1 /* Conversion from and to IBM1364.
2 Copyright (C) 2005-2021 Free Software Foundation, Inc.
3 This file is part of the GNU C Library.
4 Contributed by Masahide Washizawa <washi@jp.ibm.com>, 2005.
6 The GNU C Library is free software; you can redistribute it and/or
7 modify it under the terms of the GNU Lesser General Public
8 License as published by the Free Software Foundation; either
9 version 2.1 of the License, or (at your option) any later version.
11 The GNU C Library is distributed in the hope that it will be useful,
12 but WITHOUT ANY WARRANTY; without even the implied warranty of
13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 Lesser General Public License for more details.
16 You should have received a copy of the GNU Lesser General Public
17 License along with the GNU C Library; if not, see
18 <https://www.gnu.org/licenses/>. */
26 /* This is really the IBM1364 converter, not another module sharing
28 # define DATA_HEADER "ibm1364.h"
29 # define CHARSET_NAME "IBM1364//"
30 # define FROM_LOOP from_ibm1364
31 # define TO_LOOP to_ibm1364
32 # define SB_TO_UCS4 __ibm1364sb_to_ucs4
33 # define DB_TO_UCS4_IDX __ibm1364db_to_ucs4_idx
34 # define DB_TO_UCS4 __ibm1364db_to_ucs4
35 # define UCS4_TO_SB_IDX __ucs4_to_ibm1364sb_idx
36 # define UCS4_TO_SB __ucs4_to_ibm1364sb
37 # define UCS4_TO_DB_IDX __ucs4_to_ibm1364db_idx
38 # define UCS4_TO_DB __ucs4_to_ibm1364db
39 # define UCS_LIMIT 0xffff
45 /* The shift sequences for this charset (it does not use ESC). */
46 #define SI 0x0F /* Shift In, host code to turn DBCS off. */
47 #define SO 0x0E /* Shift Out, host code to turn DBCS on. */
49 /* Definitions used in the body of the `gconv' function. */
50 #define MIN_NEEDED_FROM 1
51 #define MAX_NEEDED_FROM 2
52 #define MIN_NEEDED_TO 4
54 # define MAX_NEEDED_TO 8
56 # define MAX_NEEDED_TO 4
58 #define ONE_DIRECTION 0
59 #define PREPARE_LOOP \
61 int *curcsp = &data->__statep->__count;
62 #define EXTRA_LOOP_ARGS , curcsp
64 /* Definitions of initialization and destructor function. */
69 /* Since this is a stateful encoding we have to provide code which resets
70 the output state to the initial state. This has to be done during the
72 #define EMIT_SHIFT_TO_INIT \
73 if ((data->__statep->__count & ~7) != sb) \
76 data->__statep->__count &= 7; \
79 /* We are not in the initial state. To switch back we have \
81 if (__glibc_unlikely (outbuf >= outend)) \
82 /* We don't have enough room in the output buffer. */ \
83 status = __GCONV_FULL_OUTPUT; \
86 /* Write out the shift sequence. */ \
88 data->__statep->__count &= 7; \
94 /* Since we might have to reset input pointer we must be able to save
95 and retore the state. */
96 #define SAVE_RESET_STATE(Save) \
98 save_curcs = *curcsp; \
103 /* Current codeset type. */
111 /* Subroutine to write out converted UCS4 from IBM-13XX. */
113 # define SUB_COMBINED_UCS_FROM_IBM13XX \
115 if (res != UCS_LIMIT || ch < __TO_UCS4_COMBINED_MIN \
116 || ch > __TO_UCS4_COMBINED_MAX) \
118 put32 (outptr, res); \
123 /* This is a combined character. Make sure we have room. */ \
124 if (__glibc_unlikely (outptr + 8 > outend)) \
126 result = __GCONV_FULL_OUTPUT; \
130 const struct divide *cmbp \
131 = &DB_TO_UCS4_COMB[ch - __TO_UCS4_COMBINED_MIN]; \
132 assert (cmbp->res1 != 0 && cmbp->res2 != 0); \
134 put32 (outptr, cmbp->res1); \
136 put32 (outptr, cmbp->res2); \
141 # define SUB_COMBINED_UCS_FROM_IBM13XX \
143 put32 (outptr, res); \
146 #endif /* HAS_COMBINED */
149 /* First, define the conversion function from IBM-13XX to UCS4. */
150 #define MIN_NEEDED_INPUT MIN_NEEDED_FROM
151 #define MAX_NEEDED_INPUT MAX_NEEDED_FROM
152 #define MIN_NEEDED_OUTPUT MIN_NEEDED_TO
153 #define MAX_NEEDED_OUTPUT MAX_NEEDED_TO
154 #define LOOPFCT FROM_LOOP
157 uint32_t ch = *inptr; \
159 if (__builtin_expect (ch, 0) == SO) \
161 /* Shift OUT, change to DBCS converter (redundant escape okay). */ \
166 if (__builtin_expect (ch, 0) == SI) \
