Re: pkg/56785: Python 3.9 build fails with pkgsrc-2022Q1 on Linux

[Yosuke Kawasaki <you.kawasaki%gmail.com@localhost>]

The following reply was made to PR pkg/56785; it has been noted by GNATS.

From: Yosuke Kawasaki <you.kawasaki%gmail.com@localhost>
To: gnats-bugs%netbsd.org@localhost
Cc: pkg-manager%netbsd.org@localhost, gnats-admin%netbsd.org@localhost, pkgsrc-bugs%netbsd.org@localhost
Subject: Re: pkg/56785: Python 3.9 build fails with pkgsrc-2022Q1 on Linux
Date: Sun, 24 Apr 2022 06:45:05 +0900

 --0000000000003a066205dd5942fa
 Content-Type: text/plain; charset="UTF-8"
 Content-Transfer-Encoding: quoted-printable
 
 Following are sys.base_prefix and sys.path under my build environment.
 
 sys.base_prefix:/var/pkg
 sys.path:['/var/pkg/work/lang/python39/work/Python-3.9.12',
 '/var/pkg/lib/python39.zip',
 '/var/pkg/work/lang/python39/work/Python-3.9.12/Lib',
 '/var/pkg/work/lang/python39/work/Python-3.9.12/build/lib.linux-x86_64-3.9'=
 ,
 '/var/pkg/lib/python3.9/site-packages']
 
 My PREFIX is /var/pkg and WRKOBJDIR is /var/pkg/work.
 The patched operation "sys.path =3D [p for p..." removes all the sys.path
 members.
 
 It seems that the patched operation should keep paths under WRKOBJDIR in
 sys.path.
 When I disabled WRKOBJDIR setting, the build finished without problem.
 
 Best regards,
 
 Yosuke
 
 2022=E5=B9=B44=E6=9C=8821=E6=97=A5(=E6=9C=A8) 11:55 David Holland <dholland=
 -pbugs%netbsd.org@localhost>:
 
 > The following reply was made to PR pkg/56785; it has been noted by GNATS.
 >
 > From: David Holland <dholland-pbugs%netbsd.org@localhost>
 > To: gnats-bugs%netbsd.org@localhost
 > Cc:
 > Subject: Re: pkg/56785: Python 3.9 build fails with pkgsrc-2022Q1 on Linu=
 x
 > Date: Thu, 21 Apr 2022 02:50:06 +0000
 >
 >  On Sun, Apr 10, 2022 at 12:35:00AM +0000, you.kawasaki%gmail.com@localhost wrote:
 >   > I found commenting out the following line from patch/patch-setup is
 > effective in completing the build.
 >   >
 >   > +sys.path =3D [p for p in sys.path if not re.compile('^' +
 > sys.base_prefix + '/').match(p)]
 >   >
 >   > Under my environment, this line makes the sys.path an empty list.
 >   > Moreover, it worked fine if I change "if not" to "if".
 >
 >  What's in sys.base_prefix? Sounds like it's got some characters in it
 >  that make Python think it's an invalid regexp.
 >
 >  There should probably be some kind of quoting operation on there to
 >  protect it.
 >
 >  --
 >  David A. Holland
 >  dholland%netbsd.org@localhost
 >
 >
 
 --=20
 Yosuke Kawasaki
 
 you.kawasaki%gmail.com@localhost
 
 --0000000000003a066205dd5942fa
 Content-Type: text/html; charset="UTF-8"
 Content-Transfer-Encoding: quoted-printable
 
