1 |
frodo |
2 |
/* |
2 |
|
|
parse_driver.c - Part of psiconv, a PSION 5 file formats converter |
3 |
|
|
Copyright (c) 1999 Frodo Looijaard <frodol@dds.nl> |
4 |
|
|
|
5 |
|
|
This program is free software; you can redistribute it and/or modify |
6 |
|
|
it under the terms of the GNU General Public License as published by |
7 |
|
|
the Free Software Foundation; either version 2 of the License, or |
8 |
|
|
(at your option) any later version. |
9 |
|
|
|
10 |
|
|
This program is distributed in the hope that it will be useful, |
11 |
|
|
but WITHOUT ANY WARRANTY; without even the implied warranty of |
12 |
|
|
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
13 |
|
|
GNU General Public License for more details. |
14 |
|
|
|
15 |
|
|
You should have received a copy of the GNU General Public License |
16 |
|
|
along with this program; if not, write to the Free Software |
17 |
|
|
Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. |
18 |
|
|
*/ |
19 |
|
|
|
20 |
|
|
#include "config.h" |
21 |
|
|
#include <stdlib.h> |
22 |
|
|
|
23 |
|
|
#include "parse.h" |
24 |
|
|
#include "parse_routines.h" |
25 |
|
|
#include "data.h" |
26 |
|
|
|
27 |
|
|
psiconv_file_type_t psiconv_file_type(psiconv_buffer buf) |
28 |
|
|
{ |
29 |
|
|
psiconv_header_section header; |
30 |
|
|
psiconv_file_type_t res; |
31 |
|
|
|
32 |
|
|
psiconv_parse_header_section(buf,0,0,NULL,&header); |
33 |
|
|
res = header->file; |
34 |
|
|
psiconv_free_header_section(header); |
35 |
|
|
return res; |
36 |
|
|
} |
37 |
|
|
|
38 |
|
|
int psiconv_parse(const psiconv_buffer buf,psiconv_file *result) |
39 |
|
|
{ |
40 |
|
|
int res=0; |
41 |
|
|
int lev=0; |
42 |
|
|
int off=0; |
43 |
|
|
|
44 |
|
|
(*result) = malloc(sizeof(**result)); |
45 |
|
|
|
46 |
|
|
(*result)->type = psiconv_file_type(buf); |
47 |
|
|
if ((*result)->type == psiconv_unknown_file) { |
48 |
|
|
psiconv_warn(lev+1,off,"Unknown file type: can't parse!"); |
49 |
|
|
(*result)->file = NULL; |
50 |
|
|
res = -1; |
51 |
|
|
} else if ((*result)->type == psiconv_word_file) |
52 |
|
|
res = psiconv_parse_word_file(buf,lev+2,off, |
53 |
|
|
(psiconv_word_f *)(&((*result)->file))); |
54 |
|
|
else if ((*result)->type == psiconv_texted_file) |
55 |
|
|
res = psiconv_parse_texted_file(buf,lev+2,off, |
56 |
|
|
(psiconv_texted_f *)(&((*result)->file))); |
57 |
frodo |
12 |
else if ((*result)->type == psiconv_mbm_file) |
58 |
|
|
res = psiconv_parse_mbm_file(buf,lev+2,off, |
59 |
|
|
(psiconv_mbm_f *)(&((*result)->file))); |
60 |
frodo |
2 |
else { |
61 |
|
|
psiconv_warn(lev+1,off,"Can't parse this file yet!"); |
62 |
|
|
(*result)->file = NULL; |
63 |
|
|
} |
64 |
|
|
res = -1; |
65 |
|
|
|
66 |
|
|
return res; |
67 |
|
|
} |
68 |
|
|
|
69 |
frodo |
12 |
int psiconv_parse_mbm_file(const psiconv_buffer buf,int lev, psiconv_u32 off, |
70 |
|
|
psiconv_mbm_f *result) |
71 |
|
|
{ |
72 |
|
|
int res=0; |
73 |
|
|
int i; |
74 |
|
|
psiconv_mbm_jumptable_section table; |
75 |
|
|
psiconv_paint_data_section paint; |
