Date: Fri, 14 Oct 2022 13:17:25 -0600 From: Adam Weinberger <adamw@freebsd.org> To: Jan Beich <jbeich@freebsd.org> Cc: Adam Weinberger <adamw@freebsd.org>, ports-committers@freebsd.org, dev-commits-ports-all@freebsd.org, dev-commits-ports-main@freebsd.org Subject: Re: git: 9d241a90370c - main - Mk: Add USES=luajit Message-ID: <CAP7rwciyjDPbQnKJEuTxGyzs73N0ORQCsKyXZt8x5e-vDi3ZYA@mail.gmail.com> In-Reply-To: <pmeu-6yr3-wny@FreeBSD.org> References: <202210141707.29EH7i08013647@gitrepo.freebsd.org> <k052-b9ml-wny@FreeBSD.org> <CAP7rwciigyJxVJZ5=e=m8=X6u0PpDP=urfp_jtrgfEJHHMj9Eg@mail.gmail.com> <pmeu-6yr3-wny@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
--000000000000d122d705eb037a05
Content-Type: text/plain; charset="UTF-8"
On Fri, Oct 14, 2022 at 12:33 PM Jan Beich <jbeich@freebsd.org> wrote:
> Adam Weinberger <adamw@freebsd.org> writes:
>
> > On Fri, Oct 14, 2022 at 11:25 AM Jan Beich <jbeich@freebsd.org> wrote:
> >
> >> Adam Weinberger <adamw@FreeBSD.org> writes:
> >>
> >> > For the vast majority of situations, USES=luajit will be
> sufficient.
> >> > Ports that use openresty's API additions can specify
> >> > USES=luajit:openresty.
> >>
> >> Are consumers to be switched "en masse" or "one by one"?
> >>
> >
> > Hi Jan,
> >
> > I'm going to switch a couple ports (neovim and luv) right now, but my
> > thinking is that perhaps the best approach is two exp-runs. The first run
> > can switch everything depending on lang/luajit to USES=luajit. For the
> > second run, I'll need to look at everything that depends on
> > luajit-openresty to find out whether it uses it because it uses the
> > openresty API or because what they actually wanted was "anything other
> than
> > the old version in lang/luajit."
> >
> > Does that sound sensible?
>
> Probably:
> - exp-runs can be reduced via local testing as USES=luajit only has ~50
> candidates
> - luajit -> luajit-devel is safe due to same upstream and tested by other
> distros
> - luajit-openresty -> luajit-devel build test may show false negatives[1]
> but
> ARCH-specific cases are likely safe
>
> [1] www/openresty and {devel,www}/lua-resty-* build fine with
> lang/luajit-devel.
> However, not every instance maybe as obvious whether to ask maintainer.
>
Everything you said sounds very sensible. It sounds more like one-by-one
conversion makes the most sense. I could really use help on that if anyone
has the time. I don't have a great development machine that I can use to
test stuff, so if anybody wants to help with converting it'd be a big help.
# Adam
--
Adam Weinberger
adamw@adamw.org // adamw@FreeBSD.org
https://www.adamw.org
--000000000000d122d705eb037a05
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
<div dir=3D"ltr"><div dir=3D"ltr"><div class=3D"gmail_default" style=3D"fon=
t-family:arial,sans-serif"><br></div></div><br><div class=3D"gmail_quote"><=
div dir=3D"ltr" class=3D"gmail_attr">On Fri, Oct 14, 2022 at 12:33 PM Jan B=
eich <<a href=3D"mailto:jbeich@freebsd.org">jbeich@freebsd.org</a>> w=
rote:<br></div><blockquote class=3D"gmail_quote" style=3D"margin:0px 0px 0p=
x 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">Adam Weinb=
erger <<a href=3D"mailto:adamw@freebsd.org" target=3D"_blank">adamw@free=
bsd.org</a>> writes:<br>
<br>
> On Fri, Oct 14, 2022 at 11:25 AM Jan Beich <<a href=3D"mailto:jbeic=
h@freebsd.org" target=3D"_blank">jbeich@freebsd.org</a>> wrote:<br>
><br>
>> Adam Weinberger <adamw@FreeBSD.org> writes:<br>
>><br>
>> >=C2=A0 =C2=A0 =C2=A0For the vast majority of situations, USES=
=3Dluajit will be sufficient.<br>
>> >=C2=A0 =C2=A0 =C2=A0Ports that use openresty's API additio=
ns can specify<br>
>> >=C2=A0 =C2=A0 =C2=A0USES=3Dluajit:openresty.<br>
>><br>
>> Are consumers to be switched "en masse" or "one by =
one"?<br>
>><br>
><br>
> Hi Jan,<br>
><br>
> I'm going to switch a couple ports (neovim and luv) right now, but=
my<br>
> thinking is that perhaps the best approach is two exp-runs. The first =
run<br>
> can switch everything depending on lang/luajit to USES=3Dluajit. For t=
he<br>
> second run, I'll need to look at everything that depends on<br>
> luajit-openresty to find out whether it uses it because it uses the<br=
>
> openresty API or because what they actually wanted was "anything =
other than<br>
> the old version in lang/luajit."<br>
><br>
> Does that sound sensible?<br>
<br>
Probably:<br>
- exp-runs can be reduced via local testing as USES=3Dluajit only has ~50 c=
andidates<br>
- luajit -> luajit-devel is safe due to same upstream and tested by othe=
r distros<br>
- luajit-openresty -> luajit-devel build test may show false negatives[1=
] but<br>
=C2=A0 ARCH-specific cases are likely safe<br>
<br>
[1] www/openresty and {devel,www}/lua-resty-* build fine with lang/luajit-d=
evel.<br>
=C2=A0 =C2=A0 However, not every instance maybe as obvious whether to ask m=
aintainer.<br>
</blockquote></div><div><br></div><div><div style=3D"font-family:arial,sans=
-serif" class=3D"gmail_default">Everything you said sounds very sensible. I=
t sounds more like one-by-one conversion makes the most sense. I could real=
ly use help on that if anyone has the time. I don't have a great develo=
pment machine that I can use to test stuff, so if anybody wants to help wit=
h converting it'd be a big help.</div><div style=3D"font-family:arial,s=
ans-serif" class=3D"gmail_default"><br></div><div style=3D"font-family:aria=
l,sans-serif" class=3D"gmail_default"># Adam</div><br></div><br>-- <br><div=
dir=3D"ltr" class=3D"gmail_signature">Adam Weinberger<br><a href=3D"mailto=
:adamw@adamw.org" target=3D"_blank">adamw@adamw.org</a> // adamw@FreeBSD.or=
g<br><a href=3D"https://www.adamw.org" target=3D"_blank">https://www.adamw.=
org</a></div></div>
--000000000000d122d705eb037a05--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAP7rwciyjDPbQnKJEuTxGyzs73N0ORQCsKyXZt8x5e-vDi3ZYA>