. */
*d = '\0';
/* Now we can start flushing everything and clean everything up. */
d = mapping->Flush(LanguageMap::GetLanguageMap(grfid, language_id));
delete mapping;
mapping = NULL;
}
break;
case 0x13:
case 0x14:
case 0x15:
if (str[0] == '\0') goto string_end;
if (mapping != NULL) {
grfmsg(1, "choice lists can't be stacked, it's going to get messy now...");
if (code != 0x14) str++;
} else {
static const StringControlCode mp[] = { SCC_GENDER_LIST, SCC_SWITCH_CASE, SCC_PLURAL_LIST };
mapping = new UnmappedChoiceList(mp[code - 0x13], d, code == 0x14 ? 0 : *str++);
}
break;
case 0x16:
case 0x17:
case 0x18:
case 0x19:
case 0x1A:
case 0x1B:
case 0x1C:
case 0x1D:
d += Utf8Encode(d, SCC_NEWGRF_PRINT_DWORD_DATE_LONG + code - 0x16);
break;
default:
grfmsg(1, "missing handler for extended format code");
break;
}
break;
}
case 0x9E: d += Utf8Encode(d, 0x20AC); break; // Euro
case 0x9F: d += Utf8Encode(d, 0x0178); break; // Y with diaeresis
case 0xA0: d += Utf8Encode(d, SCC_UP_ARROW); break;
case 0xAA: d += Utf8Encode(d, SCC_DOWN_ARROW); break;
case 0xAC: d += Utf8Encode(d, SCC_CHECKMARK); break;
case 0xAD: d += Utf8Encode(d, SCC_CROSS); break;
case 0xAF: d += Utf8Encode(d, SCC_RIGHT_ARROW); break;
case 0xB4: d += Utf8Encode(d, SCC_TRAIN); break;
case 0xB5: d += Utf8Encode(d, SCC_LORRY); break;
case 0xB6: d += Utf8Encode(d, SCC_BUS); break;
case 0xB7: d += Utf8Encode(d, SCC_PLANE); break;
case 0xB8: d += Utf8Encode(d, SCC_SHIP); break;
case 0xB9: d += Utf8Encode(d, SCC_SUPERSCRIPT_M1); break;
case 0xBC: d += Utf8Encode(d, SCC_SMALL_UP_ARROW); break;
case 0xBD: d += Utf8Encode(d, SCC_SMALL_DOWN_ARROW); break;
default:
/* Validate any unhandled character */
if (!IsValidChar(c, CS_ALPHANUMERAL)) c = '?';
d += Utf8Encode(d, c);
break;
}
}
string_end:
if (mapping != NULL) {
grfmsg(1, "choice list was incomplete, the whole list is ignored");
delete mapping;
}
*d = '\0';
if (olen != NULL) *olen = d - tmp + 1;
tmp = ReallocT(tmp, d - tmp + 1);
return tmp;
}
/**
* Add a GRFText to a GRFText list.
* @param list The list where the text should be added to.
* @param text_to_add The GRFText to add to the list.
*/
void AddGRFTextToList(GRFText **list, GRFText *text_to_add)
{
GRFText **ptext, *text;
/* Loop through all languages and see if we can replace a string */
for (ptext = list; (text = *ptext) != NULL; ptext = &text->next) {
if (text->langid == text_to_add->langid) {
text_to_add->next = text->next;
*ptext = text_to_add;
delete text;
return;
}
}
/* If a string wasn't replaced, then we must append the new string */
*ptext = text_to_add;
}
/**
* Add a string to a GRFText list.
* @param list The list where the text should be added to.
* @param langid The language of the new text.
* @param grfid The grfid where this string is defined.
* @param allow_newlines Whether newlines are allowed in this string.
* @param text_to_add The text to add to the list.
* @note All text-codes will be translated.
*/
void AddGRFTextToList(struct GRFText **list, byte langid, uint32 grfid, bool allow_newlines, const char *text_to_add)
{
int len;
char *translatedtext = TranslateTTDPatchCodes(grfid, langid, allow_newlines, text_to_add, &len);
GRFText *newtext = GRFText::New(langid, translatedtext, len);
free(translatedtext);
AddGRFTextToList(list, newtext);
}
/**
* Add a GRFText to a GRFText list. The text should not contain any text-codes.