76 |
|
|
psiconv_header_section header; |
77 |
|
|
psiconv_u32 *entry; |
78 |
|
|
|
79 |
|
|
psiconv_progress(lev+1,off,"Going to read a mbm file"); |
80 |
|
|
*result = malloc(sizeof(**result)); |
81 |
|
|
|
82 |
|
|
psiconv_progress(lev+2,off, "Going to read the header section"); |
83 |
|
|
res |= psiconv_parse_header_section(buf,lev+2,off,NULL,&header); |
84 |
|
|
|
85 |
|
|
psiconv_progress(lev+2,off,"Going to read the MBM jumptable"); |
86 |
|
|
res |= psiconv_parse_mbm_jumptable_section(buf,lev+2, |
87 |
|
|
header->section_table_offset, |
88 |
|
|
NULL,&table); |
89 |
|
|
|
90 |
frodo |
13 |
psiconv_progress(lev+2,off,"Going to read the picture sections"); |
91 |
frodo |
12 |
(*result)->sections = psiconv_list_new(sizeof(*paint)); |
92 |
|
|
for (i = 0; i < psiconv_list_length(table); i ++) { |
93 |
|
|
entry = psiconv_list_get(table,i); |
94 |
frodo |
13 |
psiconv_progress(lev+3,off,"Going to read picture section %i",i); |
95 |
|
|
psiconv_parse_paint_data_section(buf,lev+3,*entry,NULL,&paint); |
96 |
|
|
psiconv_list_add((*result)->sections,paint); |
97 |
frodo |
12 |
} |
98 |
|
|
|
99 |
|
|
psiconv_free_header_section(header); |
100 |
|
|
psiconv_free_mbm_jumptable_section(table); |
101 |
|
|
psiconv_progress(lev+1,off,"End of mbm file"); |
102 |
|
|
return res; |
103 |
|
|
} |
104 |
|
|
|
105 |
frodo |
2 |
int psiconv_parse_texted_file(const psiconv_buffer buf,int lev, psiconv_u32 off, |
106 |
|
|
psiconv_texted_f *result) |
107 |
|
|
{ |
108 |
|
|
int res=0; |
109 |
|
|
psiconv_header_section header; |
110 |
|
|
psiconv_section_table_section table; |
111 |
|
|
psiconv_application_id_section appl_id; |
112 |
|
|
char *temp_str; |
113 |
|
|
psiconv_character_layout base_char; |
114 |
|
|
psiconv_paragraph_layout base_para; |
115 |
|
|
psiconv_u32 page_sec = 0; |
116 |
|
|
psiconv_u32 texted_sec = 0; |
117 |
|
|
psiconv_u32 applid_sec = 0; |
118 |
|
|
psiconv_section_table_entry entry; |
119 |
|
|
int i; |
120 |
|
|
|
121 |
|
|
psiconv_progress(lev+1,off,"Going to read a texted file"); |
122 |
|
|
*result = malloc(sizeof(**result)); |
123 |
|
|
|
124 |
|
|
psiconv_progress(lev+2,off, "Going to read the header section"); |
125 |
|
|
res |= psiconv_parse_header_section(buf,lev+2,off,NULL,&header); |
126 |
|
|
|
127 |
|
|
psiconv_progress(lev+2,header->section_table_offset, |
128 |
|
|
"Going to read the section table section"); |
129 |
|
|
res |= psiconv_parse_section_table_section(buf,lev+2, |
130 |
|
|
header->section_table_offset, |
131 |
|
|
NULL,&table); |
132 |
|
|
|
133 |
|
|
for (i = 0; i < psiconv_list_length(table); i ++) { |
134 |
|
|
psiconv_progress(lev+2,header->section_table_offset, |
135 |
|
|
"Going to read entry %d",i); |
136 |
|
|
entry = psiconv_list_get(table,i); |
137 |
|
|
if (entry->id == PSICONV_ID_APPL_ID_SECTION) { |
138 |
|
|
applid_sec = entry->offset; |
139 |
|
|
psiconv_debug(lev+3,header->section_table_offset, |
140 |
|
|
"Found the Application ID section at %08x",applid_sec); |
141 |
|
|
