Re: [ECB-list] ecb-activate -> uninterruptable loop. . .

[<kla...@sd...>]

Henry S. Thompson wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>=20
> [anon] writes:
>=20
>> could you please explain me some more details why defining macros
>> "on the fly" during ecb-activate--impl???
>=20
> So, here's my analysis:
>=20
>  XEmacs 21.5 has changed save-window-excursion from a builtin special
>  form/subr to a macro.  That macro calls current-window-configuration
>  and set-window-configuration.  The ecb _advice_ to
>  current-window-configuration includes a call to
>  save-window-excursion.  The result is an infinite loop.
>=20
>  There are also similar problems with set-window-configuration calling
>  current-window-configuration, and restore-saved-window calling
>  split-window.
>=20
>  The best way I could see to fix these problems was to ensure that
>  those three (one macro, two functions) _always_ used the original
>  definitions of current-window-configuration and split-window.  I used
>  ecb-activate--impl as a point in the activation process when the
>  original definitions (ad-Orig-...) were available, but the ecb advice
>  hadn't yet been activated.
>=20
> As I said, doing this improves things considerably, but doesn't get us
> all the way -- there is still a problem, possibly an XEmacs problem,
> can't tell yet, somewhere in the area of window management.  The
> closest I can get to something odd is the following stack trace:
>=20
>   mark-marker(t t)
>   root-window->saved-window(#<window on #<INTERNAL OBJECT (XEmacs
>   bug?) (symbol-value-forward type 13) 0x61bc98> 0x13d753>)
>   root-window->saved-window(#<window 0x13a27c>)
>   root-window->saved-window(#<window 0x13d74b>)
>   ad-Orig-current-window-configuration() (window-configuration-equal
> configuration (ad-Orig-current-window-configuration))=20
>=20
> Both the 'XEmacs bug?' and the second arg't to mark-marker being 't'
> are very odd. . .

Thanks for your valuable work, Henry... i'm planning to release another =
release
of ECB during the next weeks. It would be good if we could ECB enable to =
run also with
XEmacs 21.5...

So my question: Did you get success for the problem above or what is =
current status?
Would you say, with your recent changes ECB could be ready for XEmacs =
21.5? Or still to
buggy??

Thanks.,
Klaus

>=20
> ht
> - --
>  Henry S. Thompson, HCRC Language Technology Group, University of
>                      Edinburgh Half-time member of W3C Team
>     2 Buccleuch Place, Edinburgh EH8 9LW, SCOTLAND -- (44) 131
>             650-4440 Fax: (44) 131 650-4587, e-mail: ht...@in...
>                    URL: http://www.ltg.ed.ac.uk/~ht/
> [mail really from me _always_ has this .sig -- mail without it is
> forged spam] -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.2.6 (GNU/Linux)
>=20
> iD8DBQFGijEukjnJixAXWBoRAoDZAJ0XOP/NoUPEYoWlPn6WwUq/7R9RBQCeLGWN
> +aIuxdj+F/TGSxtKDy/FIk0=3D
> =3DuMl3
> -----END PGP SIGNATURE-----