On 25.02.2012 13:59 (UTC+1), Paul Harris wrote:

On 25 February 2012 20:48, Rainer Hurling <[email protected]
<mailto:[email protected]>> wrote:

    On 25.02.2012 12:24 (UTC+1), Doug Barton wrote:

        Ok, I did the clang run, the log is here:

        http://people.freebsd.org/~__dougb/WindowMaker-clang-build.__log
        <http://people.freebsd.org/~dougb/WindowMaker-clang-build.log>

        Search for "warning" to see what it found. I created a patch for the
        super obvious ones:

        http://people.freebsd.org/~__dougb/patch-clang-warnings.txt
        <http://people.freebsd.org/~dougb/patch-clang-warnings.txt>

        Rainer, you might want to give this a try, in the half dozen
        attempts I
        ran it seems to make restoring the session work.

        I think the reason is the obvious typo in src/osdep_bsd.c. I can't
        necessarily tell you *why* this caused the problem (although I
        have a
        theory) but it fits the pattern of the bug report in that a)
        this file
        is new in 0.95, and b) it's FreeBSD specific.

        After applying the patch above the only warnings left are the
        several
        "implicit truncation from 'int' to bitfield" and one unused label in
        usermenu.c. The former look like they should probably be fixed,t
        but the
        right fix wasn't obvious to me. The unused label looks like it
        can just
        be removed, since that's the only use of the word "keyover"
        anywhere in
        the sources. What wasn't clear to me is how much of the rest of the
        block should go, so I left it alone.


    Doug,

    thank you very much for the clang run and the new patch. I removed
    your older patch, commented out the two lines in Makefile and put
    this patch under files. It patches, builds and installs fine.

    I am sorry, but there is no change in the behaviour of restoring old
    sessions. I am not able to restore sessions, which obviously were
    saved before in WMState.

    I am wondering, if the next step should be to debug the whole part
    (of restoring), beginning with a bp somewhere in src/startup.c? But
    I have almost no experiences with that.

    Rainer


I wonder if its something outside of wmaker that is causing the trouble.

I also thought about this possibility. At least for one of my boxes, because on two other boxes all is fine now. But I really have no idea.

I assume saving a session involves writing a file somewhere.
Can you confirm that the session is successfully written to disk, and
can be read back from disk?

WMState is located at ~/GNUstep/Defaults. It is definitely written when saving sessions.

I will give an example. After creating a fresh configuratio with wmaker.inst and starting wmaker I first cleared the session, then opened a xterm windows and leaved the session with saving. The difference in WMState is:

  333c333,345
  <   Applications = ();
  ---
  >   Applications = (
  >     {
  >       Dock = Dock;
  >       Name = xterm.XTerm;
  >       Shaded = No;
  >       Miniaturized = No;
  >       Command = "xterm";
  >       Geometry = "1081x709+239+327";
  >       Hidden = No;
  >       Workspace = System;
  >       ShortcutMask = 0;
  >     }
  >   );

Is it normal, that wmaker even touches the file WMState, when starting? The date entry of WMState changes on starting and leaving wmaker (without saving anything).

Rainer


--
To unsubscribe, send mail to [email protected].

Reply via email to