Some of you may have noticed that a lot has been broken by recent revisions. But of course a lot has been fixed as well. This creates a conflict in some games where you're not sure what revision to use since newer revisions will replace some of the issues you had with older revisions with different issues. I believe that if the devs just stopped fixing issue on googlecode, made a list of every issue in newer revisions that wasn't in 2.0, and then fixed every problem on that list we would have a nearly perfect dolphin revision, all of the good and none of the bad. Obviously we would still have a lot of issues since many issues have never been fixed but I'm not talking about those, I'm just talking about the ones that have popped up recently. Obviously this is never going to happen since dolphin's development style/method is quite different from this. However we can still make it happen.
Many of these problems are well known, have been around for hundreds of revisions, and are already reported on googlecode. Why are they still not fixed then? Simply because no one has taken the time to track down the issue to the single revision that caused it. I have done this before, and I won't lie, it is a pain in the ass and very time consuming. But it must be done if we ever want to fix these problems.
So since many of you want to help dolphin grow but don't know how and can't code I hereby give you another option (besides posting issues on googlecode of course). Either:
1. Find an issue with a game that is present in the current revision but not present in 2.0 (r5350) and post it here
or
2. Pick one of the issues on this list and track it down to a particular revision. Then post the issue and which revision caused it here.
I know that what I am asking is annoying and time consuming. But I think you'll agree with me that the results will be worth it.
How do I do this? First you test both the latest revision and 2.0 and make sure that the issue is present in the latest revision but not 2.0. LOW PERFORMANCE DOES NOT COUNT AS AN ISSUE. FIFO ISSUES DO NOT COUNT EITHER (since it is being worked on). Then you test a revision halfway in between. Then keep cutting it in half. So let's say for example the issue exists in 6500 but not in 5350. I would then test 5900. If the issue exists in 5900 then you know it was a revision between 5900-6500 that broke it. So then you test halfway between them, 6200. And so on. Make a list as you do this so you don't lost track. Just number the revisions you testing and write "PROBLEM" or "NO PROBLEM" next to them until you get it down to a specific revision.
Some issues we already know about that come to mind:
AR Codes broken (broken inbetween 5350 and 6331)
Non-fifo related stuttering (likely caused by changes to the shader caching system, framerate and vps do not drop when it happens, noticable in NSMB world 2-3, happened between 6000 and 6500)
I'm incredible tired right now so I'll add some more later when I can think clearly. Unfortunately we've gotten used to a lot of these problems
Many of these problems are well known, have been around for hundreds of revisions, and are already reported on googlecode. Why are they still not fixed then? Simply because no one has taken the time to track down the issue to the single revision that caused it. I have done this before, and I won't lie, it is a pain in the ass and very time consuming. But it must be done if we ever want to fix these problems.
So since many of you want to help dolphin grow but don't know how and can't code I hereby give you another option (besides posting issues on googlecode of course). Either:
1. Find an issue with a game that is present in the current revision but not present in 2.0 (r5350) and post it here
or
2. Pick one of the issues on this list and track it down to a particular revision. Then post the issue and which revision caused it here.
I know that what I am asking is annoying and time consuming. But I think you'll agree with me that the results will be worth it.
How do I do this? First you test both the latest revision and 2.0 and make sure that the issue is present in the latest revision but not 2.0. LOW PERFORMANCE DOES NOT COUNT AS AN ISSUE. FIFO ISSUES DO NOT COUNT EITHER (since it is being worked on). Then you test a revision halfway in between. Then keep cutting it in half. So let's say for example the issue exists in 6500 but not in 5350. I would then test 5900. If the issue exists in 5900 then you know it was a revision between 5900-6500 that broke it. So then you test halfway between them, 6200. And so on. Make a list as you do this so you don't lost track. Just number the revisions you testing and write "PROBLEM" or "NO PROBLEM" next to them until you get it down to a specific revision.
Some issues we already know about that come to mind:
AR Codes broken (broken inbetween 5350 and 6331)
Non-fifo related stuttering (likely caused by changes to the shader caching system, framerate and vps do not drop when it happens, noticable in NSMB world 2-3, happened between 6000 and 6500)
I'm incredible tired right now so I'll add some more later when I can think clearly. Unfortunately we've gotten used to a lot of these problems
"Normally if given a choice between doing something and nothing, I’d choose to do nothing. But I would do something if it helps someone else do nothing. I’d work all night if it meant nothing got done."
-Ron Swanson
"I shall be a good politician, even if it kills me. Or if it kills anyone else for that matter. "
-Mark Antony
-Ron Swanson
"I shall be a good politician, even if it kills me. Or if it kills anyone else for that matter. "
-Mark Antony