} else if (entry->id == PSICONV_ID_PAGE_LAYOUT_SECTION) { |
142 |
|
|
page_sec = entry->offset; |
143 |
|
|
psiconv_debug(lev+3,header->section_table_offset, |
144 |
|
|
"Found the Page Layout section at %08x",page_sec); |
145 |
|
|
} else if (entry->id == PSICONV_ID_TEXTED) { |
146 |
|
|
texted_sec = entry->offset; |
147 |
|
|
psiconv_debug(lev+3,header->section_table_offset, |
148 |
|
|
"Found the TextEd section at %08x",texted_sec); |
149 |
|
|
} else { |
150 |
|
|
psiconv_warn(lev+3,header->section_table_offset, |
151 |
|
|
"Found unknown section in the Section Table"); |
152 |
|
|
psiconv_debug(lev+3,header->section_table_offset, |
153 |
|
|
"Section ID %08x, offset %08x",entry->id,entry->offset); |
154 |
|
|
res = -1; |
155 |
|
|
} |
156 |
|
|
} |
157 |
|
|
|
158 |
|
|
psiconv_progress(lev+2,header->section_table_offset, |
159 |
|
|
"Looking for the Application ID section"); |
160 |
|
|
if (! applid_sec) { |
161 |
|
|
psiconv_warn(lev+2,header->section_table_offset, |
162 |
|
|
"Application ID section not found in the section table"); |
163 |
|
|
res = -1; |
164 |
|
|
} else { |
165 |
|
|
psiconv_debug(lev+2,header->section_table_offset, |
166 |
|
|
"Application ID section at offset %08x",applid_sec); |
167 |
|
|
res |= psiconv_parse_application_id_section(buf,lev+2,applid_sec,NULL, |
168 |
|
|
&appl_id); |
169 |
|
|
} |
170 |
|
|
if ((appl_id->id != PSICONV_ID_TEXTED) || |
171 |
|
|
strcmp(appl_id->name,"TextEd.app")) { |
172 |
|
|
psiconv_warn(lev+2,applid_sec, |
173 |
|
|
"Application ID section contains unexpected data"); |
174 |
|
|
psiconv_debug(lev+2,applid_sec,"ID: %08x expected, %08x found", |
175 |
|
|
PSICONV_ID_TEXTED,appl_id->id); |
176 |
|
|
temp_str = psiconv_make_printable(appl_id->name); |
177 |
|
|
psiconv_debug(lev+2,applid_sec,"Name: `%s' expected, `%s' found", |
178 |
|
|
PSICONV_ID_TEXTED,temp_str); |
179 |
|
|
free(temp_str); |
180 |
|
|
} |
181 |
|
|
|
182 |
|
|
psiconv_progress(lev+2,header->section_table_offset, |
183 |
|
|
"Looking for the Page layout section"); |
184 |
|
|
if (! page_sec) { |
185 |
|
|
psiconv_warn(lev+2,header->section_table_offset, |
186 |
|
|
"Page layout section not found in the section table"); |
187 |
|
|
(*result)->page_sec = NULL; |
188 |
|
|
res = -1; |
189 |
|
|
} else { |
190 |
|
|
psiconv_debug(lev+2,header->section_table_offset, |
191 |
|
|
"Page layout section at offset %08x",page_sec); |
192 |
|
|
res |= psiconv_parse_page_layout_section(buf,lev+2,page_sec,NULL, |
193 |
|
|
&(*result)->page_sec); |
194 |
|
|
} |
195 |
|
|
|
196 |
|
|
base_char = psiconv_basic_character_layout(); |
197 |
|
|
base_para = psiconv_basic_paragraph_layout(); |
198 |
|
|
|
199 |
|
|
psiconv_progress(lev+2,header->section_table_offset, |
200 |
|
|
"Looking for the TextEd section"); |
201 |
|
|
if (! texted_sec) { |
202 |
|
|
psiconv_warn(lev+2,header->section_table_offset, |
203 |
|
|
"TextEd section not found in the section table"); |
204 |
|
|
(*result)->texted_sec = NULL; |
205 |
|
|
res = -1; |
206 |
