Version 0.5 preview

[sleepless]

The register state display is quite nice visually, but I have 2 gripes with it: It does not display MMC5 channels, and the grayscale bars displaying pitch have higher notes toward the left, which is the opposite of what I'm used to seeing in Impulse Tracker 2.

[Drillimation]

I can't even open it because it says there's a virus in it.

[garvalf]

I've exported a song to .nes format with 5beta8, and all I get is a high pitched noise. If I export to text so I can load it into 4.2, then when I export to .nes from FT4.2, the .nes file is running fine.

[HertzDevil]

Loading any FTM created since 0.3.5 stable with more than 128 VRC6 sequences will crash the tracker:

Code: Select all

bool CFamiTrackerDoc::ReadBlock_SequencesVRC6(CDocumentFile *pDocFile)
{
   int Version = pDocFile->GetBlockVersion();
   unsigned int Count = pDocFile->GetBlockInt();
   ASSERT_FILE_DATA(Count < (MAX_SEQUENCES * SEQ_COUNT));       // 639 (actually should be <=)
   if (Version < 4) { /* ... */ }                               // 0.3.5 beta 0 or before
   else {                                                       // since 0.3.5 stable
      int Indices[MAX_SEQUENCES];                               // 128
      int Types[MAX_SEQUENCES];                                 // 128
      // ...
      for (unsigned int i = 0; i < Count; ++i) {
         unsigned int Index     = pDocFile->GetBlockInt();
         unsigned int Type     = pDocFile->GetBlockInt();
         // ...
         Indices[i] = Index;                                    // access violation on i >= 128
         Types[i] = Type;
// ...


This is not fixed in 0.5.0 beta 5 yet. (Version 0.4.2 and before hid this by asserting that Count < MAX_SEQUENCES, so that each expansion chip may only hold 127 sequences, but that was incorrect as well as long as FamiTracker intends to allow 128 sequences in all 5 sequence types.)

[Stratelier]

Quick question - does hitting F9 (with no modifier) still set the current octave to 7? Because I'm on a laptop whose keyboard does not have its function keys nicely grouped by fours. Me fat-fingering the F8/9 keys happens at least once every session, and it seriously wrecks my mood when I'm unexpectedly calling dogs instead of tracking notes. I don't mind that Ctrl+Fkeys are mapped for setting the current octave, but the Fkeys by themselves....

Also, I'm starting to think that the way FT always assigns file associations upon startup may be a factor in why we have so many unmarked beta .ftms posted in the music sections.

[HertzDevil]

I've exported a song to .nes format with 5beta8, and all I get is a high pitched noise. If I export to text so I can load it into 4.2, then when I export to .nes from FT4.2, the .nes file is running fine.

Probably due to this.

...FT always assigns file associations upon startup...

In fact it is already like that since 0.3.7. Here is the code that registers the file extension in 0.4.6:

Code: Select all

BOOL CFamiTrackerApp::InitInstance()
{
   // ...
   // Skip this if in wip/beta mode
#if /*!defined(WIP) &&*/ !defined(_DEBUG)
   // Add shell options
   RegisterShellFileTypes(TRUE);
   // Add an option to play files
   // ...
#endif

These versions block debug builds rather than beta builds. In contrast here is the same part in the 0.3.6 source code:

Code: Select all

   // Skip this if in wip/beta mode
#ifndef WIP
   RegisterShellFileTypes(TRUE);
#endif

EDIT: When loading an FTM whose vibrato style is different from the current one, the tracker continues to use the old vibrato amplitudes. Following these steps could replicate the bug since 0.3.5 (the first version to use different vibrato styles):

1. Launch FamiTracker.
2. Create a blank instrument, then fill the pattern with:

   Code: Select all

        # : Pulse 1
ROW 00 : C-3 00 . ...
ROW 01 : ... .. . ...
ROW 02 : ... .. . 41F
3. Set the vibrato style to "Old style (bend up)". Play the module.
4. Create a new module and repeat the instrument and pattern there. Play the module.

Both vibratos have the same amplitude even though this should not be the case for the new style.

This is because, during loading a module, CFamiTrackerDoc::m_iVibratoStyle is modified directly without calling the setter method; here is a fix. (Actually the beta build might have already done this for the tuning settings.) The callback method is already called in the module properties dialog on finishing, so the vibrato tables will be regenerated as usual.

[garvalf]

I've exported a song to .nes format with 5beta8, and all I get is a high pitched noise. If I export to text so I can load it into 4.2, then when I export to .nes from FT4.2, the .nes file is running fine.

Probably due to this.

ok, thanks for the info, so we can expect it to be fixed in the final version...

[MrKyurem]

happy birthday 0.5 beta

here's to another year of no stable release

[TheMudkipMaster12]

And today marks the release of a new stable build of 0CC

[iYamWhatIYam]

jsr where the fuck are you jesus christ
we've had a ton of stable releases of 0cc-ft and not one of official ft
is this what you want from life