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
|
1.08.20090918
=============
* User-visible changes:
- Correctly locates engine translations under Debian
- Fix crash when quitting under ms woe
* Developer notes:
- Fixed slight memory leak in font lookup
- Updated bundled DLLs
1.08.20090916
=============
* User-visible changes:
- NSIS (NullSoft Install System) script to generate a complete ms woe
installer (with FreeDink, DFArc and freedink-data)
- Toggle fullscreen works on more platforms (not just X11)
- New architecture: PSP (playstation portable)
- New logging system, enabled with '-d' (instead of writing debug
information on standard output by default), with different
priorities (debug/info/warn/error/fatal). You also can easily check
for DinkC errors in your scripts by searching for "[DinkC]" in
DEBUG.TXT. In addition DinkC errors often carry the line number
(rather than the binary offset).
- Dink doesn't crash on very long lines in dink.ini or DinkC scripts
- DinkC: map_tile(...) can modify the first tile square (tile_index=0)
- Lower startup RAM requirements (-7MB so far, and less memory
bandwidth for sounds); this is mostly useful for small devices.
- Add support for contextual translations: when there are several
lines of dialog that are identical in a few scripts, but need to be
translated differently, you can add a context in your PO file such
as 'msgctxt "s3-chick:57"' (name of the script in lowercase without
extension + the line number).
* Developer notes:
- Dropped libffi dependency to support more architectures
- Fixed SDL_ttf version check (it failed for macport's 2.1.0)
- Code clean-ups, namely in the input subsystem (joystick) and
shutdown process.
- Fixed numerous buffer overflows in the DinkC interpreter, and a
potential infinite loop during search/replace (when the replacement
countained search term).
1.08.20090120
=============
* User-visible changes:
- Fix game freeze in Pilgrim Quest (return value for DinkC functions
'get_sprite_with_this_brain', 'get_rand_sprite_with_this_brain',
'get_next_sprite_with_this_brain' and 'sp')
* Developer notes:
- Added scripts to rebuild snapshots, .deb, .rpm and .exe unattended
1.08.20090109
=============
* User-visible changes:
- i18n ("internationalization") - or support for translating D-Mods.
- Fix some crashes due to improper D-Mod scripts (e.g. attempt to
modify sprite #1000): DinkC scripts are now better validated to
avoid this.
* Developer notes:
- DinkC functions are now called dynamically using libffi, for cleaner
bindings to the C engine core. The DinkC parser code was also
clarified and a few string functions were optimized.
1.08.20080920
=============
* User-visible changes:
- Supports playing Ogg Vorbis for background music (will look for .ogg
extension before .mid extension)
* Developer notes:
- Supports /usr/share/games alternate prefix for game data.
- Can search and use system fonts under FreeDesktop-compliant systems
(using fontconfig)
1.08.20080828: first official release
=====================================
FreeDink is based on version 1.07 of the original engine and also
incorporates v1.08 improvements.
The main change is that FreeDink is portable, so it can work under
operating systems (e.g. GNU/Linux) and architectures (e.g. PowerPC).
The following describe some differences with the original game engine.
* User-visible changes:
- Backward-incompatible changes in v1.08 can be disabled using a
'v1.07 mode' (--v1.07 option -- more information in doc/v1.08.txt).
This allows playing D-Mods that weren't upgraded to v1.08 yet.
- Game graphics
- Better windowed mode: supports non-truecolor mode, and can switch
between windowed and full-screen using Alt + Enter (X11 only).
- Avoid color/palette temporary glitch after quitting show_bmp()
mode.
- Smoother truecolor fades, and avoid some glitches in palette mode.
- Use a different but metric-compatible fonts (alignment and
wrapping are conserved), due to copyright issues (see
doc/fonts.txt).
- Widen debug-mode line to avoid some unwanted line wrapping.
- Gamepad: the threshold for joystick axis is slightly greater - with
the same threshold than the original game (~10%) Dink is going diags
too easily. Feedback welcome :)
- Game speed: set maximum framerate to 50 FPS, which is compatible
with v1.08 (see doc/framerate.txt)
- Installation: supports separate directories for original data and
additional D-Mods. This permits installing the game as
administrator in a write-protected directory, while allowing normal
users to install D-Mods in their personal directory.
