Date: Mon, 17 Jul 2023 10:13:48 +0100 From: Nuno Teixeira <eduardo@freebsd.org> To: Dimitry Andric <dim@freebsd.org> Cc: FreeBSD Mailing List <freebsd-ports@freebsd.org>, Po-Chuan Hsieh <sunpoet@freebsd.org> Subject: Re: www/node18 fails to build on aarch64/current Message-ID: <CAFDf7UKQFF9W3i7PbPbu2qRaCwiVFHBxHa89gu_BiKnpEsF=%2BQ@mail.gmail.com> In-Reply-To: <3500818A-B09B-48F1-80FD-7BC419E386BF@FreeBSD.org> References: <CAFDf7ULB281n-aiWt5MVuo1XPjwbYhPf2GpVHnZonN0PF4YquQ@mail.gmail.com> <06E66577-0A59-4BCD-BCD8-9C1BC2B6F17C@FreeBSD.org> <CAFDf7UJe0pLKUag4%2BSXs-iS2vWQ45P-Av-BM9EPO2aXLpE7X1g@mail.gmail.com> <CAFDf7U%2BHW%2BeKREA4wwvVjGk%2BtR6FF3hRtQxg6r%2BPDh39Ye3HAw@mail.gmail.com> <CAFDf7UJpjZkNXktcRZAUY3KQpCge9vJb3abYj7uuqKTwucAeNw@mail.gmail.com> <3500818A-B09B-48F1-80FD-7BC419E386BF@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
--00000000000009a7180600ab384e Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable Should a temporary fix to unbreak clang16 on aarch64 acceptable? # See also: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D272013 # See also: https://chromium.googlesource.com/v8/v8.git/+/d15d49b09dc7aef9edcc4cf6a0cb2= b77a0db203f .if ${OPSYS} =3D=3D FreeBSD && ${OSVERSION} >=3D 1400091 && ${ARCH} =3D=3D = aarch64 CXXFLAGS+=3D -Wno-error=3Denum-constexpr-conversion .endif I think we got required info to open a PR. Dimitry Andric <dim@freebsd.org> escreveu no dia segunda, 17/07/2023 =C3=A0= (s) 09:48: > Probably the particular enums it complains about are only hit on aarch64. > It may also be that it already partially has the actual fixes from v8 for > this issue: > > > https://chromium.googlesource.com/v8/v8.git/+/d15d49b09dc7aef9edcc4cf6a0c= b2b77a0db203f > > -Dimitry > > > On 17 Jul 2023, at 10:33, Nuno Teixeira <eduardo@freebsd.org> wrote: > > > > (...) > > > > More strange is that I can only reproduce this on aarch64. amd64 seems > to build fine... > > Any clues? > > > > Nuno Teixeira <eduardo@freebsd.org> escreveu no dia segunda, 17/07/2023 > =C3=A0(s) 07:53: > > I can confirm that it builds. > > > > Should a PR be open or this is being taken already? > > > > Thanks, > > > > Nuno Teixeira <eduardo@freebsd.org> escreveu no dia domingo, 16/07/2023 > =C3=A0(s) 14:24: > > Hello Dimitry! > > > > Thanks, I didn't see that PR since I only searched node18... > > > > I'm building it right now with: > > > > +.if ${OPSYS} =3D=3D FreeBSD && ${OSVERSION} >=3D 1400091 > > +CXXFLAGS+=3D -Wno-error=3Denum-constexpr-conversion > > +.endif > > > > Tomorrow I will have results. > > > > Thanks! > > > > Dimitry Andric <dim@freebsd.org> escreveu no dia domingo, 16/07/2023 > =C3=A0(s) 14:10: > > On 16 Jul 2023, at 14:58, Nuno Teixeira <eduardo@freebsd.org> wrote: > > > > > > main 1400093 aarch64 > > > > > > Anyone seeing www/node18 failing (needed for www/firefox)? > > > I tried reverting latest 2 committs and same result with 18.16.0. > > > > > > Any hints? > > > > > > https://people.freebsd.org/~eduardo/logs/node18/node18-18.16.1.log > > > https://people.freebsd.org/~eduardo/logs/node18/node18-18.16.0.log > > > > These are similar to <https://bugs.freebsd.org/272013>, but instead of > properly fixing it, the can was kicked down the road in < > https://cgit.freebsd.org/ports/commit/?id=3Dee3e6d5a17a0c78bb56f8d5719de8= 2b8dd49950d > >. > > > > These enums should be fixed by either removing the values outside of th= e > enum's valid range, or by specifying the enum's underlying type explicitl= y > to allow for such values. In the near future the > -Wno-enum-constexpr-conversion flag is going to go away! > > > > -Dimitry > > > > > > > > -- > > Nuno Teixeira > > FreeBSD Committer (ports) > > > > > > -- > > Nuno Teixeira > > FreeBSD Committer (ports) > > > > > > -- > > Nuno Teixeira > > FreeBSD Committer (ports) > > --=20 Nuno Teixeira FreeBSD Committer (ports) --00000000000009a7180600ab384e Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable <div dir=3D"ltr"><div><br></div><div>Should a temporary fix to unbreak clan= g16 on aarch64 acceptable?</div><div><br></div><div># See also: <a href=3D"= https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D272013">https://bugs.fr= eebsd.org/bugzilla/show_bug.cgi?id=3D272013</a></div><div># See also: <a hr= ef=3D"https://chromium.googlesource.com/v8/v8.git/+/d15d49b09dc7aef9edcc4cf= 6a0cb2b77a0db203f">https://chromium.googlesource.com/v8/v8.git/+/d15d49b09d= c7aef9edcc4cf6a0cb2b77a0db203f</a></div><div>.if ${OPSYS} =3D=3D FreeBSD &a= mp;& ${OSVERSION} >=3D 1400091 && ${ARCH} =3D=3D aarch64</di= v><div><div class=3D"gmail-add">CXXFLAGS+=3D -Wno-error=3Denum-constexpr-co= nversion</div><div class=3D"gmail-add">.endif</div><div class=3D"gmail-add"= ><br></div><div class=3D"gmail-add">I think we got required info to open a = PR.<br></div></div></div><br><div class=3D"gmail_quote"><div dir=3D"ltr" cl= ass=3D"gmail_attr">Dimitry Andric <<a href=3D"mailto:dim@freebsd.org">di= m@freebsd.org</a>> escreveu no dia segunda, 17/07/2023 =C3=A0(s) 09:48:<= br></div><blockquote class=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8e= x;border-left:1px solid rgb(204,204,204);padding-left:1ex">Probably the par= ticular enums it complains about are only hit on aarch64. It may also be th= at it already partially has the actual fixes from v8 for this issue:<br> <br> <a href=3D"https://chromium.googlesource.com/v8/v8.git/+/d15d49b09dc7aef9ed= cc4cf6a0cb2b77a0db203f" rel=3D"noreferrer" target=3D"_blank">https://chromi= um.googlesource.com/v8/v8.git/+/d15d49b09dc7aef9edcc4cf6a0cb2b77a0db203f</a= ><br> <br> -Dimitry<br> <br> > On 17 Jul 2023, at 10:33, Nuno Teixeira <<a href=3D"mailto:eduardo@= freebsd.org" target=3D"_blank">eduardo@freebsd.org</a>> wrote:<br> > <br> > (...)<br> > <br> > More strange is that I can only reproduce this on aarch64. amd64 seems= to build fine...<br> > Any clues?<br> > <br> > Nuno Teixeira <<a href=3D"mailto:eduardo@freebsd.org" target=3D"_bl= ank">eduardo@freebsd.org</a>> escreveu no dia segunda, 17/07/2023 =C3=A0= (s) 07:53:<br> > I can confirm that it builds.