[meta-oe][PATCH] fio: Remove --disable-static from EXTRA_OECONF

[meta-oe][PATCH] fio: Remove --disable-static from EXTRA_OECONF

[Fabio Berton]

As of commit OE-Core:773c9e18071d71454473dd81aff911104a2e9bc6
EXTRA_OECONF is appended with the option --disable-static on
DISABLE_STATIC variable and this cause the error:

    DEBUG: Python function sysroot_cleansstate finished
    DEBUG: Executing shell function do_configure
    Bad option --disable-static
    --cpu=                 Specify target CPU if auto-detect fails
    --cc=                  Specify compiler to use
    --extra-cflags=        Specify extra CFLAGS to pass to compiler
    --build-32bit-win      Enable 32-bit build on Windows
    --build-static         Build a static fio
    --esx                  Configure build options for esx
    --enable-gfio          Enable building of gtk gfio
    --disable-numa         Disable libnuma even if found
    --enable-libhdfs       Enable hdfs support
    WARNING: exit code 1 from a shell command.

So, we need to disable this option.

Signed-off-by: Fabio Berton <fabio.berton@ossystems.com.br>
---
 meta-oe/recipes-benchmark/fio/fio_2.2.6.bb | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/meta-oe/recipes-benchmark/fio/fio_2.2.6.bb b/meta-oe/recipes-benchmark/fio/fio_2.2.6.bb
index dca0e64..5bf387c 100644
--- a/meta-oe/recipes-benchmark/fio/fio_2.2.6.bb
+++ b/meta-oe/recipes-benchmark/fio/fio_2.2.6.bb
@@ -25,6 +25,9 @@ SRC_URI = "git://git.kernel.dk/fio.git"
 
 S = "${WORKDIR}/git"
 
+# avoids build breaks when using no-static-libs.inc
+DISABLE_STATIC = ""
+
 EXTRA_OEMAKE = "CC='${CC}' LDFLAGS='${LDFLAGS}'"
 
 do_configure() {
-- 
2.1.4

Re: [meta-oe][PATCH] fio: Remove --disable-static from EXTRA_OECONF

[Christopher Larson]

I think the convention is to add such workarounds to no-static-libs.inc,
not the recipe, at this time.

On Thu, Feb 18, 2016 at 10:30 AM Fabio Berton <fabio.berton@ossystems.com.br>
wrote:

> As of commit OE-Core:773c9e18071d71454473dd81aff911104a2e9bc6
> EXTRA_OECONF is appended with the option --disable-static on
> DISABLE_STATIC variable and this cause the error:
>
>     DEBUG: Python function sysroot_cleansstate finished
>     DEBUG: Executing shell function do_configure
>     Bad option --disable-static
>     --cpu=                 Specify target CPU if auto-detect fails
>     --cc=                  Specify compiler to use
>     --extra-cflags=        Specify extra CFLAGS to pass to compiler
>     --build-32bit-win      Enable 32-bit build on Windows
>     --build-static         Build a static fio
>     --esx                  Configure build options for esx
>     --enable-gfio          Enable building of gtk gfio
>     --disable-numa         Disable libnuma even if found
>     --enable-libhdfs       Enable hdfs support
>     WARNING: exit code 1 from a shell command.
>
> So, we need to disable this option.
>
> Signed-off-by: Fabio Berton <fabio.berton@ossystems.com.br>
> ---
>  meta-oe/recipes-benchmark/fio/fio_2.2.6.bb | 3 +++
>  1 file changed, 3 insertions(+)
>
> diff --git a/meta-oe/recipes-benchmark/fio/fio_2.2.6.bb
> b/meta-oe/recipes-benchmark/fio/fio_2.2.6.bb
> index dca0e64..5bf387c 100644
> --- a/meta-oe/recipes-benchmark/fio/fio_2.2.6.bb
> +++ b/meta-oe/recipes-benchmark/fio/fio_2.2.6.bb
> @@ -25,6 +25,9 @@ SRC_URI = "git://git.kernel.dk/fio.git"
>
>  S = "${WORKDIR}/git"
>
> +# avoids build breaks when using no-static-libs.inc
> +DISABLE_STATIC = ""
> +
>  EXTRA_OEMAKE = "CC='${CC}' LDFLAGS='${LDFLAGS}'"
>
>  do_configure() {
> --
> 2.1.4
>
> --
> _______________________________________________
> Openembedded-devel mailing list
> Openembedded-devel@lists.openembedded.org
> http://lists.openembedded.org/mailman/listinfo/openembedded-devel
>