* The text will be added as a 'default language'-text.
* @param list The list where the text should be added to.
* @param text_to_add The text to add to the list.
*/
void AddGRFTextToList(struct GRFText **list, const char *text_to_add)
{
AddGRFTextToList(list, GRFText::New(0x7F, text_to_add, strlen(text_to_add) + 1));
}
/**
* Create a copy of this GRFText list.
* @param orig The GRFText list to copy.
* @return A duplicate of the given GRFText.
*/
GRFText *DuplicateGRFText(GRFText *orig)
{
GRFText *newtext = NULL;
GRFText **ptext = &newtext;
for (; orig != NULL; orig = orig->next) {
*ptext = GRFText::Copy(orig);
ptext = &(*ptext)->next;
}
return newtext;
}
/**
* Add the new read string into our structure.
*/
StringID AddGRFString(uint32 grfid, uint16 stringid, byte langid_to_add, bool new_scheme, bool allow_newlines, const char *text_to_add, StringID def_string)
{
char *translatedtext;
uint id;
/* When working with the old language scheme (grf_version is less than 7) and
* English or American is among the set bits, simply add it as English in
* the new scheme, i.e. as langid = 1.
* If English is set, it is pretty safe to assume the translations are not
* actually translated.
*/
if (!new_scheme) {
if (langid_to_add & (GRFLB_AMERICAN | GRFLB_ENGLISH)) {
langid_to_add = GRFLX_ENGLISH;
} else {
StringID ret = STR_EMPTY;
if (langid_to_add & GRFLB_GERMAN) ret = AddGRFString(grfid, stringid, GRFLX_GERMAN, true, allow_newlines, text_to_add, def_string);
if (langid_to_add & GRFLB_FRENCH) ret = AddGRFString(grfid, stringid, GRFLX_FRENCH, true, allow_newlines, text_to_add, def_string);
if (langid_to_add & GRFLB_SPANISH) ret = AddGRFString(grfid, stringid, GRFLX_SPANISH, true, allow_newlines, text_to_add, def_string);
return ret;
}
}
for (id = 0; id < _num_grf_texts; id++) {
if (_grf_text[id].grfid == grfid && _grf_text[id].stringid == stringid) {
break;
}
}
/* Too many strings allocated, return empty */
if (id == lengthof(_grf_text)) return STR_EMPTY;
int len;
translatedtext = TranslateTTDPatchCodes(grfid, langid_to_add, allow_newlines, text_to_add, &len);
GRFText *newtext = GRFText::New(langid_to_add, translatedtext, len);
free(translatedtext);
/* If we didn't find our stringid and grfid in the list, allocate a new id */
if (id == _num_grf_texts) _num_grf_texts++;
if (_grf_text[id].textholder == NULL) {
_grf_text[id].grfid = grfid;
_grf_text[id].stringid = stringid;
_grf_text[id].def_string = def_string;
}
AddGRFTextToList(&_grf_text[id].textholder, newtext);
grfmsg(3, "Added 0x%X: grfid %08X string 0x%X lang 0x%X string '%s'", id, grfid, stringid, newtext->langid, newtext->text);
return (GRFTAB << TABSIZE) + id;
}
/**
* Returns the index for this stringid associated with its grfID
*/
StringID GetGRFStringID(uint32 grfid, uint16 stringid)
{
for (uint id = 0; id < _num_grf_texts; id++) {
if (_grf_text[id].grfid == grfid && _grf_text[id].stringid == stringid) {
return (GRFTAB << TABSIZE) + id;
}
}
return STR_UNDEFINED;
}
/**
* Get a C-string from a GRFText-list. If there is a translation for the
* current language it is returned, otherwise the default translation
* is returned. If there is neither a default nor a translation for the
* current language NULL is returned.
* @param text The GRFText to get the string from.
