Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
Although it's a novel implementation (nobody has made a 4-player movie here before; we haven't yet even had a true 3-player movie either), I think I will have to vote "no", because there seems to be many obvious improvements that can be done.
Speedwise: Kill enemies earlier so they cause less lag. Avoid bumping against walls, use the characters more optimally.
Stylewise: Let each character act uniquely, don't group them together.
Still, I want to congratulate you on going ahead with that challenge.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
Thereisnospoon wrote:
how do I fix those warnings(or do I)?
By following what they say.
> warning: 'operator=' should return a reference to '*this'
Solution: Change your operator= so that it returns a reference to *this.
Instead of void operator=(const matrix& rhs) you should use: matrix& operator= (const matrix& rhs).
In case you are wondering what that function should return, it's return *this;.
> matrix.cpp:4: warning: 'matrix::array' should be initialized in the member initialization list
> matrix.cpp:4: warning: 'matrix::m' should be initialized in the member initialization list
> matrix.cpp:4: warning: 'matrix::n' should be initialized in the member initialization list
Solution: Instead of
matrix::matrix()
{
m = 0;
n = 0;
}
you should write:
matrix::matrix()
: array(0), m(0), n(0)
{
}
And so on.
The warnings help avert common design / implementation mistakes, so one should heed them.
In this case, the array(0) prevents your program from crashing if you do void foo() { matrix a; }. With your previous code, it would leave the "array" property uninitialized (i.e. its value would be random), and when the destructor runs, it would attempt to delete the memory pointed by that undefined pointer. It is however safe to attempt to delete with a null pointer, which is why you should set it to 0 if you haven't allocated anything.
Ps: Do not forget to deallocate the previous contents of the matrix in the assignment operator! Otherwise, your code has a resource leak.
PPs: Do not forget that "a = a;" is legal code. The "rhs" of the assignment operator might actually be the object itself. Verify that your program does not do anything unexpected if you do that.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
Your code has:
typedef double* doubleArrayPtr;
doubleArrayPtr* array;
In other words, you're defining array as double**, that is, a pointer to a pointer to a double. Is this really what you want?
Hmm. I guess it is, because you handle it properly.
Your code is missing the assignment operator: matrix& operator= (const matrix& ref);
It should be implemented in the same way as the copy constructor, which by the way seems to be allright (except for a missing semicolon before "for").
You would have noticed this if you had used the -Weffc++warning option. I recommend you add -Wall -W -Weffc++ to your Makefile in the CXXFLAGS.
Why did you put using namespace std; into matrix.h? There's nothing in matrix.h that uses anything from namespace std. You can use it in matrix.cpp if you like, but don't put it into the header file for it corrupts the namespace of everyone who uses that header.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
I'm locking this thread. Follow-up to the existing Mega Man 3 thread.
Duplicate threads for the same game are not encouraged.
Ps: This image demonstrates well the advantages of the "falling faster" technique. As you see, the difference made by an individual jump is very small, but doing it consistently accumulates into clearly visible differences.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
Cmuxmt wrote:
My belief that the voting system has anything to do with the publication of movies is now officially dead.
Majority of the votes for this movie were casted before the full speedrun was completed. Back then, based on those votes, it was decided that this movie will be delayed at least until the full movie is ready.
Some time after the full speedrun was released, this submission was rejudged as "accepting as a concept demo", and hence, I published it today.
I hope this answers your anxiosity.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
Zerotech wrote:
utilize a NEW Jump Trick effectively that allows Mega Man to jump higher
As BagOfMagicFood already pointed out, this second-controller-assisted debug feature is not allowed on this site for this game.
The first time it popped up on this site was March 2004 in this thread.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
NeonTiger wrote:
I guess the game has somekind of safe mechanism if you skip a plot important person somehow...
Rather, it seems like the game does not have that mechanism.
<hypothesis>When the first situation in the game occured when dialog required Celes to be present, it probably used a function that searches for the character slot where Celes is placed in. The function returned "no character found", but the caller did not verify the return value (because the programmers were believing it can't fail), and thus, it used whatever slot the function returned. In that slot, was still remaining the data that was placed there when Kutan was part of the party. Therefore, Kutan became the substitute Celes. Little by little, in different situations, more of the supposed aspects of Celes were assigned into Kutan (such as "visibility"), and ultimately in the Opera house, when Kutan's appearance was changed for the purpose of the opera, and later changed back ("back" meaning whatever Celes is supposed to look like), Kutan got Celes's appearance permanently on him, in the same way Locke has the merchant apperance on him permanently (until-if something happens that changes it again). But Kutan did not have Runic in his ability list, and there is no event that forces it into the list, and thus, he still can't do it. </hypothesis>
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
Catastrophe wrote:
I start the Ifrit battle [...]
This battle suffers from occasional desync when frame 440834 is delayed by a single frame and becomes 440835 and I'm not sure why. Once it desynched by two frames and I saved it by hex-editing.
The movie seems to desync for everyone here.
