[PATCH] os-release: add more variables to list

[PATCH] os-release: add more variables to list

[Armin Kuster]

If someone extends the OS_RELEASE_FIELDS in a bbappend, it will cause the hash to change which is
caught by the yocto-check-layer

Signed-off-by: Armin Kuster <akuster808@gmail.com>
---
 meta/recipes-core/os-release/os-release.bb | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/meta/recipes-core/os-release/os-release.bb b/meta/recipes-core/os-release/os-release.bb
index 7f3d9cb..88bf8f8 100644
--- a/meta/recipes-core/os-release/os-release.bb
+++ b/meta/recipes-core/os-release/os-release.bb
@@ -12,7 +12,7 @@ do_configure[noexec] = "1"
 
 # Other valid fields: BUILD_ID ID_LIKE ANSI_COLOR CPE_NAME
 #                     HOME_URL SUPPORT_URL BUG_REPORT_URL
-OS_RELEASE_FIELDS = "ID ID_LIKE NAME VERSION VERSION_ID PRETTY_NAME"
+OS_RELEASE_FIELDS = "ID ID_LIKE NAME VERSION VERSION_ID PRETTY_NAME BUILD_ID HOME_URL BUG_REPORT_URL SUPPORT_URL CPE_NAME"
 
 ID = "${DISTRO}"
 NAME = "${DISTRO_NAME}"
-- 
2.7.4

Re: [PATCH] os-release: add more variables to list

[Richard Purdie]

On Thu, 2018-10-25 at 08:20 +0100, Armin Kuster wrote:
> If someone extends the OS_RELEASE_FIELDS in a bbappend, it will cause
> the hash to change which is
> caught by the yocto-check-layer
> 
> Signed-off-by: Armin Kuster <akuster808@gmail.com>
> ---
>  meta/recipes-core/os-release/os-release.bb | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/meta/recipes-core/os-release/os-release.bb
> b/meta/recipes-core/os-release/os-release.bb
> index 7f3d9cb..88bf8f8 100644
> --- a/meta/recipes-core/os-release/os-release.bb
> +++ b/meta/recipes-core/os-release/os-release.bb
> @@ -12,7 +12,7 @@ do_configure[noexec] = "1"
>  
>  # Other valid fields: BUILD_ID ID_LIKE ANSI_COLOR CPE_NAME
>  #                     HOME_URL SUPPORT_URL BUG_REPORT_URL
> -OS_RELEASE_FIELDS = "ID ID_LIKE NAME VERSION VERSION_ID PRETTY_NAME"
> +OS_RELEASE_FIELDS = "ID ID_LIKE NAME VERSION VERSION_ID PRETTY_NAME BUILD_ID HOME_URL BUG_REPORT_URL SUPPORT_URL CPE_NAME"

I'm not sure I agree with this. It would usually be a policy decision,
e.g. a ditro config which would change this and you'd not be doing this
with a bbappend?

Cheers,

Richard

Re: [PATCH] os-release: add more variables to list

[akuster808]

On 10/25/18 10:25 AM, Richard Purdie wrote:
> On Thu, 2018-10-25 at 08:20 +0100, Armin Kuster wrote:
>> If someone extends the OS_RELEASE_FIELDS in a bbappend, it will cause
>> the hash to change which is
>> caught by the yocto-check-layer
>>
>> Signed-off-by: Armin Kuster <akuster808@gmail.com>
>> ---
>>  meta/recipes-core/os-release/os-release.bb | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/meta/recipes-core/os-release/os-release.bb
>> b/meta/recipes-core/os-release/os-release.bb
>> index 7f3d9cb..88bf8f8 100644
>> --- a/meta/recipes-core/os-release/os-release.bb
>> +++ b/meta/recipes-core/os-release/os-release.bb
>> @@ -12,7 +12,7 @@ do_configure[noexec] = "1"
>>  
>>  # Other valid fields: BUILD_ID ID_LIKE ANSI_COLOR CPE_NAME
>>  #                     HOME_URL SUPPORT_URL BUG_REPORT_URL
>> -OS_RELEASE_FIELDS = "ID ID_LIKE NAME VERSION VERSION_ID PRETTY_NAME"
>> +OS_RELEASE_FIELDS = "ID ID_LIKE NAME VERSION VERSION_ID PRETTY_NAME BUILD_ID HOME_URL BUG_REPORT_URL SUPPORT_URL CPE_NAME"
> I'm not sure I agree with this. It would usually be a policy decision,
> e.g. a ditro config which would change this and you'd not be doing this
> with a bbappend?

