[Buildroot] Failing bind 9.11.1-P3 configure failure

[Buildroot] Failing bind 9.11.1-P3 configure failure

[grunpferd at netscape.net]

Currently unable to build bind 9.11.1-P3.



It looks like there is some confusion over the local existence of header files resulting in the following error message.
Previous version of bind compiles successfully.


...
checking for using OpenSSL for hash functions... no
checking for PKCS11 support... no
checking for PKCS11 tools... disabled
checking for native PKCS11... disabled
checking for clock_gettime... yes
checking for lmdb library... yes
checking for library containing mdb_env_create... no
configure: error: found lmdb include but not library.
package/pkg-generic.mk:217: recipe for target '/u/rfg/w/buildroot_tk1_gcc7_4.12/output/build/bind-9.11.1-P3/.stamp_configured' faile
d
make: *** [/u/rfg/w/buildroot_tk1_gcc7_4.12/output/build/bind-9.11.1-P3/.stamp_configured] Error 1

...


So buildroot can't configure bind-9.11.1-P3.


Ross








-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.busybox.net/pipermail/buildroot/attachments/20170806/da756c3d/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: .config
Type: application/octet-stream
Size: 135378 bytes
Desc: not available
URL: <http://lists.busybox.net/pipermail/buildroot/attachments/20170806/da756c3d/attachment.obj>

[Buildroot] Failing bind 9.11.1-P3 configure failure

[Peter Seiderer]

Hello 'grunpferd'(*),

On Sun, 6 Aug 2017 21:53:01 -0400, grunpferd at netscape.net wrote:

> Currently unable to build bind 9.11.1-P3.
> 
> 
> 
> It looks like there is some confusion over the local existence of header files resulting in the following error message.
> Previous version of bind compiles successfully.
> 

Could be, mind to try with BR2_COMPILER_PARANOID_UNSAFE_PATH enabled?

> 
> ...
> checking for using OpenSSL for hash functions... no
> checking for PKCS11 support... no
> checking for PKCS11 tools... disabled
> checking for native PKCS11... disabled
> checking for clock_gettime... yes
> checking for lmdb library... yes
> checking for library containing mdb_env_create... no
> configure: error: found lmdb include but not library.
> package/pkg-generic.mk:217: recipe for target '/u/rfg/w/buildroot_tk1_gcc7_4.12/output/build/bind-9.11.1-P3/.stamp_configured' faile
> d
> make: *** [/u/rfg/w/buildroot_tk1_gcc7_4.12/output/build/bind-9.11.1-P3/.stamp_configured] Error 1
> 
> ...

Local build of bind here (with a random configuration) succeeds, mind to try
the following buildroot patch:

diff --git a/package/bind/bind.mk b/package/bind/bind.mk
index 988f40e1c2..c65568e9e3 100644
--- a/package/bind/bind.mk
+++ b/package/bind/bind.mk
@@ -24,6 +24,7 @@ BIND_CONF_ENV = \
        BUILD_CC="$(TARGET_CC)" \
        BUILD_CFLAGS="$(TARGET_CFLAGS)"
 BIND_CONF_OPTS = \
+       --without-lmdb \
        --with-libjson=no \
        --with-randomdev=/dev/urandom \
        --enable-epoll \


Regards,
Peter

(*) Using a real name instead of any pseudonym or fake name is preferred

> 
> 
> So buildroot can't configure bind-9.11.1-P3.
> 
> 
> Ross
> 
> 
> 
> 
> 
> 
> 
> 

[Buildroot] Failing bind 9.11.1-P3 configure failure

[Arnout Vandecappelle]

 Hi Peter,

On 08-08-17 00:19, Peter Seiderer wrote:
> Local build of bind here (with a random configuration) succeeds, mind to try
> the following buildroot patch:
> 
> diff --git a/package/bind/bind.mk b/package/bind/bind.mk
> index 988f40e1c2..c65568e9e3 100644
> --- a/package/bind/bind.mk
> +++ b/package/bind/bind.mk
> @@ -24,6 +24,7 @@ BIND_CONF_ENV = \
>         BUILD_CC="$(TARGET_CC)" \
>         BUILD_CFLAGS="$(TARGET_CFLAGS)"
>  BIND_CONF_OPTS = \
> +       --without-lmdb \
>         --with-libjson=no \
>         --with-randomdev=/dev/urandom \
>         --enable-epoll \

 Could you send this as an official patch?


> (*) Using a real name instead of any pseudonym or fake name is preferred

 Actually, only when submitting patches. For bug reporting, asking questions, on
IRC, etc. use pseudonyms all you like.

