Trex-generated Show instances dodgy

--===============2941707021485464493==
Content-Type: multipart/alternative; boundary="00000000000089283a05e71b5681"

--00000000000089283a05e71b5681
Content-Type: text/plain; charset="UTF-8"

Strange: on loading this module, Hugs complains (this is standard-issue Sep
2006 Hugs in Hugsmode)

> - An instance of ShowRecRow (x :: a, y :: a) is required to derive Show
(NPoint a)

Just loading it again and the complaint disappears.

> module NPoint  where
>
>     import Hugs.Trex
>
>     newtype Num a => NPoint a  = MkNPoint ( Rec( x :: a, y :: a))
 deriving (Eq, Show)
>

If I ask for :info ShowRecRow after the failed load, I see an instance each
for label x, y -- as well as the EmptyRow instance (standard with
Hugs.Trex). If I :load Hugs.Trex after the failed load, those two instances
have gone; I can load NPoint ok.

On firing up Hugs, if I :load Hugs.Trex before :load NPoint, the error
still happens. But if I

>print (z = 5)                                -- some arbitrary label

then :load NPrint, all is fine. (And there's no ghost ShowRecRow instance
for z.)

If I 'clear modules' after the failed load; then :load NPoint, no error.

Same for an ordinary data decl wrapping a Trex record, with deriving. Same
if I include some arbitrary `silly = show (silly = "shilly")`.

AntC

--00000000000089283a05e71b5681
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr">Strange: on loading this module, Hugs complains (this is s=
tandard-issue Sep 2006 Hugs in Hugsmode)<div><br></div><div>&gt;=C2=A0- An =
instance of ShowRecRow (x :: a, y :: a) is required to derive Show (NPoint =
a)</div><div><br></div><div>Just loading it again and the complaint disappe=
ars.</div><div><br></div><div>&gt; module NPoint =C2=A0where</div>&gt;<br>&=
gt;=C2=A0 =C2=A0 =C2=A0import Hugs.Trex<br>&gt;<br>&gt;=C2=A0 =C2=A0 =C2=A0=
newtype Num a =3D&gt; NPoint a =C2=A0=3D MkNPoint ( Rec( x :: a, y :: a)) =
=C2=A0deriving (Eq, Show)<br>&gt;<div><br></div><div>If I ask for :info Sho=
wRecRow after the failed load, I see an instance each for label x, y -- as =
well as the EmptyRow instance (standard with Hugs.Trex). If I :load Hugs.Tr=
ex=C2=A0after the failed load, those two instances have gone; I can load NP=
oint ok.</div><div><br></div><div>On firing up Hugs, if I :load Hugs.Trex=
=C2=A0before :load NPoint, the error still happens. But if I=C2=A0</div><di=
v><br></div><div>&gt;print (z =3D 5)=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=
=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 -=
- some arbitrary label</div><div><br></div><div>then :load NPrint, all is f=
ine. (And there&#39;s no ghost ShowRecRow instance for z.)</div><div><br></=
div><div>If I &#39;clear modules&#39; after the failed load; then :load NPo=
int, no error.<br></div><div><br></div><div>Same for an ordinary data decl =
wrapping a Trex record, with deriving. Same if I include some arbitrary `si=
lly =3D show (silly =3D &quot;shilly&quot;)`.</div><div><br></div><div><div=
>AntC</div><br class=3D"gmail-Apple-interchange-newline"></div></div>

--00000000000089283a05e71b5681--

--===============2941707021485464493==
Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: base64
Content-Disposition: inline

X19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX18KSHVncy1CdWdz
IG1haWxpbmcgbGlzdApIdWdzLUJ1Z3NAaGFza2VsbC5vcmcKaHR0cDovL21haWwuaGFza2VsbC5v
cmcvY2dpLWJpbi9tYWlsbWFuL2xpc3RpbmZvL2h1Z3MtYnVncw==

--===============2941707021485464493==--