BlackBox Component Builder Forums

Unfortunately the center currently does not have any code signing certificate at its disposal. That's why the link cannot be resolved.

Josef

Did someone from oberoncore ever look at this bug?
Is it true that it is only related to wine? So far, I was not ale to
produce it under Windows.

- Josef

It seems to be a clipping/updating problem under wine. You can also produce pixel garbage if lines := 1. It looks like the problem occurs when the screen update is not yet finished when the next scroll event arrives, or something like that. Try scrolling multiple lines with a single mouse wheel turn...

I have no idea what may cause such a behavior.

- Josef

it seems that you are loading incompatible versions of some modules. If you have the sources, recompile them all. BlackBox 1.6 is a pretty old version, by the way, but it is the latest from oberon microsystems. Newer versions are available from the BlackBox Framework Center at http://blackboxframewo...

by setting a smaller stack size (Coroutines.SetStackSize) it should be possible to start a larger number of coroutines. golang starts with a size of only 4KB and increases that on demand. With the native Windows fibers being used for BlackBox coroutines the minimum stack size is somewhere at 64 KB, ...

@Ivan: can you please describe the issue with the multiLine Field Update?
The coding looks quite tricky, I mean the coding of the old code.
Therefore it is not obvious to me what the fix is all about.

- Josef

Does it mean that WinApi.SystemParametersInfoW is not supprted under wine?

What is the value in "lines" after calling it?
It should be 3 (the Windows default) or may be some user configured value.

What if you set lines := 3; under wine without calling WinApi.SystemParametersInfoW?

- Josef

@X512 Please have a look at http://blackboxframework.org/unstable/issue-%23197/blackbox-1.7.3-a1.1120.zip for the fix as you proposed. The list of changes is in https://redmine.blackboxframework.org/projects/blackbox/repository/revisions/93acb1d2fc68882b5f8382a78b3bbad0a8f88ab6/diff/ . Thanks again ...

I think I do understand now what is going on. In Kernel.ExecFinalizer there is an extra check for finalizing objects of unloaded modules. IF (fin # NIL) & (f.blk.tag.mod.refcnt >= 0) THEN fin(S.VAL(ANYPTR, S.ADR(f.blk.last))) END; If your test module A is unloaded, then the finalizer of the coro...