- Memory usage
- Requires less memory in truecolor mode (by not converting paletted
graphics to truecolor)
- Fixes memory leaks, in particular when loading a savegame
(reloaded graphics were duplicated in memory)
- Editor fixes:
- To avoid issues with non-US keyboards in the editor, F1->F10-F11
and NumPad 0->9-. are used in addition to 1->0-` for tiles
selection.
- When vision is changed from the minimap mode in the editor, don't
display the current screen's sprites over the minimap.
- When text or number is asked in the editor, user now can press
'escape' to discard changes.
- In editor "Alternative Tile Hardness Selector", when going
completely to the right or to the bottom of the screen, avoid a
postponement for the square selector. Reported at
http://www.dinknetwork.com/forum.cgi?MID=57829
- When editing hardness, print the hardness index actually being
edited, instead of the one associated by the current background
tile.
- Display real-time information about hardness in hardness mode.
* Developer notes:
- Compressed BMP support: afaics in some situations, the original Dink
engines does not support compressed (RLE?) BMPs. Such BMPs are
notably used in "Prophecy of the Ancients" (like the 3 alternate
colored orbs) - when support is lacking, graphics are simply not
displayed which much complaining in DEBUG.TXT. I believe this
support depends on which DX version is used. More surprisingly:
there's a screen, still in PotA, where you can pass above and under
and bridge (you need to pass under the bridge to blow up a rock in
the above screen and thus reach the secret island). There's no way
to pass under the bridge when there's no compressed BMP support - I
haven't investigated much, but I guess either a missing sprite does
not clear the hardness for Dink to pass, either the missing sprite
is causing a DinkC script to stop. So, vanilla SDL does not support
compressed BMPs, so I used SDL_image everywhere to do so. This means
that FreeDink always support compressed BMPs.
- Fix several array-out-of-bounds issues while reproducing the
original de-facto engine behavior.
- For example, in 'init_scripts', the script may check the active
status of sprite 1000, while the sprite array has 300
elements. This caused bugs after the port and is now
fixed. Curiously it didn't seem to cause bugs before. Such bug did
'appear' in v1.08 too.
- Sometimes, a init() or dink.ini line would be more than 80 chars,
which is its allocated length in the 'sequence' structure. This
would overwrite parts of the next sequence. Possibly this never
was an issue, because sequences were initialized incrementally,
but an init() line was more likely to cause breakage. This is
fixed anyway.
- When reloading a sequence through init() with a longer sequence,
v1.07 will overwrite bits of the next sequence, for dir.ff loading
only. v1.08 overwrites for both BMPs and dir.ff sequences to avoid
to a bug with mixed dir.ff and BMPs sequences. This bug is
documented in dink.ini and loads the sword (longer) hit sequence
before the first (shorter) sequence to work-around it. See
http://www.dinknetwork.com/forum.cgi?MID=92733 for details. In
FreeDink, sequences are never overwritten.
- Some D-Mods have sequences with more than 50 frames. This triggers
a buffer overflow in the original game BUT the overwritten data is
the animation metadata ('delay' and 'special' fields), which is
not critical for non-animated sprite sequences. Example: "The
Green Voice in my Head - Part 1 - Hangover & Agony", first save
game then east then house on the right: it shows a shelf with
broken vases with is seq 66 frame 65 (> 50). FreeDink reproduces
this behavior by allowing up to 1000 frames in a sequence but
without the buffer overflow part.
- Command line: fails on invalid options. Supports undocumented
'-nomovie' for compatibility with DFArc2. Supports short aliases
(e.g. -window -> -w) and double-dash options (e.g. -window ->
--window).
- Search path: Dink search data in additional places to support
separating reference data and user D-Mods. Check
doc/directory_layout.txt for information.
- Dinkedit
- Dinkedit creates a custom sequence #1 but mark the 4th frame as
'-1'. This made FreeDinkEdit crash, so it's set to 0, which means
"end of sequence".
- The editor palette is now Ts01.bmp instead of esplash.bmp - just
like the game. This is simpler that way, and nobody changed the
editor palette anyway.
|