File: Bugs

package info (click to toggle)
gnurobbo 0.68+dfsg-4
  • links: PTS, VCS
  • area: main
  • in suites: bullseye, buster, sid
  • size: 5,212 kB
  • sloc: ansic: 21,128; sh: 301; makefile: 185; awk: 98
file content (200 lines) | stat: -rw-r--r-- 11,950 bytes parent folder | download | duplicates (5)
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
x = fixed
p = partially
- = not fixed

>>> neurocyp asked if we could use the SF bug tracker from now onwards,
	so please use that now to record and manage bugs <<<

[-] - The Original skin's robbo.ttf doesn't show 'ñ'. It's not a
	problem at the moment because there's no Spanish translation yet.
[X] - The designer appears to be remembering settings from previous
	editing sessions. Try this :-
	* Play the Designer level set and press F9.
	* Single-click the Save icon and it says "overwrite on save".
	* Single-click the New icon and it says "clear & insert on save".
	* Single-click again the New icon so that the level is cleared.
	* Double-click the "X" icon top-right and play another level/pack.
	* Press F9 to activate the designer.
	* Single-click the Save icon and it says "Click again to insert on
	save" i.e. the previous settings of the designer are not being
	cleared so that "insert" becomes the only option. Notice too that I
	haven't actually saved anything to file. * neurocyp: well, now,
	when you leave designer all settings like lastclick,insert on save
	are cleared. so this should be ok now
[x] - Not initialising game_timer_id to NULL and not checking that it
	is not NULL before removing it which causes a segfault when
	quitting early due to a file not being found.
[x] - I've just noticed that the window manager icon has disappeared.
	I think it's about time the wm_icon was set as per SDL docs i.e.
	before setting the video mode. This code really hasn't been touched
	since the start of the project.
[x] - k_icons is being ignored if not found i.e. like bgrnd.
[x] - bgrnd (if one was loaded) isn't being freed on exit.
[x] - k_icons isn't being initialised to NULL before load_selected_skin
	attempts to free non-NULL surfaces, nor is it being freed on exit.
[x] - Robbo2009 levels 10, 15, 16 and 17 crash the game on my GP2X.
[x] - Robbo98:12 crashes my GP2X. Maybe this will help with the above.
	UPDATE: I have noticed all the above levels have one thing in
	common: >> they all have force fields bordering the map edges <<
	UPDATE: 20091218-1208 I've found the bug and I'm going to attempt
	to fix it now. FIXED: It was a combination of missing boundary
	checks and referencing invalid array elements before checking that
	the index was valid. I tested east and west barriers and they work.
[x] - Original:6 and RobboVII:2 demos are now out of sync due to
	teleport modifications. They will require fixing at some point.
[x] - Sound: The Volume +/- message box is limiting the volume to 100
	and needs to go up to 128.
	[x] - It really should be updating Sfx Volume when Options is shown
		and HAVE_MUSIC hasn't been defined.
[x] - Is sound supposed to be enabled by default? "int sound" isn't
	being initialised anywhere so I don't know. There should be a
	"sound = 0;" or "sound = 1;" in game.c:195 to accompany volume.
[x] - Delete or rename ~/.gnurobborc so that the game starts from the
	beginning. Now go to Options and enable sound. You can hear the
	sfx but it says "0". Exit Options and play the game. You will hear
	loud sfx. Go back to Options and Sfx Volume is 0. If you press the
	+/- keys to see the volume messagebox you can see the real value of
	the volume (see following bug for possible answer).
[x] - The volume controls +/- do not synchronise with int sfx_vol.
	Setting sfx_vol updates volume, but setting volume does not update
	sfx_vol. Also in Options, using +/- to modify the volume does not
	update the Sfx Volume on-screen.
[x] - rcfile.save_frequency isn't being initialised at the start.
[x] - Disable Options: Primary Click for the GP2X.
[x] - Having the rcfile called .gnurobborc on the GP2X is not such a
		good idea because if Linux users look at their SD card they
		won't know it's there and they'll delete it. From now on look
		for gnurobborc first, then .gnurobborc and save gnurobborc.
[x] - Ok as Zaurus Confirm isn't much use to non SL-5500 users, so set
		Confirm by default to be something common like SPACE.
[x] - Fix incredibly fast Help Screen with redraw_everything always on
[x] - Fixed the problem with the colour cycling cycling too fast when
		holding down controls.
[x] - In Atari Robbo the bomb blows-up empty space but not in GNU Robbo.
[x] - Solid Lasers: In Atari Robbo the solid lasers are not blowable but
		in GNU Robbo they are. In Atari Robbo the solid lasers are left
		live when the gun is destroyed (same as GNU Robbo) which is a
		bit nonsensical. FIXED: Added an rcfile overrideable
		game_mechanic.sensible_solid_lasers; defaults to true.
[x] - Bears again ;) Fix bears of the same type going around one another
		forever. I think a courtesy protocol should make one of them
		wait for another DELAY_BEAR. The other bear can check his moved
		property and say hey, he's letting me pass, let's go. FIXED:
		It was easier to fix than expected, simply checking for empty
		space with another bear of the same type in the opposite corner.
[x] - Teleporting short distances should not feel sluggish.
[x] - Since forcing speed to 1 and using a new game_cycle_limit to
		control the game's speed, the Robbo-drawn-in-the-borders-when-
		teleporting-large-distances bug has reappeared. I think that
		the reason this wasn't happening was because the screen always
		finished scrolling before Robbo was being drawn in his new
		location. Obviously nothing should be being drawn off-board so
		this bug is likely to be in show_game_area(). In fact I have
		experienced this problem before when playing with deeper boards
		and this is to do with x/y_rest and mapoffsetx/y. Even so, care
		should be taken to limit the area sprites are being drawn to.
		Looking at show_game_area() now, I'd say it's due for a rewrite.
		I am going to study mapoffsetx/y now which is actually screen