168 /* Shift IN, change to SBCS converter (redundant escape okay). */ \
176 /* Use the IBM13XX table for single byte. */ \
177 uint32_t res = SB_TO_UCS4[ch]; \
178 if (__builtin_expect (res, L'\1') == L'\0' && ch != '\0') \
180 /* This is an illegal character. */ \
181 if (! ignore_errors_p ()) \
183 result = __GCONV_ILLEGAL_INPUT; \
190 put32 (outptr, res); \
197 assert (curcs == db); \
199 if (__glibc_unlikely (inptr + 1 >= inend)) \
201 /* The second character is not available. Store the \
202 intermediate result. */ \
203 result = __GCONV_INCOMPLETE_INPUT; \
207 ch = (ch * 0x100) + inptr[1]; \
209 /* Use the IBM1364 table for double byte. */ \
210 const struct gap *rp2 = DB_TO_UCS4_IDX; \
211 while (ch > rp2->end) \
215 if (__builtin_expect (rp2->start == 0xffff, 0) \
216 || __builtin_expect (ch < rp2->start, 0) \
217 || (res = DB_TO_UCS4[ch + rp2->idx], \
218 __builtin_expect (res, L'\1') == L'\0' && ch != '\0')) \
220 /* This is an illegal character. */ \
221 if (! ignore_errors_p ()) \
223 result = __GCONV_ILLEGAL_INPUT; \
230 SUB_COMBINED_UCS_FROM_IBM13XX; \
235 #define LOOP_NEED_FLAGS
236 #define EXTRA_LOOP_DECLS , int *curcsp
237 #define INIT_PARAMS int curcs = *curcsp & ~7
238 #define UPDATE_PARAMS *curcsp = curcs
239 #include <iconv/loop.c>
242 /* Subroutine to convert two UCS4 codes to IBM-13XX. */
244 # define SUB_COMBINED_UCS_TO_IBM13XX \
246 const struct combine *cmbp = UCS4_COMB_TO_DB; \
247 while (cmbp->res1 < ch) \
249 /* XXX if last char is beginning of combining store in state */ \
250 if (cmbp->res1 == ch && inptr + 4 < inend) \
252 /* See if input is part of a combined character. */ \
253 uint32_t ch_next = get32 (inptr + 4); \
254 while (cmbp->res2 != ch_next) \
257 if (cmbp->res1 != ch) \
261 /* It is a combined character. First make sure we are in \
262 double byte mode. */ \
265 /* We know there is room for at least one byte. */ \
270 if (__glibc_unlikely (outptr + 2 > outend)) \
272 result = __GCONV_FULL_OUTPUT; \
275 *outptr++ = cmbp->ch[0]; \
276 *outptr++ = cmbp->ch[1]; \
284 # define SUB_COMBINED_UCS_TO_IBM13XX
285 #endif /* HAS_COMBINED */
288 /* Next, define the other direction. */
289 #define MIN_NEEDED_INPUT MIN_NEEDED_TO
290 #define MAX_NEEDED_INPUT MAX_NEEDED_TO
291 #define MIN_NEEDED_OUTPUT MIN_NEEDED_FROM
292 #define MAX_NEEDED_OUTPUT MAX_NEEDED_FROM
293 #define LOOPFCT TO_LOOP
296 uint32_t ch = get32 (inptr); \
298 if (__glibc_unlikely (ch >= UCS_LIMIT)) \
300 UNICODE_TAG_HANDLER (ch, 4); \
302 if (! ignore_errors_p ()) \
304 result = __GCONV_ILLEGAL_INPUT; \
312 SUB_COMBINED_UCS_TO_IBM13XX; \
314 const struct gap *rp1 = UCS4_TO_SB_IDX; \
315 while (ch > rp1->end) \
318 /* Use the UCS4 table for single byte. */ \
320 if (__builtin_expect (ch < rp1->start, 0) \
321 || (cp = UCS4_TO_SB[ch + rp1->idx], \
322 __builtin_expect (cp[0], L'\1') == L'\0' && ch != '\0')) \
324 /* Use the UCS4 table for double byte. */ \
325 const struct gap *rp2 = UCS4_TO_DB_IDX; \
326 while (ch > rp2->end) \
329 if (__builtin_expect (ch < rp2->start, 0) \
330 || (cp = UCS4_TO_DB[ch + rp2->idx], \
331 __builtin_expect (cp[0], L'\1') == L'\0' && ch != '\0')) \
333 /* This is an illegal character. */ \
334 if (! ignore_errors_p ()) \
336 result = __GCONV_ILLEGAL_INPUT; \
345 /* We know there is room for at least one byte. */ \
350 if (__glibc_unlikely (outptr + 2 > outend)) \
352 result = __GCONV_FULL_OUTPUT; \
361 if (__glibc_unlikely (curcs == db)) \
363 /* We know there is room for at least one byte. */ \
367 if (__glibc_unlikely (outptr >= outend)) \
369 result = __GCONV_FULL_OUTPUT; \
377 /* Now that we wrote the output increment the input pointer. */ \
380 #define LOOP_NEED_FLAGS
381 #define EXTRA_LOOP_DECLS , int *curcsp
382 #define INIT_PARAMS int curcs = *curcsp & ~7
383 #define REINIT_PARAMS curcs = *curcsp & ~7
384 #define UPDATE_PARAMS *curcsp = curcs
385 #include <iconv/loop.c>
387 /* Now define the toplevel functions. */
388 #include <iconv/skeleton.c>