Re: [freesci-develop] Release Candidate 2: Please test

[Christoph Reichenbach]

Robert,

On Sun, Feb 24, 2008 at 04:40:46PM -0600, Robert Wenner wrote:
> Sorry for the late email...

  No worries!

> freesci builds, but there are some warnings, mostly about comparisons 
> being always true or false due to limited range of data types involved.
> (Do you want the compiler output?)

  These are known issues that we can all reproduce (the warnings about
`limited ranges' are merely due to the compiler trying to outsmart our
built-in sanity checks, so they are and nothing to worry about.)

> make check seemed to run only one test, is that ok?

  We have only very recently begun incorporating systematic unit
tests.  Technically speaking, this is not OK, but there's nothing we
can do about it other than to increase our test coverage ;->

> I tried running some games, and here's the results:
[...]

  That's a very substantial amount of testing; thank you very much!

> Quest for Glory 2:
> 
> Sometimes freezes when I try to import my saved character from part
> 1.

  To the best of my knowledge, you are the first person to test this.
Can you submit a bug report to our bugzilla (freesci.org, then go to
`Bug Tracker') and attach your saved character, please?  That way we
can try to reproduce the problem.

[...]

> At this point the game froze but more underruns happened.

  I'll have to look into these underruns in more detail at one point.
Did you have audio problems (cracking noises, gaps in the music)?

> First, it doesn't load the file.
> When it asked me if I want to try importing again and I click Yes, it 
> dumps me to a console, with the game window is frozen:
[...]

  That's our internal debugger.  We'll have to attempt to reproduce
this problem ourselves to figure out exactly what is going wrong,
because these games can be rather complex internally :-)

[LSL3]
> FSCI: kNOP: Warning (kernel.c L737): Warning: Kernel function 0x71 
> invoked: unmapped

  These are usually harmless but occasionally useful to us.  As long
as we're unstable, they will probably stay in.

> SQ3:
> Seg faults during intro before the glass slides away from Roger's face.
> Here is the last messages on the shell. 
> The core dump is at
> http://de.geocities.com/robertwenner/core-sq3.zip

  Your core file does not appear to include debug symbols (or
otherwise by gdb is confused).  Could you try re-running this, and
this time run with 0.6.3 compiled via

  CFLAGS="-g -O0" ./configure && make

please?  This will give a noticeably slower executable, but I should
be able to figure out what I need from there.

> In the intro, a blond guy chases another guy to a street lamp. (Shouldn't 
> they run further?)
> That blond guy (and his shadow) are still in the next intro screen with 
> the night skyline.
> Doesn't look like that's correct...

  This is the same as #2837, I think.

> PQ3:
> When the officer in the introduction speaks, the text is cut off on the 
> right (probably too wide for the screen).

  Added as #2838.  I don't have PQ3, so someone else might want to
check that some time.

[...]

> SQ4:
> Leaves me with a black game window and a console prompt:

[...]

  Currently, only the disk version of SQ4 works.  This is a known issue.

> KQ4:
> 
> Intro is fine.

  It was pretty broken just two weeks ago, so I'm glad to hear that ;-)

> Sometimes in scenes like Rosella cleaning up the dwarves' house, it seems 
> the game hogs all the processor resources (admittedly, this is an old 
> machine). The game briefly freezes and the console has lots of these 
> messages:
> Looping song iterator 00230014.
> and sometimes 
> Looping song iterator 001a0014.

  This may be due to recent changes I made to the song iterators.
Submitted as #2839.

> It crashed when changing screens, the core dump is at 
> http://de.geocities.com/robertwenner/core-kq4.zip

  For some reason I can't download that one...

  If possible and convenient, please try to reproduce this with 0.6.3
built as described above, and feel free to e-mail the core files to me
directly.


  To the best of my knowledge, none of our recent changes will affect
any of the issues you reported.

  Again, thank you very much for your thorough testing, this is
greatly appreciated!

-- Christoph