Re: [meta-oe][PATCH] fio: Remove --disable-static from EXTRA_OECONF

[Otavio Salvador]

On Thu, Feb 18, 2016 at 6:43 PM, Christopher Larson <kergoth@gmail.com> wrote:
> I think the convention is to add such workarounds to no-static-libs.inc,
> not the recipe, at this time.

For OE-Core, yes. For Meta-OE? I think the recipe is the easiest to
notice when doing the upgrade.

-- 
Otavio Salvador                             O.S. Systems
http://www.ossystems.com.br        http://code.ossystems.com.br
Mobile: +55 (53) 9981-7854            Mobile: +1 (347) 903-9750

Re: [meta-oe][PATCH] fio: Remove --disable-static from EXTRA_OECONF

[Richard Purdie]

On Thu, 2016-02-18 at 20:43 +0000, Christopher Larson wrote:
> I think the convention is to add such workarounds to no-static
> -libs.inc,
> not the recipe, at this time.

This is a good question. I was just doing some world builds with meta
-oe in and have a few more of these fixes.

Referencing meta-oe recipes in OE-Core doesn't seem right somehow. We
could create an .inc file in meta-oe included by something like
layer.conf which set these, setting it in the recipe does somehow seem
like the better solution though.

Does anyone have any other thoughts on this?

Cheers,

Richard

Re: [meta-oe][PATCH] fio: Remove --disable-static from EXTRA_OECONF

[Martin Jansa]

What's the advantage of more global .inc file?

IMHO it's only increasing chances for conflicts when merging/cherry-picking.

On Sun, Feb 21, 2016 at 11:30 AM, Richard Purdie <
richard.purdie@linuxfoundation.org> wrote:

> On Thu, 2016-02-18 at 20:43 +0000, Christopher Larson wrote:
> > I think the convention is to add such workarounds to no-static
> > -libs.inc,
> > not the recipe, at this time.
>
> This is a good question. I was just doing some world builds with meta
> -oe in and have a few more of these fixes.
>
> Referencing meta-oe recipes in OE-Core doesn't seem right somehow. We
> could create an .inc file in meta-oe included by something like
> layer.conf which set these, setting it in the recipe does somehow seem
> like the better solution though.
>
> Does anyone have any other thoughts on this?
>
> Cheers,
>
> Richard
> --
> _______________________________________________
> Openembedded-devel mailing list
> Openembedded-devel@lists.openembedded.org
> http://lists.openembedded.org/mailman/listinfo/openembedded-devel
>

Re: [meta-oe][PATCH] fio: Remove --disable-static from EXTRA_OECONF

[Richard Purdie]

On Sun, 2016-02-21 at 11:47 +0100, Martin Jansa wrote:
> What's the advantage of more global .inc file?

In OE-Core it started as a standalone experiment which made sense as a
separate .inc as we've done with some other things of this nature. I
thought it was going to take longer to work through and establish. 

Having looked at the effect, I decided to adopt it for poky. At this
point I suspect marking up the recipes is the best way to do things.

I'd note that if we had a way to know if a recipe was really autotools
or just acts a bit like it, it would help a lot...

> IMHO it's only increasing chances for conflicts when merging/cherry
> -picking.

I think the central class did make sense for an initial evaluation of
this but probably not going forward...

Cheers,

Richard

Re: [meta-oe][PATCH] fio: Remove --disable-static from EXTRA_OECONF

[Burton, Ross]

On 21 February 2016 at 22:44, Richard Purdie <
richard.purdie@linuxfoundation.org> wrote:

> I'd note that if we had a way to know if a recipe was really autotools
> or just acts a bit like it, it would help a lot...
>

Something I've been meaning to do for a long time is review everything that
is using autotools but isn't really autotools, and stop them using
autotools.  For example, cmake.bbclass was inheriting autotools.bbclass and
only using the autotools_do_install function, which is a glorified "make
install".

Ross