|
|
} else { |
207 |
|
|
psiconv_debug(lev+2,header->section_table_offset, |
208 |
|
|
"TextEd section at offset %08x",texted_sec); |
209 |
|
|
res |= psiconv_parse_texted_section(buf,lev+2,texted_sec,NULL, |
210 |
|
|
&(*result)->texted_sec, |
211 |
|
|
base_char,base_para); |
212 |
|
|
} |
213 |
|
|
psiconv_free_character_layout(base_char); |
214 |
|
|
psiconv_free_paragraph_layout(base_para); |
215 |
|
|
|
216 |
|
|
psiconv_free_application_id_section(appl_id); |
217 |
|
|
psiconv_free_header_section(header); |
218 |
|
|
psiconv_free_section_table_section(table); |
219 |
|
|
|
220 |
|
|
psiconv_progress(lev+1,off,"End of word file"); |
221 |
|
|
return res; |
222 |
|
|
} |
223 |
|
|
|
224 |
|
|
int psiconv_parse_word_file(const psiconv_buffer buf,int lev, psiconv_u32 off, |
225 |
|
|
psiconv_word_f *result) |
226 |
|
|
{ |
227 |
|
|
int res=0; |
228 |
|
|
psiconv_header_section header; |
229 |
|
|
psiconv_section_table_section table; |
230 |
|
|
psiconv_application_id_section appl_id; |
231 |
|
|
char *temp_str; |
232 |
|
|
psiconv_u32 pwd_sec = 0; |
233 |
|
|
psiconv_u32 status_sec = 0; |
234 |
|
|
psiconv_u32 styles_sec = 0; |
235 |
|
|
psiconv_u32 page_sec = 0; |
236 |
|
|
psiconv_u32 text_sec = 0; |
237 |
|
|
psiconv_u32 layout_sec = 0; |
238 |
|
|
psiconv_u32 applid_sec = 0; |
239 |
|
|
psiconv_section_table_entry entry; |
240 |
|
|
int i; |
241 |
|
|
|
242 |
|
|
psiconv_progress(lev+1,off,"Going to read a word file"); |
243 |
|
|
*result = malloc(sizeof(**result)); |
244 |
|
|
|
245 |
|
|
psiconv_progress(lev+2,off, "Going to read the header section"); |
246 |
|
|
res |= psiconv_parse_header_section(buf,lev+2,off,NULL,&header); |
247 |
|
|
|
248 |
|
|
psiconv_progress(lev+2,header->section_table_offset, |
249 |
|
|
"Going to read the section table section"); |
250 |
|
|
res |= psiconv_parse_section_table_section(buf,lev+2, |
251 |
|
|
header->section_table_offset, |
252 |
|
|
NULL,&table); |
253 |
|
|
|
254 |
|
|
for (i = 0; i < psiconv_list_length(table); i ++) { |
255 |
|
|
psiconv_progress(lev+2,header->section_table_offset, |
256 |
|
|
"Going to read entry %d",i); |
257 |
|
|
entry = psiconv_list_get(table,i); |
258 |
|
|
if (entry->id == PSICONV_ID_APPL_ID_SECTION) { |
259 |
|
|
applid_sec = entry->offset; |
260 |
|
|
psiconv_debug(lev+3,header->section_table_offset, |
261 |
|
|
"Found the Application ID section at %08x",applid_sec); |
262 |
|
|
} else if (entry->id == PSICONV_ID_PAGE_LAYOUT_SECTION) { |
263 |
|
|
page_sec = entry->offset; |
264 |
|
|
psiconv_debug(lev+3,header->section_table_offset, |
265 |
|
|
"Found the Page Layout section at %08x",page_sec); |
266 |
|
|
} else if (entry->id == PSICONV_ID_TEXT_SECTION) { |
267 |
|
|
text_sec = entry->offset; |
268 |
|
|
psiconv_debug(lev+3,header->section_table_offset, |
269 |
|
|
"Found the Text section at %08x",text_sec); |
270 |
|
|
} else if (entry->id == PSICONV_ID_PASSWORD_SECTION) { |
271 |
|
|
pwd_sec = entry->offset; |
272 |
|
|
psiconv_debug(lev+3,header->section_table_offset, |