*/
const char *GetGRFStringFromGRFText(const GRFText *text)
{
const char *default_text = NULL;
/* Search the list of lang-strings of this stringid for current lang */
for (; text != NULL; text = text->next) {
if (text->langid == _currentLangID) return text->text;
/* If the current string is English or American, set it as the
* fallback language if the specific language isn't available. */
if (text->langid == GRFLX_UNSPECIFIED || (default_text == NULL && (text->langid == GRFLX_ENGLISH || text->langid == GRFLX_AMERICAN))) {
default_text = text->text;
}
}
return default_text;
}
/**
* Get a C-string from a stringid set by a newgrf.
*/
const char *GetGRFStringPtr(uint16 stringid)
{
assert(_grf_text[stringid].grfid != 0);
const char *str = GetGRFStringFromGRFText(_grf_text[stringid].textholder);
if (str != NULL) return str;
/* Use the default string ID if the fallback string isn't available */
return GetStringPtr(_grf_text[stringid].def_string);
}
/**
* Equivalence Setter function between game and newgrf langID.
* This function will adjust _currentLangID as to what is the LangID
* of the current language set by the user.
* This function is called after the user changed language,
* from strings.cpp:ReadLanguagePack
* @param language_id iso code of current selection
*/
void SetCurrentGrfLangID(byte language_id)
{
_currentLangID = language_id;
}
bool CheckGrfLangID(byte lang_id, byte grf_version)
{
if (grf_version < 7) {
switch (_currentLangID) {
case GRFLX_GERMAN: return (lang_id & GRFLB_GERMAN) != 0;
case GRFLX_FRENCH: return (lang_id & GRFLB_FRENCH) != 0;
case GRFLX_SPANISH: return (lang_id & GRFLB_SPANISH) != 0;
default: return (lang_id & (GRFLB_ENGLISH | GRFLB_AMERICAN)) != 0;
}
}
return (lang_id == _currentLangID || lang_id == GRFLX_UNSPECIFIED);
}
/**
* Delete all items of a linked GRFText list.
* @param grftext the head of the list to delete
*/
void CleanUpGRFText(GRFText *grftext)
{
while (grftext != NULL) {
GRFText *grftext2 = grftext->next;
delete grftext;
grftext = grftext2;
}
}
/**
* House cleaning.
* Remove all strings and reset the text counter.
*/
void CleanUpStrings()
{
uint id;
for (id = 0; id < _num_grf_texts; id++) {
CleanUpGRFText(_grf_text[id].textholder);
_grf_text[id].grfid = 0;
_grf_text[id].stringid = 0;
_grf_text[id].textholder = NULL;
}
_num_grf_texts = 0;
}
struct TextRefStack {
byte stack[0x30];
byte position;
const GRFFile *grffile;
bool used;
TextRefStack() : position(0), grffile(NULL), used(false) {}
TextRefStack(const TextRefStack &stack) :
position(stack.position),
grffile(stack.grffile),
used(stack.used)
{
memcpy(this->stack, stack.stack, sizeof(this->stack));
}
uint8 PopUnsignedByte() { assert(this->position < lengthof(this->stack)); return this->stack[this->position++]; }
int8 PopSignedByte() { return (int8)this->PopUnsignedByte(); }
uint16 PopUnsignedWord()
{
uint16 val = this->PopUnsignedByte();
return val | (this->PopUnsignedByte() << 8);
}
int16 PopSignedWord() { return (int32)this->PopUnsignedWord(); }
uint32 PopUnsignedDWord()
{
uint32 val = this->PopUnsignedWord();
return val | (this->PopUnsignedWord() << 16);
}
int32 PopSignedDWord() { return (int32)this->PopUnsignedDWord(); }
uint64 PopUnsignedQWord()
{
uint64 val = this->PopUnsignedDWord();
return val | (((uint64)this->PopUnsignedDWord()) << 32);
}
int64 PopSignedQWord() { return (int64)this->PopUnsignedQWord(); }
/** Rotate the top four words down: W1, W2, W3, W4 -> W4, W1, W2, W3 */
void RotateTop4Words()
{
byte tmp[2];
for (int i = 0; i < 2; i++) tmp[i] = this->stack[this->position + i + 6];
for (int i = 5; i >= 0; i--) this->stack[this->position + i + 2] = this->stack[this->position + i];
for (int i = 0; i < 2; i++) this->stack[this->position + i] = tmp[i];
}
void PushWord(uint16 word)
{
if (this->position >= 2) {
this->position -= 2;
} else {
for (int i = lengthof(stack) - 1; i >= this->position + 2; i--) {
this->stack[i] = this->stack[i - 2];
}
}
this->stack[this->position] = GB(word, 0, 8);
this->stack[this->position + 1] = GB(word, 8, 8);
}
void ResetStack(const GRFFile *grffile)
{
assert(grffile != NULL);
this->position = 0;
this->grffile = grffile;
this->used = true;
}
void RewindStack() { this->position = 0; }
};
/** The stack that is used for TTDP compatible string code parsing */
static TextRefStack _newgrf_textrefstack;
/**
* Check whether the NewGRF text stack is in use.
