[PATCH v2] ghostscript: Define COMPATIBLE_HOST with supported platforms

[PATCH v2] ghostscript: Define COMPATIBLE_HOST with supported platforms

[Nathan Rossi]

Setup the COMPATIBLE_HOST variable with all supported hosts based on the
provided objarch.h files. This inverts the existing pattern where
unsupported hosts (e.g. 'arc.*') are explicitly set.

This suppresses the warning message generated on hosts that are not
supported but were not listed explicitly.

  WARNING: .../ghostscript_9.25.bb: Unable to get checksum for
  ghostscript SRC_URI entry objarch.h: file could not be found

Signed-off-by: Nathan Rossi <nathan@nathanrossi.com>
---
Changes in v2:
  - Fix up to use HOST_SYS patterns instead of OVERRIDES
  - Make i586/i686 pattern common as done in other recipes
---
 meta/recipes-extended/ghostscript/ghostscript_9.25.bb | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/meta/recipes-extended/ghostscript/ghostscript_9.25.bb b/meta/recipes-extended/ghostscript/ghostscript_9.25.bb
index 35eaaeb2fa..93134dfb9d 100644
--- a/meta/recipes-extended/ghostscript/ghostscript_9.25.bb
+++ b/meta/recipes-extended/ghostscript/ghostscript_9.25.bb
@@ -121,5 +121,4 @@ do_install_class-native () {
 
 BBCLASSEXTEND = "native"
 
-# ghostscript does not supports "arc"
-COMPATIBLE_HOST = "^(?!arc).*"
+COMPATIBLE_HOST = "(aarch64|arm|i.86|microblaze|mips|nios2|powerpc|x86_64).*"
---
2.19.1

Re: [PATCH v2] ghostscript: Define COMPATIBLE_HOST with supported platforms

[Khem Raj]

On Thu, Nov 8, 2018 at 12:43 AM Nathan Rossi <nathan@nathanrossi.com> wrote:
>
> Setup the COMPATIBLE_HOST variable with all supported hosts based on the
> provided objarch.h files. This inverts the existing pattern where
> unsupported hosts (e.g. 'arc.*') are explicitly set.
>
> This suppresses the warning message generated on hosts that are not
> supported but were not listed explicitly.
>
>   WARNING: .../ghostscript_9.25.bb: Unable to get checksum for
>   ghostscript SRC_URI entry objarch.h: file could not be found
>
> Signed-off-by: Nathan Rossi <nathan@nathanrossi.com>
> ---
> Changes in v2:
>   - Fix up to use HOST_SYS patterns instead of OVERRIDES
>   - Make i586/i686 pattern common as done in other recipes
> ---
>  meta/recipes-extended/ghostscript/ghostscript_9.25.bb | 3 +--
>  1 file changed, 1 insertion(+), 2 deletions(-)
>
> diff --git a/meta/recipes-extended/ghostscript/ghostscript_9.25.bb b/meta/recipes-extended/ghostscript/ghostscript_9.25.bb
> index 35eaaeb2fa..93134dfb9d 100644
> --- a/meta/recipes-extended/ghostscript/ghostscript_9.25.bb
> +++ b/meta/recipes-extended/ghostscript/ghostscript_9.25.bb
> @@ -121,5 +121,4 @@ do_install_class-native () {
>
>  BBCLASSEXTEND = "native"
>
> -# ghostscript does not supports "arc"
> -COMPATIBLE_HOST = "^(?!arc).*"
> +COMPATIBLE_HOST = "(aarch64|arm|i.86|microblaze|mips|nios2|powerpc|x86_64).*"

this ignores riscv now. since you wsnt to
ignore for just one arch you can use

COMPATIBLE_HOST_arc = "(none)"

or somesuch mechanism.

> ---
> 2.19.1
> --
> _______________________________________________
> Openembedded-core mailing list
> Openembedded-core@lists.openembedded.org
> http://lists.openembedded.org/mailman/listinfo/openembedded-core

Re: [PATCH v2] ghostscript: Define COMPATIBLE_HOST with supported platforms

[Burton, Ross]

On Thu, 8 Nov 2018 at 13:30, Khem Raj <raj.khem@gmail.com> wrote:
> > --- a/meta/recipes-extended/ghostscript/ghostscript_9.25.bb
> > +++ b/meta/recipes-extended/ghostscript/ghostscript_9.25.bb
> > @@ -121,5 +121,4 @@ do_install_class-native () {
> >
> >  BBCLASSEXTEND = "native"
> >
> > -# ghostscript does not supports "arc"
> > -COMPATIBLE_HOST = "^(?!arc).*"
> > +COMPATIBLE_HOST = "(aarch64|arm|i.86|microblaze|mips|nios2|powerpc|x86_64).*"
>
> this ignores riscv now. since you wsnt to
> ignore for just one arch you can use
>
> COMPATIBLE_HOST_arc = "(none)"
>
> or somesuch mechanism.

How is that different to the original assignment?

The real point here is that all of these variations have flaws.
Listing the architectures which don't work means the list isn't
complete, but listing the working architectures means that layers
which provide a bbappend instead of submitting to oe-core [1] will
then refuse to build ghostscript unless they also do something clever
with COMPATIBLE_HOST.

Ross

[1] meta-riscv, I'm looking at you
https://github.com/riscv/meta-riscv/tree/master/recipes-extended/ghostscript

Re: [PATCH v2] ghostscript: Define COMPATIBLE_HOST with supported platforms

[Nathan Rossi]

On Thu, 8 Nov 2018 at 23:44, Burton, Ross <ross.burton@intel.com> wrote:
>
> On Thu, 8 Nov 2018 at 13:30, Khem Raj <raj.khem@gmail.com> wrote:
> > > --- a/meta/recipes-extended/ghostscript/ghostscript_9.25.bb
> > > +++ b/meta/recipes-extended/ghostscript/ghostscript_9.25.bb
> > > @@ -121,5 +121,4 @@ do_install_class-native () {
> > >
> > >  BBCLASSEXTEND = "native"
> > >
> > > -# ghostscript does not supports "arc"
> > > -COMPATIBLE_HOST = "^(?!arc).*"
> > > +COMPATIBLE_HOST = "(aarch64|arm|i.86|microblaze|mips|nios2|powerpc|x86_64).*"
> >
> > this ignores riscv now. since you wsnt to
> > ignore for just one arch you can use
> >
> > COMPATIBLE_HOST_arc = "(none)"
> >
> > or somesuch mechanism.
>
> How is that different to the original assignment?
>
> The real point here is that all of these variations have flaws.
> Listing the architectures which don't work means the list isn't
> complete, but listing the working architectures means that layers
> which provide a bbappend instead of submitting to oe-core [1] will
> then refuse to build ghostscript unless they also do something clever
> with COMPATIBLE_HOST.

Ross beat me to it. But I was making this change as inverting due to
the desire to remove the warning and disable the package for at least
three architectures (or1k, riscv32 and lm32).

Regards,
Nathan

>
> Ross
>
> [1] meta-riscv, I'm looking at you
> https://github.com/riscv/meta-riscv/tree/master/recipes-extended/ghostscript