 <div dir=3D"ltr"><div>Following are sys.base_prefix and sys.path under my b=
 uild environment.</div><div><br></div><div>sys.base_prefix:/var/pkg<br>sys.=
 path:[&#39;/var/pkg/work/lang/python39/work/Python-3.9.12&#39;, &#39;/var/p=
 kg/lib/python39.zip&#39;, &#39;/var/pkg/work/lang/python39/work/Python-3.9.=
 12/Lib&#39;, &#39;/var/pkg/work/lang/python39/work/Python-3.9.12/build/lib.=
 linux-x86_64-3.9&#39;, &#39;/var/pkg/lib/python3.9/site-packages&#39;]</div=
 ><div><br></div><div>My PREFIX is /var/pkg and WRKOBJDIR is /var/pkg/work.<=
 /div><div>The patched operation &quot;sys.path =3D [p for p...&quot; remove=
 s all the sys.path members.</div><div><br></div><div>It seems that the patc=
 hed operation should keep paths under WRKOBJDIR in sys.path.</div><div><div=
 >When I disabled WRKOBJDIR setting, the build finished without problem.<br>=
 </div></div><div><br></div><div>Best regards,</div><div><br></div><div>Yosu=
 ke<br></div><div><br></div><div><div class=3D"gmail_quote"><div dir=3D"ltr"=
  class=3D"gmail_attr">2022=E5=B9=B44=E6=9C=8821=E6=97=A5(=E6=9C=A8) 11:55 D=
 avid Holland &lt;<a href=3D"mailto:dholland-pbugs%netbsd.org@localhost"; target=3D"_bl=
 ank">dholland-pbugs%netbsd.org@localhost</a>&gt;:<br></div><blockquote class=3D"gmail=
 _quote" style=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204=
 ,204);padding-left:1ex">The following reply was made to PR pkg/56785; it ha=
 s been noted by GNATS.<br>
 <br>
 From: David Holland &lt;<a href=3D"mailto:dholland-pbugs%netbsd.org@localhost"; target=
 =3D"_blank">dholland-pbugs%netbsd.org@localhost</a>&gt;<br>
 To: <a href=3D"mailto:gnats-bugs%netbsd.org@localhost"; target=3D"_blank">gnats-bugs@n=
 etbsd.org</a><br>
 Cc: <br>
 Subject: Re: pkg/56785: Python 3.9 build fails with pkgsrc-2022Q1 on Linux<=
 br>
 Date: Thu, 21 Apr 2022 02:50:06 +0000<br>
 <br>
 =C2=A0On Sun, Apr 10, 2022 at 12:35:00AM +0000, <a href=3D"mailto:you.kawas=
 aki%gmail.com@localhost" target=3D"_blank">you.kawasaki%gmail.com@localhost</a> wrote:<br>
 =C2=A0 &gt; I found commenting out the following line from patch/patch-setu=
 p is effective in completing the build.<br>
 =C2=A0 &gt; <br>
 =C2=A0 &gt; +sys.path =3D [p for p in sys.path if not re.compile(&#39;^&#39=
 ; + sys.base_prefix + &#39;/&#39;).match(p)]<br>
 =C2=A0 &gt;<br>
 =C2=A0 &gt; Under my environment, this line makes the sys.path an empty lis=
 t.<br>
 =C2=A0 &gt; Moreover, it worked fine if I change &quot;if not&quot; to &quo=
 t;if&quot;.<br>
 <br>
 =C2=A0What&#39;s in sys.base_prefix? Sounds like it&#39;s got some characte=
 rs in it<br>
 =C2=A0that make Python think it&#39;s an invalid regexp.<br>
 <br>
 =C2=A0There should probably be some kind of quoting operation on there to<b=
 r>
 =C2=A0protect it.<br>
 <br>
 =C2=A0-- <br>
 =C2=A0David A. Holland<br>
 =C2=A0<a href=3D"mailto:dholland%netbsd.org@localhost"; target=3D"_blank">dholland@net=
 bsd.org</a><br>
 <br>
 </blockquote></div></div><br clear=3D"all"><br>-- <br><div dir=3D"ltr" data=
 -smartmail=3D"gmail_signature"><div dir=3D"ltr"><div>Yosuke Kawasaki</div><=
 div><br></div><div><a href=3D"mailto:you.kawasaki%gmail.com@localhost"; target=3D"_bla=
 nk">you.kawasaki%gmail.com@localhost</a><br></div><div><p class=3D"MsoNormal" style=
 =3D"margin-top:1.2pt;line-height:12.0pt"><br></p></div></div></div></div>
 
 --0000000000003a066205dd5942fa--