273 |
|
|
"Found the Password section at %08x",pwd_sec); |
274 |
|
|
psiconv_warn(lev+3,header->section_table_offset, |
275 |
|
|
"Password section found - can't read encrypted data"); |
276 |
|
|
res = -1; |
277 |
|
|
} else if (entry->id == PSICONV_ID_WORD_STATUS_SECTION) { |
278 |
|
|
status_sec = entry->offset; |
279 |
|
|
psiconv_debug(lev+3,header->section_table_offset, |
280 |
|
|
"Found the Word Status section at %08x",status_sec); |
281 |
|
|
} else if (entry->id == PSICONV_ID_WORD_STYLES_SECTION) { |
282 |
|
|
styles_sec = entry->offset; |
283 |
|
|
psiconv_debug(lev+3,header->section_table_offset, |
284 |
|
|
"Found the Word Styles section at %08x",styles_sec); |
285 |
|
|
} else if (entry->id == PSICONV_ID_LAYOUT_SECTION) { |
286 |
|
|
layout_sec = entry->offset; |
287 |
|
|
psiconv_debug(lev+3,header->section_table_offset, |
288 |
|
|
"Found the Layout section at %08x",layout_sec); |
289 |
|
|
} else { |
290 |
|
|
psiconv_warn(lev+3,header->section_table_offset, |
291 |
|
|
"Found unknown section in the Section Table"); |
292 |
|
|
psiconv_debug(lev+3,header->section_table_offset, |
293 |
|
|
"Section ID %08x, offset %08x",entry->id,entry->offset); |
294 |
|
|
res = -1; |
295 |
|
|
} |
296 |
|
|
} |
297 |
|
|
|
298 |
|
|
|
299 |
|
|
psiconv_progress(lev+2,header->section_table_offset, |
300 |
|
|
"Looking for the Status section"); |
301 |
|
|
if (!status_sec) { |
302 |
|
|
psiconv_warn(lev+2,header->section_table_offset, |
303 |
|
|
"Status section not found in the section table"); |
304 |
|
|
res = -1; |
305 |
|
|
} else { |
306 |
|
|
psiconv_debug(lev+2,header->section_table_offset, |
307 |
|
|
"Status section at offset %08x",status_sec); |
308 |
|
|
res |= psiconv_parse_word_status_section(buf,lev+2,status_sec,NULL, |
309 |
|
|
&((*result)->status_sec)); |
310 |
|
|
} |
311 |
|
|
|
312 |
|
|
psiconv_progress(lev+2,header->section_table_offset, |
313 |
|
|
"Looking for the Application ID section"); |
314 |
|
|
if (! applid_sec) { |
315 |
|
|
psiconv_warn(lev+2,header->section_table_offset, |
316 |
|
|
"Application ID section not found in the section table"); |
317 |
|
|
res = -1; |
318 |
|
|
} else { |
319 |
|
|
psiconv_debug(lev+2,header->section_table_offset, |
320 |
|
|
"Application ID section at offset %08x",applid_sec); |
321 |
|
|
res |= psiconv_parse_application_id_section(buf,lev+2,applid_sec,NULL, |
322 |
|
|
&appl_id); |
323 |
|
|
} |
324 |
|
|
if ((appl_id->id != PSICONV_ID_WORD) || |
325 |
|
|
strcmp(appl_id->name,"Word.app")) { |
326 |
|
|
psiconv_warn(lev+2,applid_sec, |
327 |
|
|
"Application ID section contains unexpected data"); |
328 |
|
|
psiconv_debug(lev+2,applid_sec,"ID: %08x expected, %08x found", |
329 |
|
|
PSICONV_ID_WORD,appl_id->id); |
330 |
|
|
temp_str = psiconv_make_printable(appl_id->name); |
331 |
|
|
psiconv_debug(lev+2,applid_sec,"Name: `%s' expected, `%s' found", |
332 |
|
|
PSICONV_ID_WORD,temp_str); |
333 |
|
|
free(temp_str); |
334 |
|
|
} |
335 |
|
|
|
336 |
|