By doing what you described, you actually made it to desync. Probably, you first loaded a wrong save, and thus, observed that the movie desynced, and you attempted repairing it by replaying from that point, but instead, you actually broke it.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
Genisto’s Princess-only run was originally created because I pointed out that in my Super Mario Bros 2 movies, I never used Princess. So Genisto went ahead and created his movie. Which happened to also obsolete my Mario+Luigi+Toad movie back then. Back then, it was novel.
Today, there is really no need for that movie like it was back then, and as everyone have already pointed out, it is technically inferior to all other Super Mario Bros 2 movies on this site.
However, I see no reason to remove it ― it would set an unpleasant precedent, because after I created the site engine about two years ago, I have never removed a published movie without replacing it with another.
In fact, I cannot even do that without destroying all traces that the movie ever existed from the site.
I would mark the movie obsoleted (by some other movie) if I had a good enough excuse doing so, but currently I do not have one. (Maybe someone could help this fact to change?)
With these circumstances, I am allowing submissions that attempt to improve that movie. However, even if they do beat it, they can not be inferior to other movies of the same game on this site. Therefore, I can not accept this submission.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
Meow.
In other words, I'm locking this thread, because it failed to pitch those hacks convincingly.
Specifically, it started the thread with two posts and failed to provide any useful information in either one of them. It also sported credibility-depriving language.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
Your posts are very insightful reading. You must have been doing FF6 ROM hacks or something at some point of time.
Hmm, a form of Celes appears in Zozo. Must be a Celes-looking object, because if it was Celes, it would look like Kutan :)
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
Bisqwit wrote:
EDIT 3: Surprisingly, this was a good thing. It saved 4 additional frames. However, now it desyncs in the second BisqBot-assisted scene. Time to fire up the bot again.
The bot didn't do as well as the last time, shrinking the 19-frame gain into mere 3 frames.
However, there were other improvements in Crashman stage, which combined resulted in Crashman's door entered 31 frames earlier than in the previous WIP.
I tried fighting the boss, but I lost there 5 frames compared to Shinryuu's version. So it's Shinryuu's turn now.
After he's done, it's time to redo Quickman and Metalman.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
Bisqwit wrote:
I believe we could use this technique twice in Woodman stage, saving 20-40 frames.
The first place is out of question. Having three Items on screen in the push-up phase causes more lag (because of the dog's fireballs) than the trick would possibly save, and there's no way to do it without the first Item.
Checking the second place.
EDIT: The second place is a success, and it savessaved 15 frames.
EDIT 2: Now it desyncs in Crashman stage _after_ the first horizontal scrolling.
Details on the desync:
This is mystical.
[img_left]http://bisqwit.iki.fi/kala/snap/rm2cdes1.png[/img_left]This is the old movie.
Play #32310 - X=295.25+0.000 Y=84.00-0.250
Frame 32310: 1 on for 0
Play #32311 - X=295.25+0.000 Y=84.00+4.871
Play #32312 - X=296.56+1.312 Y=79.13+4.621
Play #32313 - X=297.88+1.312 Y=74.51+4.371
Play #32314 - X=299.19+1.312 Y=70.14+4.121
Frame 32314: 2 on for 0
Play #32315 - X=300.50+1.312 Y=66.02+3.871
Play #32316 - X=301.81+1.312 Y=62.14+3.621 (this is the frame shown in the image)
Play #32317 - X=303.12+1.312 Y=58.52+3.371
[img_left]http://bisqwit.iki.fi/kala/snap/rm2cdes2.png[/img_left]This is the new movie.
Play #32295 - X=295.25+0.000 Y=84.00-0.250
Frame 32295: 1 on for 0
Play #32296 - X=295.25+0.000 Y=84.00+4.871
Play #32297 - X=296.56+1.312 Y=79.13+4.621
Play #32298 - X=297.88+1.312 Y=74.51+4.371
Play #32299 - X=299.19+1.312 Y=70.14+4.121
Frame 32299: 2 on for 0
Play #32300 - X=300.50+1.312 Y=66.02+3.871
Play #32301 - X=301.81+1.312 Y=62.14+3.621 (this is the frame shown in image)
Play #32302 - X=303.12+1.312 Y=58.52+3.371
There is a 1 pixel difference between the heights of the Item1s in these images. Yet, the input are identical, and the coordinates of Rockman are identical. I have no idea why it happens.
EDIT 3: Surprisingly, this was a good thing. It saved 4 additional frames. However, now it desyncs in the second BisqBot-assisted scene. Time to fire up the bot again.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
I always use -slbegin and -slend to adjust the scanlines according to my whim anyway. :P
For example, with Rockman it's sometimes helpful to have full 240 scanlines even with NTSC, to be able to see Rockman even when he's outside "normal" screen boundaries.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
Luke wrote:
You can tell if this bug is happening: the text at the bottom of the screen (e.g. "Playing frame 11034") will be in the wrong position, either too high or too low.
Huh? Well that's hardly a problem. Besides, I'm using a different text rendering engine, as indicated in my patch changelog.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
GuanoBowl wrote:
crap! I thought I stopped my movie sooner than it was originally stopped. I believe the movie should be stopped at 12236 which was my intention. sorry.
And it should be ~10 or 11 seconds faster.
If you send me the fixed FCM, I'll update the submission.
Do not cancel this submission. Send me the fixed FCM.
[Edit: Oh well.]
