Fix beat clips being created with a lenght of 0 in the pattern editor

[Itreza2]

closes #8101
The context and motivations for this pr can be found in the issue's thread

[bratpeki]

Could we get a little explanation as to why this works? It'd help with the code review!

[bratpeki]

I asked GH Copilot and got this, so feel free to expand on this.

---

Issue #8101 is about the pattern editor showing blank tracks (no steps) for newly created sample tracks. The error messages you're seeing (called Track::getClip(X), but Clip X doesn't exist) indicate that the code is trying to access clips that haven't been created yet.

The original init() method unconditionally calls:

saveJournallingState(false);
updateLength();
restoreJournallingState();

This happens for all MidiClip instances, including those created in the PatternStore (the pattern editor's internal container).

When a new MidiClip is created in the pattern store:

1. updateLength() calls updatePatternTrack()
2. updatePatternTrack() triggers pattern regeneration
3. During this process, other tracks try to access clips via getClip() that don't exist yet
4. The journalling state save/restore causes issues with incomplete clip initialization

The PR wraps the problematic code in a conditional check:

if (getTrack()->trackContainer() != Engine::patternStore())
{
	saveJournallingState(false);
	updateLength();
	restoreJournallingState();
}

This means:

• For regular tracks (song editor): Process normally with journalling
• For pattern store tracks (pattern editor): Skip the journalling and length update

By skipping these operations for pattern store clips, the code avoids triggering clip access during initialization, preventing the "Clip doesn't exist" errors. The pattern store handles its own initialization separately, so this code path isn't needed there anyway.

This is why the PR works! It prevents the premature clip access that was causing blank pattern editor tracks.

[Itreza2]

Copilot (almost) got it perfectly right !

As I detailed in my original comment on #8101, updateLenght()'s call chain is in fine trying to access clips that don't yet exist and this is what is causing the bug.
Fortunately, updateLenght() is completely superfluous for tracks in the pattern store, so adding a if clause is an easy fix.

(And no, this PR should not affect journalling in any way, I think copilot hallucinated this point)

[bratpeki]

Thanks for the awesome explanation!

[bratpeki]

Approved for code and testing. Will merge soon.

[regulus79]

I tested this, and it seems to work fine.

[bratpeki]

Okay, testing now!