From kjsembed  Fri Oct 06 00:40:55 2006
From: Ian Reinhart Geiser <geiseri () yahoo ! com>
Date: Fri, 06 Oct 2006 00:40:55 +0000
To: kjsembed
Subject: Re: [kjsembed] eval returnvalue in
Message-Id: <200610052041.02861.geiseri () yahoo ! com>
X-MARC-Message: https://marc.info/?l=kjsembed&m=116010401526422
MIME-Version: 1
Content-Type: multipart/mixed; boundary="--===============1449568803=="

--===============1449568803==
Content-Type: multipart/signed;
  boundary="nextPart1457618.3DRKyjaIOS";
  protocol="application/pgp-signature";
  micalg=pgp-sha1
Content-Transfer-Encoding: 7bit

--nextPart1457618.3DRKyjaIOS
Content-Type: text/plain;
  charset="iso-8859-15"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

=46irst thanks for the patch.  Its close, but not quite.  With some minor=20
modifications it can handle anything handled by the metatype system. (Basic=
ly=20
you do not create the variant until the end, but only if the metatype is=20
handled as a QVariant.  Now this gets "funny" because anything that has a=20
metatype in theory can be handled as a variant.  So because of this we need=
=20
to check if the usertype is in fact covered by the range of QVariant.  At=20
that point we know its  a QVariant and we deal with it as such.  If it is n=
ot=20
a variant type then we need to play some further games with the meta object=
s=20
and see if we can QObject cast it.  This is were I am wandering in the=20
wilderness now.  Because of some other personal problems that have taken=20
priority this week I have not had enough time to finish this)

The patch is good though, and it may be an adequate stopgap measure until t=
he=20
full solution is in place.   Matt, Erik thoughts?

Cheers
	-ian reinhart geiser

On Thursday 05 October 2006 19:35, Sebastian Sauer wrote:
> hi * :)
>
> the attached patch tries to add support for handling of slot returnvalues
> at the QObject binding.
>
> * Unittest I used to test the returnvalues (bytearray does not work, how =
to
> define a QStringList in javascript?);
> http://websvn.kde.org/trunk/koffice/libs/kross/test/unittest.js?rev=3D586=
618&
>view=3Dmarkup
>
> * I didn't regressiontest it so much.
>
> * It may crash or do some other dirty things if e.g. a QVariant::UserType
> got passed.
>
> * Any feedback is welcome :)

--nextPart1457618.3DRKyjaIOS
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.2 (GNU/Linux)

iD8DBQBFJaYePy62TRm8dvgRAmfyAJwPzckpLKv7YphdNyh3cPYXTonTLwCg2vsB
3esB+a5lXhPLQfInbYoEBC4=
=hboo
-----END PGP SIGNATURE-----

--nextPart1457618.3DRKyjaIOS--


--===============1449568803==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

_______________________________________________
kjsembed mailing list
kjsembed@kde.org
https://mail.kde.org/mailman/listinfo/kjsembed

--===============1449568803==--