Full Version:
OpenMPT v1.17.02.52
Has the bug occured in previous versions? If yes, please specify version(s): (This option is optional, and doesn't need research. But in case you know...)
I am not sure, but it may happen in the newer versions as well.
Description of the bug:
When using replace, if I try to replace an instrument with either 252 or 253, it is not replaced with that instrument. 252 = instrument minus 1 and 253 = instrument plus 1. All of the other instruments work just fine, however.
Another problem: My incorrect insrument is 121. When I change instrument 121 into instrument 250 and then do a replace with Instrument plus 1 twice, my 251s that are correct are also changed even with the Only in channels box checked and the range set to the proper channel (in my case, 19 to 0) (The correct instruments are in channel 26.)
How often does it happen?:
Always
How to reproduce (step by step description):
Step 1: Open the Find dialog.
Step 2: Click on the Instrument checkbox.
Step 3 (recommendation, but not required): Click the Find in the whole song radio button.
Step 4: Assuming that you have done everything in the Find tab, click the Replace tab, click the Replace By: checkbox and change the instrument combo box to either 252 or 253 (checking the Instrument checkbox again, of course).
My recommendation: Either:
1. Fix the bug (in this version) (without decreasing the sequence limit) or
2. Extend the sequence limit again (as the bug does not appear to exist in version 1.17.03.01, however, I cannot open the song in which I need to fix the problem because it has 16,974 patterns!)
Now you see what I am talking about all the time - Extending the limits to some big values that were not intended when the program was written are never a good idea, if you do it *something* will not work.
Hang on, I am going to investigate this a little bit further. Per my further investigation, I discovered that version 1.17.02.50 has the same problem. I tried 1.17.02.48 and 1.17.02.47, but neither of them even opened the song, although, I did not actually open the song to do the test, so I guess that I could have done it still. Rather than doing it though, I am going to go out on a limb and say that those vesions likely have that problem as well. And that was caused by adding the OpenMPT format? Hang on again, even further investigation is under way! Per further investigation, I have discovered that version 1.17.02.45 (just prior to the OpenMPT format's existance) also has the instrument 252/253 problem. That leads me to believe that the OpenMPT format (as well as the extended sequence limit) is indeed not the problem, but rather, perhaps faulty code somewhere else. Sorry to burst your bubble once again, Jojo. I am also if this post sounds either offensive or rude to you.
It's simply because the instrument count has been increased from 200 (MPT 1.16) to 256.
Actually, I just read the last line of your first post and it indeed doesn't happen with the current SVN code. We agreed on only fixing severe bugs in the 1.17.02 branch, so I will close this report if Relabs agrees that it's not worth fixing. Because, you know, we actually prefer working on the 1.17.03 branch. :)
OK, I was unaware of that (and by the way, it is actually 255, not 256). This one is a tad bit more challenging to fight logically. Based on the wording, however, it would appear that the problem (according to you) is actually the fact that the instrument quantity was increased. The bug does not seem to exist in the newest version (though it does exists in verions 1.17.02.53 and 1.17.02.54), which indicates that you found it wherever it was and corrected it. Perhaps what you could do in that case is increse the sequence limt again (as well as the channel limit back to 128 so it is not odd) and test the program to see if the bug comes back. If it does, then remove the code. If it doesn't, then keep the code. It seems that simple to me.
Actually, the bug is still there, but due to the addition of a few new "special" notes, it's not 252 and 253 anymore, but 247 and 248 or something. The fact that this bug exists has actually nothing to do with the number of instruments (and it has not been fixed in 1.17.03.01), but when the number of instruments was increased, a few constants were conflicting with the number of instruments. That is the stuff that always breaks when any limits are increased. In this case, using the different constants for different limits would not have caused the bug, but the same constants were used for note and instrument limit (which is bad style of course), so it broke.
Oh, I did not know that. I did discover, however, that the Only in channels part does still exist. And, you were correct. It is indeed instrumens 247 and 248. Perhaps that bug is unfixable then. I am not sure.
Please confirm that this is fixed in OpenMPT 1.17.03.02. You may set the S=C flag in the first post yourself then, without further comments.
Hmm, I see that the next version's name is there, but it is not a hyperlink. Perhaps you can note that for future reference, Jojo. By the way, where is the file located. I looked at SourceForge's index and it did not appear to be there. I also looked at the german site and the Software Downloads section of the forum, but those did not seem to help either.
It is in the development forum (http://forum.openmpt.org/index.php?topic=3432.0).
I did not think to look there. Thanks
It would appear that that bug has been fixed (?) (The reason that the question mark is there is because it may have actually transferred to different instruments once again, but I am not sure of this.)
Well, the instruments or number of instruments didn't matter. I'll close it as you can confirm what I have suspected to be the reason of fail at least. :)