MIME-Version: 1.0
References: <CA+HiwqGGJtcVP=EGP0hH1Q6EqJE99KnOQvibC18QnepsWVu9fQ@mail.gmail.com>
 <202310061001.ndbl6o4dl3zh@alvherre.pgsql>
In-Reply-To: <202310061001.ndbl6o4dl3zh@alvherre.pgsql>
From: Amit Langote <amitlangote09@gmail.com>
Date: Fri, 6 Oct 2023 20:27:12 +0900
Message-ID: <CA+HiwqF6ekhOrg704xz3==aeizF2dRpEnGiBZkoU+wSj93R6BA@mail.gmail.com>
Subject: Re: remaining sql/json patches
To: Alvaro Herrera <alvherre@alvh.no-ip.org>
Cc: Andrew Dunstan <andrew@dunslane.net>, Erik Rijkers <er@xs4all.nl>, 
	PostgreSQL-development <pgsql-hackers@postgresql.org>, jian he <jian.universality@gmail.com>
Content-Type: multipart/alternative; boundary="00000000000048dd0306070a8609"
Archived-At: <https://www.postgresql.org/message-id/CA%2BHiwqF6ekhOrg704xz3%3D%3DaeizF2dRpEnGiBZkoU%2BwSj93R6BA%40mail.gmail.com>
Precedence: bulk

--00000000000048dd0306070a8609
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Fri, Oct 6, 2023 at 19:01 Alvaro Herrera <alvherre@alvh.no-ip.org> wrote=
:

> On 2023-Oct-06, Amit Langote wrote:
>
> > 2. Assignment of op->d.iocoerce.escontext needed to be changed like thi=
s:
> >
> >                         v_params[4] =3D
> l_ptr_const(op->d.iocoerce.escontext,
> > -
> > l_ptr(StructErrorSaveContext));
> > +                                                 l_ptr(StructNode));
>
> Oh, so you had to go back to using StructNode in order to get this
> fixed?  That's weird.  Is it just because InputFunctionCallSafe is
> defined to take fmNodePtr?  (I still fail to see that a pointer to
> ErrorSaveContext would differ in any material way from a pointer to
> Node).


The difference matters to LLVM=E2=80=99s type system, which considers Node =
to be a
type with 1 sub-type (struct member) and ErrorSaveContext with 4 sub-types.
It doesn=E2=80=99t seem to understand that both share the first member.


Another think I thought was weird is that it would only crash in LLVM5
> debug and not the other LLVM-enabled animals, but looking closer at the
> buildfarm results, I think that may have been only because you reverted
> too quickly, and phycodorus and petalura didn't actually run with
> 7fbc75b26ed8 before you reverted it.  Dragonet did make a run with it,
> but it's marked as "LLVM optimized" instead of "LLVM debug".  I suppose
> that must be making a difference.


AFAICS, only assert-enabled LLVM builds crash.

>

--00000000000048dd0306070a8609
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div>On Fri, Oct 6, 2023 at 19:01 Alvaro Herrera &lt;<a href=3D"mailto:alvh=
erre@alvh.no-ip.org">alvherre@alvh.no-ip.org</a>&gt; wrote:<br></div><div><=
div class=3D"gmail_quote"><blockquote class=3D"gmail_quote" style=3D"margin=
:0px 0px 0px 0.8ex;border-left-width:1px;border-left-style:solid;padding-le=
ft:1ex;border-left-color:rgb(204,204,204)" dir=3D"auto">On 2023-Oct-06, Ami=
t Langote wrote:<br>
<br>
&gt; 2. Assignment of op-&gt;d.iocoerce.escontext needed to be changed like=
 this:<br>
&gt; <br>
&gt;=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=A0v_params[4] =3D l_ptr_const(op-&gt;d.iocoerce.escontext=
,<br>
&gt; -<br>
&gt; l_ptr(StructErrorSaveContext));<br>
&gt; +=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 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=
=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0l_ptr(StructNode));<br>
<br>
Oh, so you had to go back to using StructNode in order to get this<br>
fixed?=C2=A0 That&#39;s weird.=C2=A0 Is it just because InputFunctionCallSa=
fe is<br>
defined to take fmNodePtr?=C2=A0 (I still fail to see that a pointer to<br>
ErrorSaveContext would differ in any material way from a pointer to<br>
Node).</blockquote><div dir=3D"auto"><br></div><div dir=3D"auto">The differ=
ence matters to LLVM=E2=80=99s type system, which considers Node to be a ty=
pe with 1 sub-type (struct member) and ErrorSaveContext with 4 sub-types. I=
t doesn=E2=80=99t seem to understand that both share the first member.</div=
><div dir=3D"auto"><br></div><div dir=3D"auto"><br></div><blockquote class=
=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-left-width:1px;bo=
rder-left-style:solid;padding-left:1ex;border-left-color:rgb(204,204,204)" =
dir=3D"auto">
Another think I thought was weird is that it would only crash in LLVM5<br>
debug and not the other LLVM-enabled animals, but looking closer at the<br>
buildfarm results, I think that may have been only because you reverted<br>
too quickly, and phycodorus and petalura didn&#39;t actually run with<br>
7fbc75b26ed8 before you reverted it.=C2=A0 Dragonet did make a run with it,=
<br>
but it&#39;s marked as &quot;LLVM optimized&quot; instead of &quot;LLVM deb=
ug&quot;.=C2=A0 I suppose<br>
that must be making a difference.</blockquote><div dir=3D"auto"><br></div><=
div dir=3D"auto">AFAICS, only assert-enabled LLVM builds crash.</div><block=
quote class=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-left-w=
idth:1px;border-left-style:solid;padding-left:1ex;border-left-color:rgb(204=
,204,204)" dir=3D"auto"></blockquote></div></div>

--00000000000048dd0306070a8609--