<br> > <br> > Should a PR be open or this is being taken already?<br> > <br> > Thanks,<br> > <br> > Nuno Teixeira <<a href=3D"mailto:eduardo@freebsd.org" target=3D"_bl= ank">eduardo@freebsd.org</a>> escreveu no dia domingo, 16/07/2023 =C3=A0= (s) 14:24:<br> > Hello Dimitry!<br> > <br> > Thanks, I didn't see that PR since I only searched node18...<br> > <br> > I'm building it right now with:<br> > <br> > +.if ${OPSYS} =3D=3D FreeBSD && ${OSVERSION} >=3D 1400091<b= r> > +CXXFLAGS+=3D -Wno-error=3Denum-constexpr-conversion<br> > +.endif<br> > <br> > Tomorrow I will have results.<br> > <br> > Thanks!<br> > <br> > Dimitry Andric <<a href=3D"mailto:dim@freebsd.org" target=3D"_blank= ">dim@freebsd.org</a>> escreveu no dia domingo, 16/07/2023 =C3=A0(s) 14:= 10:<br> > On 16 Jul 2023, at 14:58, Nuno Teixeira <<a href=3D"mailto:eduardo@= freebsd.org" target=3D"_blank">eduardo@freebsd.org</a>> wrote:<br> > ><br> > > main 1400093 aarch64<br> > ><br> > > Anyone seeing www/node18 failing (needed for www/firefox)?<br> > > I tried reverting latest 2 committs and same result with 18.16.0.= <br> > ><br> > > Any hints?<br> > ><br> > > <a href=3D"https://people.freebsd.org/~eduardo/logs/node18/node18= -18.16.1.log" rel=3D"noreferrer" target=3D"_blank">https://people.freebsd.o= rg/~eduardo/logs/node18/node18-18.16.1.log</a><br> > > <a href=3D"https://people.freebsd.org/~eduardo/logs/node18/node18= -18.16.0.log" rel=3D"noreferrer" target=3D"_blank">https://people.freebsd.o= rg/~eduardo/logs/node18/node18-18.16.0.log</a><br> > <br> > These are similar to <<a href=3D"https://bugs.freebsd.org/272013" r= el=3D"noreferrer" target=3D"_blank">https://bugs.freebsd.org/272013</a>>= , but instead of properly fixing it, the can was kicked down the road in &l= t;<a href=3D"https://cgit.freebsd.org/ports/commit/?id=3Dee3e6d5a17a0c78bb5= 6f8d5719de82b8dd49950d" rel=3D"noreferrer" target=3D"_blank">https://cgit.f= reebsd.org/ports/commit/?id=3Dee3e6d5a17a0c78bb56f8d5719de82b8dd49950d</a>&= gt;.<br> > <br> > These enums should be fixed by either removing the values outside of t= he enum's valid range, or by specifying the enum's underlying type = explicitly to allow for such values. In the near future the -Wno-enum-const= expr-conversion flag is going to go away!<br> > <br> > -Dimitry<br> > <br> > <br> > <br> > --<br> > Nuno Teixeira<br> > FreeBSD Committer (ports)<br> > <br> > <br> > --<br> > Nuno Teixeira<br> > FreeBSD Committer (ports)<br> > <br> > <br> > --<br> > Nuno Teixeira<br> > FreeBSD Committer (ports)<br> <br> </blockquote></div><br clear=3D"all"><br><span class=3D"gmail_signature_pre= fix">-- </span><br><div dir=3D"ltr" class=3D"gmail_signature"><div dir=3D"l= tr"><span style=3D"color:rgb(102,102,102)">Nuno Teixeira<br>FreeBSD Committ= er (ports)</span></div></div> --00000000000009a7180600ab384e--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAFDf7UKQFF9W3i7PbPbu2qRaCwiVFHBxHa89gu_BiKnpEsF=%2BQ>