		offsets and it looks as though part of the logic is in
		set_mapoffset() with the remainder in show_game_area().
		FIXED: I rewrote show_game_area() and implemented a new
		viewport.
[x] - There's a bug whereby Robbo is not killed if under a bird that is
		at the very top of the board i.e. where the wall mostly is. The
		level is RobboVII-008 at the top. FIXED: is_robbo_killed() bugs.
[x] - Bears are being weird, going around in loops one space apart. I
		think it's time I looked at the bear logic as they could benefit
		from not chasing their tails.
		BEAR: Follow left-hand rule maze traversal with start direction
		dictating their rotation around the board.
		BLACK_B: Follow right-hand rule maze traversal with start
		direction dictating their rotation around the board.
		In the original Robbo, both bears do nothing more than rotate
		whilst looking for exits when blocked. Original-054 (52 in
		Robbo VI) is a good place to study BEARs at the start, and also
		Original-014 (13) and Original-021 (20).
[x] - The bear can end up in a tight circle if blocked (same as
		original). It would be nice though to make it more intelligent.
		I created MECHANIC_SENSIBLE_BEARS which would be nice in Options.
[x] - The black bear can end up in a tight circle if blocked (same as
		original). It would be nice though to make it more intelligent.
		I created MECHANIC_SENSIBLE_BEARS which would be nice in Options.
[x] - Since tweaking the delays, RobboIX-003 is just too difficult and
		in fact impossible using the joystick on the GP2X. It needs one
		or two bombs at the bottom to be moved left to slow it slightly.
		The other beat-the-bomb levels are Original-020 and RobboVIII-016
		and they are fine.
[x] - Questionmarks seem to be empty for some reason. FIXED: It's
		because id_questionmark must be retained in create_object()
		and I was forcing a clear_field(). I have documented the code.
[x] - The rightward retreating solid laser is slightly faster than it
		should be. FIXED: The reason for this was that the solidlaser
		logic was modifying other objects and these objects also needed
		to be marked as processed. I watched the cycle_limit when these
		objects were being processed (DELAY_LASER is 4) and the
		cycle_limit started being dumped in units less than 4 which
		caused timing problems.
[x] - The right and leftward barriers are both slightly faster than
		they should be. FIXED: Same reason as above.
[x] - After fixing some timing issues in update_game() I found that
		the Help animations were slower than they should have been so
		I've fixed those.
[x] - The bears moving around the blocks on Original-044 go out of
		sync a little on the opposite side that they started from,
		eventually returning to sync again. The code responsible for
		this is in update_game() near the top (search for DELAY_BEAR).
		The same thing can be seen with the birds on Original-009. I
		have noticed that all the board locations (objects) are processed
		in an x/y loop (cols are processed, then rows), but what happens
		when an object is moved from one board location to another that
		has not been processed yet? Does it get processed again? This 
		would cause objects moving down and/or right to have their
		.moved delays decremented again at the top of the x/y loop!
		This needs debugging. FIXED: Indeed there was a bug so I have
		introduced another object property called processed that marks
		objects that have been processed already during that cycle. I
		also fixed another bug at the top of update_game() that
		decremented rotating objects twice. Just found another bug in
		there that caused objects to be slower than Robbo. It appears as
		though decrementing their delays but not processing them on zero
		within the same cycle caused Robbo to be a little faster than
		birds etc. that have the same delay values.
[x] - Delay parameters are not perfect :-
		GNU Robbo is attracted to magnets too slowly: DELAY_ROBBO works.
[x] - Delay parameters are not perfect :-
		Linked bombs blow slower in GNU Robbo: DELAY_BIGBOOM + 1 works.
		UPDATE: Since fixing some bugs in update_game() I have found
		that this should now be DELAY_BIGBOOM + 2 :)
[x] - I've noticed SDL_FreeSurface isn't called on any of the loaded
		images or created ones for that matter.
[x] - Syntax of level are volunerable for any inconsequences and errors.
		load_level_data() (previously load_from_file()) has been made
		more robust. Also level_init() has been tidied-up and I've made
		sure that the program can cope with failed level loads.
[x] - Force fields that kill Robbo don't go boom themselves. Thunor:
		same solution as blaster below.
[x] - Blasters that kill Robbo don't go boom themselves. Thunor: I
		simply added some returns after kill_robbo() to prevent objects
		from being created after everything has been blown-up.
[x] - On level 28 - RobboVIII occassionally killing a bear behaves like
		shooting a questionmark. I haven't seen this for some time now.
		I've seen this again. I think it might be a side-effect of
		changing level packs. I saw it whilst flipping between the end
		screen levels. I now think it might be related to what I was
		doing creating the end screen levels i.e. running incomplete or
		not-so-valid levels. This problem is currently happening on
		RobboVII level 8 with the ground underneath the blaster. It seems
		to occur only when navigating to the level via F5/F6 but I can't
		see why at the moment. FIXED! clear_field() was only zeroising
		id_questionmark if type=QUESTIONMARK. The problem is that an
		object of type QUESTIONMARK becomes an object of type BIG_BOOM
		when it is first destroyed and so update_game() checks for
		id_questionmark > 0 to identify a previous QUESTIONMARK object.
[x] - The restart_timeout needs clearing if Robbo dies and the user
		immediately changes the level or pack. I also cancelled existing
		fades when viewing options or help whilst playing.
[x] - Opening a door doesn't update the scoreline
[x] - If Robbo dies whilst firing then the firing continues
[x] - Continously shooting uses more bullets than fired
[x] - Fixed missing [author]s from original.dat levels 13 to 16