|
psiconv_progress(lev+2,header->section_table_offset, |
337 |
|
|
"Looking for the Page layout section"); |
338 |
|
|
if (! page_sec) { |
339 |
|
|
psiconv_warn(lev+2,header->section_table_offset, |
340 |
|
|
"Page layout section not found in the section table"); |
341 |
|
|
(*result)->page_sec = NULL; |
342 |
|
|
res = -1; |
343 |
|
|
} else { |
344 |
|
|
psiconv_debug(lev+2,header->section_table_offset, |
345 |
|
|
"Page layout section at offset %08x",page_sec); |
346 |
|
|
res |= psiconv_parse_page_layout_section(buf,lev+2,page_sec,NULL, |
347 |
|
|
&(*result)->page_sec); |
348 |
|
|
} |
349 |
|
|
|
350 |
|
|
psiconv_progress(lev+2,header->section_table_offset, |
351 |
|
|
"Looking for the Word Style section"); |
352 |
|
|
if (!styles_sec) { |
353 |
|
|
psiconv_warn(lev+2,header->section_table_offset, |
354 |
|
|
"Word styles section not found in the section table"); |
355 |
|
|
(*result)->styles_sec = NULL; |
356 |
|
|
res = -1; |
357 |
|
|
} else { |
358 |
|
|
psiconv_debug(lev+2,header->section_table_offset, |
359 |
|
|
"Word styles section at offset %08x",styles_sec); |
360 |
|
|
res |= psiconv_parse_word_styles_section(buf,lev+2,styles_sec,NULL, |
361 |
|
|
&(*result)->styles_sec); |
362 |
|
|
} |
363 |
|
|
|
364 |
|
|
psiconv_progress(lev+2,header->section_table_offset, |
365 |
|
|
"Looking for the Text section"); |
366 |
|
|
if (!text_sec) { |
367 |
|
|
psiconv_warn(lev+2,header->section_table_offset, |
368 |
|
|
"Text section not found in the section table"); |
369 |
|
|
(*result)->paragraphs = NULL; |
370 |
|
|
res = -1; |
371 |
|
|
} else { |
372 |
|
|
psiconv_debug(lev+2,header->section_table_offset, |
373 |
|
|
"Text section at offset %08x",text_sec); |
374 |
|
|
res |= psiconv_parse_text_section(buf,lev+2,text_sec,NULL, |
375 |
|
|
&(*result)->paragraphs); |
376 |
|
|
} |
377 |
|
|
|
378 |
|
|
if (((*result)->paragraphs) && ((*result)->styles_sec)) { |
379 |
|
|
psiconv_progress(lev+2,header->section_table_offset, |
380 |
|
|
"Looking for the Layout section"); |
381 |
|
|
if (!layout_sec) { |
382 |
|
|
psiconv_debug(lev+2,header->section_table_offset, |
383 |
|
|
"Layout section not found in the section table"); |
384 |
|
|
res = -1; |
385 |
|
|
} else { |
386 |
|
|
psiconv_debug(lev+2,header->section_table_offset, |
387 |
|
|
"Layout section at offset %08x",layout_sec); |
388 |
|
|
res |= psiconv_parse_styled_layout_section(buf,lev+2,layout_sec,NULL, |
389 |
|
|
(*result)->paragraphs, |
390 |
|
|
(*result)->styles_sec); |
391 |
|
|
} |
392 |
|
|
} else |
393 |
|
|
psiconv_debug(lev+2,header->section_table_offset, |
394 |
|
|
"Skipping search for Layout section, as either the " |
395 |
|
|
"text or the word styles section was not found"); |
396 |
|
|
|
397 |
|
|
psiconv_free_application_id_section(appl_id); |
398 |
|
|
psiconv_free_header_section(header); |
399 |
|
|
psiconv_free_section_table_section(table); |
400 |
|
|
|
401 |
|
|
psiconv_progress(lev+1,off,"End of word file"); |
402 |
|
|
return res; |
403 |
|
|
} |