* @return True iff the NewGRF text stack is used.
*/
bool UsingNewGRFTextStack()
{
return _newgrf_textrefstack.used;
}
/**
* Create a backup of the current NewGRF text stack.
* @return A copy of the current text stack.
*/
struct TextRefStack *CreateTextRefStackBackup()
{
return new TextRefStack(_newgrf_textrefstack);
}
/**
* Restore a copy of the text stack to the used stack.
* @param backup The copy to restore.
*/
void RestoreTextRefStackBackup(struct TextRefStack *backup)
{
_newgrf_textrefstack = *backup;
delete backup;
}
/**
* Start using the TTDP compatible string code parsing.
*
* On start a number of values is copied on the #TextRefStack.
* You can then use #GetString() and the normal string drawing functions,
* and they will use the #TextRefStack for NewGRF string codes.
*
* However, when you want to draw a string multiple times using the same stack,
* you have to call #RewindTextRefStack() between draws.
*
* After you are done with drawing, you must disable usage of the #TextRefStack
* by calling #StopTextRefStackUsage(), so NewGRF string codes operate on the
* normal string parameters again.
*
* @param grffile the NewGRF providing the stack data
* @param numEntries number of entries to copy from the registers
* @param values values to copy onto the stack; if NULL the temporary NewGRF registers will be used instead
*/
void StartTextRefStackUsage(const GRFFile *grffile, byte numEntries, const uint32 *values)
{
extern TemporaryStorageArray<int32, 0x110> _temp_store;
_newgrf_textrefstack.ResetStack(grffile);
byte *p = _newgrf_textrefstack.stack;
for (uint i = 0; i < numEntries; i++) {
uint32 value = values != NULL ? values[i] : _temp_store.GetValue(0x100 + i);
for (uint j = 0; j < 32; j += 8) {
*p = GB(value, j, 8);
p++;
}
}
}
/** Stop using the TTDP compatible string code parsing */
void StopTextRefStackUsage()
{
_newgrf_textrefstack.used = false;
}
void RewindTextRefStack()
{
_newgrf_textrefstack.RewindStack();
}
/**
* FormatString for NewGRF specific "magic" string control codes
* @param scc the string control code that has been read
* @param buff the buffer we're writing to
* @param str the string that we need to write
* @param argv the OpenTTD stack of values
* @param argv_size space on the stack \a argv
* @param modify_argv When true, modify the OpenTTD stack.