 Regards,
 Arnout

-- 
Arnout Vandecappelle                          arnout at mind be
Senior Embedded Software Architect            +32-16-286500
Essensium/Mind                                http://www.mind.be
G.Geenslaan 9, 3001 Leuven, Belgium           BE 872 984 063 RPR Leuven
LinkedIn profile: http://www.linkedin.com/in/arnoutvandecappelle
GPG fingerprint:  7493 020B C7E3 8618 8DEC 222C 82EB F404 F9AC 0DDF

[Buildroot] Failing bind 9.11.1-P3 configure failure

[Peter Seiderer]

Hello Arnout,

On Tue, 8 Aug 2017 16:18:35 +0200, Arnout Vandecappelle <arnout@mind.be> wrote:

>  Hi Peter,
> 
> On 08-08-17 00:19, Peter Seiderer wrote:
> > Local build of bind here (with a random configuration) succeeds, mind to try
> > the following buildroot patch:
> > 
> > diff --git a/package/bind/bind.mk b/package/bind/bind.mk
> > index 988f40e1c2..c65568e9e3 100644
> > --- a/package/bind/bind.mk
> > +++ b/package/bind/bind.mk
> > @@ -24,6 +24,7 @@ BIND_CONF_ENV = \
> >         BUILD_CC="$(TARGET_CC)" \
> >         BUILD_CFLAGS="$(TARGET_CFLAGS)"
> >  BIND_CONF_OPTS = \
> > +       --without-lmdb \
> >         --with-libjson=no \
> >         --with-randomdev=/dev/urandom \
> >         --enable-epoll \
> 
>  Could you send this as an official patch?

O.k, patch is on the way...

> 
> 
> > (*) Using a real name instead of any pseudonym or fake name is preferred
> 
>  Actually, only when submitting patches. For bug reporting, asking questions, on
> IRC, etc. use pseudonyms all you like.

And thanks for clarifying...

Regards,
Peter

> 
>  Regards,
>  Arnout
> 

[Buildroot] Failing bind 9.11.1-P3 configure failure

[grunpferd at netscape.net]

It looks like there is some confusion over the local existence of header files resulting in error during configure.



> Local build of bind here (with a random configuration) succeeds, mind to try
> the following buildroot patch:
I used the +       --without-lmdb \ option and that does work.


I guess the the real issue is why it is looking at local header files at all, rather than the buildroot headers.
Also I presume why bind wants to try and use -lmdb is to speed up some object store and retrieval,
so maybe the right solution is to build a local version of mdb and have host headers and library available for the target.
I have not tested the BR2_COMPILER_PARANOID_UNSAFE_PATH facility, prefer to let the build complete so I can test the resulting rootfs first!


Ross


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.busybox.net/pipermail/buildroot/attachments/20170807/2eb3049b/attachment.html>

[Buildroot] Failing bind 9.11.1-P3 configure failure

[Peter Seiderer]

Hello 'grunpferd',

On Mon, 7 Aug 2017 20:36:58 -0400, grunpferd at netscape.net wrote:

> 
> It looks like there is some confusion over the local existence of header files resulting in error during configure.
> 
> 
> 
> > Local build of bind here (with a random configuration) succeeds, mind to try
> > the following buildroot patch:
> I used the +       --without-lmdb \ option and that does work.

Thanks for testing...

> 
> 
> I guess the the real issue is why it is looking at local header files at all, rather than the buildroot headers.
> Also I presume why bind wants to try and use -lmdb is to speed up some object store and retrieval,
> so maybe the right solution is to build a local version of mdb and have host headers and library available for the target.

Yes this would be a nice buildroot feature, but I think there is no lmdb (Lightning
Memory-mapped Database) package yet in buildroot to use...

Regards,
Peter

> I have not tested the BR2_COMPILER_PARANOID_UNSAFE_PATH facility, prefer to let the build complete so I can test the resulting rootfs first!
> 
> 
> Ross
> 
> 

[Buildroot] Failing bind 9.11.1-P3 configure failure

[grunpferd at netscape.net]

G'day Peter,

>On Mon, 7 Aug 2017 20:36:58 -0400, grunpferd at netscape.net wrote:

>> 
>> It looks like there is some confusion over the local existence of header files resulting in error during configure.
>> 
>> 
>> 
>> > Local build of bind here (with a random configuration) succeeds, mind to try
>> > the following buildroot patch:
>> I used the +       --without-lmdb \ option and that does work.

>Thanks for testing...


No problems, rather get the issue sorted while still in -rc1.
I guess the problem with random configurations is they are not "real" or "real" enough!

>> 
>> 
>> I guess the the real issue is why it is looking at local header files at all, rather than the buildroot headers.
>> Also I presume why bind wants to try and use -lmdb is to speed up some object store and retrieval,
>> so maybe the right solution is to build a local version of mdb and have host headers and library available for the target.

> Yes this would be a nice buildroot feature, but I think there is no lmdb (Lightning
> Memory-mapped Database) package yet in buildroot to use...

Bind might run a little slower but at least it builds and works.
I think bind used to use BSD db and has since shifted to mdb.


The more serious issue is that configure is using local files and failing in this cross compilation toolchain generated by buildroot.
While the patch works in terms of allowing buildroot to build bind there is a deeper problem here.
The problem is more in the configuration tools getting it wrong, why is it looking at local header files at all, rather than the buildroot generated header files!


Keep the good work coming,


Ross


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.busybox.net/pipermail/buildroot/attachments/20170808/53adc3e1/attachment.html>