On Thu 19 Dec 2013 at 01:14:43 +0100, Rhialto wrote:
> Now the big question is: what to do about it?

I see that an attempt to provide patches to fix this was made, but it
was reverted because the fix wasn't correct.

Relevant comments from there were

    ISO C++ is explicit that deque requires a fully defined type, so it is
    not portable to embeded a deque in the same type.

and

    This commit breaks pan completely -- it dies during startup.
    Revert it. Bump PKGREVISION.

See here:
http://cvsweb.netbsd.org/bsdweb.cgi/pkgsrc/news/pan/patches/?hideattic=0#dirlist

It seems to change the deque embedded in the Entry into a pointer to
such a deque. See
http://cvsweb.netbsd.org/bsdweb.cgi/pkgsrc/news/pan/patches/Attic/patch-pan_general_log.h?rev=1.1&content-type=text/x-cvsweb-markup&hideattic=0

With that as starting point, can we provide a better fix?

I would naively think that it would need a deque<Entry *>, instead of a
deque<Entry> * (a double ended queue of pointers rather than a pointer
to a double ended queue), because even with a pointer to a deque, its
type parameter is still an incomplete type.

-Olaf.
-- 
___ Olaf 'Rhialto' Seibert  -- The Doctor: No, 'eureka' is Greek for
\X/ rhialto/at/xs4all.nl    -- 'this bath is too hot.'

Attachment: pgp2dFblq9Bzw.pgp
Description: PGP signature

_______________________________________________
Pan-users mailing list
Pan-users@nongnu.org
https://lists.nongnu.org/mailman/listinfo/pan-users

Reply via email to