* @return the string control code to "execute" now
*/
uint RemapNewGRFStringControlCode(uint scc, char *buf_start, char **buff, const char **str, int64 *argv, uint argv_size, bool modify_argv)
{
switch (scc) {
default: break;
case SCC_NEWGRF_PRINT_DWORD_SIGNED:
case SCC_NEWGRF_PRINT_WORD_SIGNED:
case SCC_NEWGRF_PRINT_BYTE_SIGNED:
case SCC_NEWGRF_PRINT_WORD_UNSIGNED:
case SCC_NEWGRF_PRINT_BYTE_HEX:
case SCC_NEWGRF_PRINT_WORD_HEX:
case SCC_NEWGRF_PRINT_DWORD_HEX:
case SCC_NEWGRF_PRINT_QWORD_HEX:
case SCC_NEWGRF_PRINT_DWORD_CURRENCY:
case SCC_NEWGRF_PRINT_QWORD_CURRENCY:
case SCC_NEWGRF_PRINT_WORD_STRING_ID:
case SCC_NEWGRF_PRINT_WORD_DATE_LONG:
case SCC_NEWGRF_PRINT_DWORD_DATE_LONG:
case SCC_NEWGRF_PRINT_WORD_DATE_SHORT:
case SCC_NEWGRF_PRINT_DWORD_DATE_SHORT:
case SCC_NEWGRF_PRINT_WORD_SPEED:
case SCC_NEWGRF_PRINT_WORD_VOLUME_LONG:
case SCC_NEWGRF_PRINT_WORD_VOLUME_SHORT:
case SCC_NEWGRF_PRINT_WORD_WEIGHT_LONG:
case SCC_NEWGRF_PRINT_WORD_WEIGHT_SHORT:
case SCC_NEWGRF_PRINT_WORD_POWER:
case SCC_NEWGRF_PRINT_WORD_STATION_NAME:
if (argv_size < 1) {
DEBUG(misc, 0, "Too many NewGRF string parameters.");
return 0;
}
break;
case SCC_NEWGRF_PRINT_WORD_CARGO_LONG:
case SCC_NEWGRF_PRINT_WORD_CARGO_SHORT:
case SCC_NEWGRF_PRINT_WORD_CARGO_TINY:
if (argv_size < 2) {
DEBUG(misc, 0, "Too many NewGRF string parameters.");
return 0;
}
break;
}
if (_newgrf_textrefstack.used && modify_argv) {
switch (scc) {
default: NOT_REACHED();
case SCC_NEWGRF_PRINT_BYTE_SIGNED: *argv = _newgrf_textrefstack.PopSignedByte(); break;
case SCC_NEWGRF_PRINT_QWORD_CURRENCY: *argv = _newgrf_textrefstack.PopSignedQWord(); break;
case SCC_NEWGRF_PRINT_DWORD_CURRENCY:
case SCC_NEWGRF_PRINT_DWORD_SIGNED: *argv = _newgrf_textrefstack.PopSignedDWord(); break;
case SCC_NEWGRF_PRINT_BYTE_HEX: *argv = _newgrf_textrefstack.PopUnsignedByte(); break;
case SCC_NEWGRF_PRINT_QWORD_HEX: *argv = _newgrf_textrefstack.PopUnsignedQWord(); break;
case SCC_NEWGRF_PRINT_WORD_SPEED:
case SCC_NEWGRF_PRINT_WORD_VOLUME_LONG:
case SCC_NEWGRF_PRINT_WORD_VOLUME_SHORT:
case SCC_NEWGRF_PRINT_WORD_SIGNED: *argv = _newgrf_textrefstack.PopSignedWord(); break;
case SCC_NEWGRF_PRINT_WORD_HEX:
case SCC_NEWGRF_PRINT_WORD_WEIGHT_LONG:
case SCC_NEWGRF_PRINT_WORD_WEIGHT_SHORT:
case SCC_NEWGRF_PRINT_WORD_POWER:
case SCC_NEWGRF_PRINT_WORD_STATION_NAME:
case SCC_NEWGRF_PRINT_WORD_UNSIGNED: *argv = _newgrf_textrefstack.PopUnsignedWord(); break;
case SCC_NEWGRF_PRINT_DWORD_DATE_LONG:
case SCC_NEWGRF_PRINT_DWORD_DATE_SHORT:
case SCC_NEWGRF_PRINT_DWORD_HEX: *argv = _newgrf_textrefstack.PopUnsignedDWord(); break;
case SCC_NEWGRF_PRINT_WORD_DATE_LONG:
case SCC_NEWGRF_PRINT_WORD_DATE_SHORT: *argv = _newgrf_textrefstack.PopUnsignedWord() + DAYS_TILL_ORIGINAL_BASE_YEAR; break;
case SCC_NEWGRF_DISCARD_WORD: _newgrf_textrefstack.PopUnsignedWord(); break;
case SCC_NEWGRF_ROTATE_TOP_4_WORDS: _newgrf_textrefstack.RotateTop4Words(); break;
case SCC_NEWGRF_PUSH_WORD: _newgrf_textrefstack.PushWord(Utf8Consume(str)); break;
case SCC_NEWGRF_UNPRINT: *buff = max(*buff - Utf8Consume(str), buf_start); break;
case SCC_NEWGRF_PRINT_WORD_CARGO_LONG:
case SCC_NEWGRF_PRINT_WORD_CARGO_SHORT:
case SCC_NEWGRF_PRINT_WORD_CARGO_TINY:
argv[0] = GetCargoTranslation(_newgrf_textrefstack.PopUnsignedWord(), _newgrf_textrefstack.grffile);
argv[1] = _newgrf_textrefstack.PopUnsignedWord();
break;
case SCC_NEWGRF_PRINT_WORD_STRING_ID:
*argv = MapGRFStringID(_newgrf_textrefstack.grffile->grfid, _newgrf_textrefstack.PopUnsignedWord());
break;
}
}
switch (scc) {
default: NOT_REACHED();
case SCC_NEWGRF_PRINT_DWORD_SIGNED:
case SCC_NEWGRF_PRINT_WORD_SIGNED:
case SCC_NEWGRF_PRINT_BYTE_SIGNED:
case SCC_NEWGRF_PRINT_WORD_UNSIGNED:
return SCC_COMMA;
case SCC_NEWGRF_PRINT_BYTE_HEX:
case SCC_NEWGRF_PRINT_WORD_HEX:
case SCC_NEWGRF_PRINT_DWORD_HEX:
case SCC_NEWGRF_PRINT_QWORD_HEX:
return SCC_HEX;
case SCC_NEWGRF_PRINT_DWORD_CURRENCY:
case SCC_NEWGRF_PRINT_QWORD_CURRENCY:
return SCC_CURRENCY_LONG;
case SCC_NEWGRF_PRINT_WORD_STRING_ID:
return SCC_NEWGRF_PRINT_WORD_STRING_ID;
case SCC_NEWGRF_PRINT_WORD_DATE_LONG:
case SCC_NEWGRF_PRINT_DWORD_DATE_LONG:
return SCC_DATE_LONG;
case SCC_NEWGRF_PRINT_WORD_DATE_SHORT:
case SCC_NEWGRF_PRINT_DWORD_DATE_SHORT:
return SCC_DATE_SHORT;
case SCC_NEWGRF_PRINT_WORD_SPEED:
return SCC_VELOCITY;
case SCC_NEWGRF_PRINT_WORD_VOLUME_LONG:
return SCC_VOLUME_LONG;
case SCC_NEWGRF_PRINT_WORD_VOLUME_SHORT:
return SCC_VOLUME_SHORT;
case SCC_NEWGRF_PRINT_WORD_WEIGHT_LONG:
return SCC_WEIGHT_LONG;
case SCC_NEWGRF_PRINT_WORD_WEIGHT_SHORT:
return SCC_WEIGHT_SHORT;
case SCC_NEWGRF_PRINT_WORD_POWER:
return SCC_POWER;
case SCC_NEWGRF_PRINT_WORD_CARGO_LONG:
return SCC_CARGO_LONG;
case SCC_NEWGRF_PRINT_WORD_CARGO_SHORT:
return SCC_CARGO_SHORT;
case SCC_NEWGRF_PRINT_WORD_CARGO_TINY:
return SCC_CARGO_TINY;
case SCC_NEWGRF_PRINT_WORD_STATION_NAME:
return SCC_STATION_NAME;
case SCC_NEWGRF_DISCARD_WORD:
case SCC_NEWGRF_ROTATE_TOP_4_WORDS:
case SCC_NEWGRF_PUSH_WORD:
case SCC_NEWGRF_UNPRINT:
return 0;
}
}
|