Re: undefined symbol: PL_stack_sp

[email protected] (Ingy dot Net) Fri, 18 Jul 2014 09:28:01 -0700
Newsgroups perl.inline
Message-ID <CAHJtQJ6YZx4yynjEE2dy_J6VMwD+abd4fic9hhTaoNiTBpwKhw@mail.gmail.com>
--047d7b2eda3bb2499d04fe7a3d8a
Content-Type: text/plain; charset=UTF-8

Can someone turn this into an issue here
https://github.com/ingydotnet/inline-c-pm/issues

This type of discussion would do better there, as we'll get around to
addressing it.

Cheers


On Thu, Jul 17, 2014 at 5:58 PM, <[email protected]> wrote:

>
> From: demerphq
> Sent: Friday, July 18, 2014 7:40 AM
>
>
>  The above shows what goes wrong. Makemaker decides that PERL_LIB is
>> "../../../../" instead of "/home/yorton/perl5/perlbrew/
>> perls/perl-5.14.4/lib/5.14.4"
>>
>> below is the diff of the two Makefiles, but the relevant part is as
>> follows:
>>
>> -PERL_LIB = /home/yorton/perl5/perlbrew/perls/perl-5.14.4/lib/5.14.4
>> -PERL_ARCHLIB = /home/yorton/perl5/perlbrew/perls/perl-5.14.4/lib/5.14.4/
>> x86_64-linux
>> +PERL_LIB = ../../../../lib
>> +PERL_ARCHLIB = ../../../../lib
>>
>> I am guessing that MakerMaker decides that is being used as part of
>> building perl, and that it should use it as the location for its header
>> files. I know there is special logic in MakeMaker for this kind of purpose.
>>
>
> I didn't know that.
> If we could get a handle on precisely how to detect that this problem
> might arise, then we could probably have Inline issue a warning that
> "things might go awry, and if they do then you should move the script to a
> different (sane) location".
>
> I tried running your demo script on Windows (as try.pl) in
> C:/git_tree/perl/Porting with perl-5.14.0, EU-MM-6.62 and it still worked
> fine for me.
> Might there be a perlbrew component to this issue ? (I'm not at all
> familiar with perlbrew. I just build my perls the old fashioned way and
> call them via different shell/batch scripts.)
>
> I see that you're running EU-MM-6.57_05. If you think it's worth the
> effort I guess you could see if updating to EU-MM-6.62 fixes the problem -
> or do any other digging you think might be productive. (It's actually a bit
> interesting.)
> Otherwise we can just settle for the "Don't do that" solution ;-)
>
> Cheers,
> Rob
>

--047d7b2eda3bb2499d04fe7a3d8a
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div>Can someone turn this into an issue here <a href=3D"h=
ttps://github.com/ingydotnet/inline-c-pm/issues">https://github.com/ingydot=
net/inline-c-pm/issues</a><br><br></div><div>This type of discussion would =
do better there, as we&#39;ll get around to addressing it.<br>
<br></div><div>Cheers<br></div></div><div class=3D"gmail_extra"><br><br><di=
v class=3D"gmail_quote">On Thu, Jul 17, 2014 at 5:58 PM,  <span dir=3D"ltr"=
>&lt;<a href=3D"mailto:[email protected]" target=3D"_blank">sisyphu=
[email protected]</a>&gt;</span> wrote:<br>
<blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1p=
x #ccc solid;padding-left:1ex"><br>
From: demerphq<br>
Sent: Friday, July 18, 2014 7:40 AM<div class=3D""><br>
<br>
<blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1p=
x #ccc solid;padding-left:1ex">
The above shows what goes wrong. Makemaker decides that PERL_LIB is &quot;.=
./../../../&quot; instead of &quot;/home/yorton/perl5/perlbrew/<u></u>perls=
/perl-5.14.4/lib/5.14.4&quot;<br>
<br>
below is the diff of the two Makefiles, but the relevant part is as follows=
:<br>
<br>
-PERL_LIB =3D /home/yorton/perl5/perlbrew/<u></u>perls/perl-5.14.4/lib/5.14=
.4<br>
-PERL_ARCHLIB =3D /home/yorton/perl5/perlbrew/<u></u>perls/perl-5.14.4/lib/=
5.14.4/<u></u>x86_64-linux<br>
+PERL_LIB =3D ../../../../lib<br>
+PERL_ARCHLIB =3D ../../../../lib<br>
<br>
I am guessing that MakerMaker decides that is being used as part of buildin=
g perl, and that it should use it as the location for its header files. I k=
now there is special logic in MakeMaker for this kind of purpose.<br>
</blockquote>
<br></div>
I didn&#39;t know that.<br>
If we could get a handle on precisely how to detect that this problem might=
 arise, then we could probably have Inline issue a warning that &quot;thing=
s might go awry, and if they do then you should move the script to a differ=
ent (sane) location&quot;.<br>

<br>
I tried running your demo script on Windows (as <a href=3D"http://try.pl" t=
arget=3D"_blank">try.pl</a>) in C:/git_tree/perl/Porting with perl-5.14.0, =
EU-MM-6.62 and it still worked fine for me.<br>
Might there be a perlbrew component to this issue ? (I&#39;m not at all fam=
iliar with perlbrew. I just build my perls the old fashioned way and call t=
hem via different shell/batch scripts.)<br>
<br>
I see that you&#39;re running EU-MM-6.57_05. If you think it&#39;s worth th=
e effort I guess you could see if updating to EU-MM-6.62 fixes the problem =
- or do any other digging you think might be productive. (It&#39;s actually=
 a bit interesting.)<br>

Otherwise we can just settle for the &quot;Don&#39;t do that&quot; solution=
 ;-)<br>
<br>
Cheers,<br>
Rob <br>
</blockquote></div><br></div>

--047d7b2eda3bb2499d04fe7a3d8a--