Yeah,  its the feedback I was looking for. Isn't OE  core "NoDistro" and
I will be wanting to have the CPE_NAME available with out having to
define DISTRO. I am jet lagged so I am not seeing how to fix this for
meta-secruity.

- Armin



>
> Cheers,
>
> Richard
>

Re: [PATCH] os-release: add more variables to list

[richard.purdie]

On Sun, 2018-10-28 at 18:24 +0000, akuster808 wrote:
> On 10/25/18 10:25 AM, Richard Purdie wrote:
> > On Thu, 2018-10-25 at 08:20 +0100, Armin Kuster wrote:
> > > If someone extends the OS_RELEASE_FIELDS in a bbappend, it will
> > > cause
> > > the hash to change which is
> > > caught by the yocto-check-layer
> > > 
> > > Signed-off-by: Armin Kuster <akuster808@gmail.com>
> > > ---
> > >  meta/recipes-core/os-release/os-release.bb | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > > 
> > > diff --git a/meta/recipes-core/os-release/os-release.bb
> > > b/meta/recipes-core/os-release/os-release.bb
> > > index 7f3d9cb..88bf8f8 100644
> > > --- a/meta/recipes-core/os-release/os-release.bb
> > > +++ b/meta/recipes-core/os-release/os-release.bb
> > > @@ -12,7 +12,7 @@ do_configure[noexec] = "1"
> > >  
> > >  # Other valid fields: BUILD_ID ID_LIKE ANSI_COLOR CPE_NAME
> > >  #                     HOME_URL SUPPORT_URL BUG_REPORT_URL
> > > -OS_RELEASE_FIELDS = "ID ID_LIKE NAME VERSION VERSION_ID
> > > PRETTY_NAME"
> > > +OS_RELEASE_FIELDS = "ID ID_LIKE NAME VERSION VERSION_ID
> > > PRETTY_NAME BUILD_ID HOME_URL BUG_REPORT_URL SUPPORT_URL
> > > CPE_NAME"
> > 
> > I'm not sure I agree with this. It would usually be a policy
> > decision, e.g. a ditro config which would change this and you'd not
> > be doing this with a bbappend?
> 
> Yeah,  its the feedback I was looking for. Isn't OE  core "NoDistro"
> and I will be wanting to have the CPE_NAME available with out having
> to define DISTRO. I am jet lagged so I am not seeing how to fix this
> for meta-secruity.

This sounds very much like "I want to be a distro but I don't want to
be a distro".

Changing os-release just because a layer was included is a really bad
idea. Its distro policy what would be included there, including what
the cpe name might be.

Whether the nodistro defaults are right is a different question but
meta-security should not be poking there.

Cheers,

Richard

Re: [PATCH] os-release: add more variables to list

[akuster808]

On 10/29/18 10:12 AM, richard.purdie@linuxfoundation.org wrote:
> On Sun, 2018-10-28 at 18:24 +0000, akuster808 wrote:
>> On 10/25/18 10:25 AM, Richard Purdie wrote:
>>> On Thu, 2018-10-25 at 08:20 +0100, Armin Kuster wrote:
>>>> If someone extends the OS_RELEASE_FIELDS in a bbappend, it will
>>>> cause
>>>> the hash to change which is
>>>> caught by the yocto-check-layer
>>>>
>>>> Signed-off-by: Armin Kuster <akuster808@gmail.com>
>>>> ---
>>>>  meta/recipes-core/os-release/os-release.bb | 2 +-
>>>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>>>
>>>> diff --git a/meta/recipes-core/os-release/os-release.bb
>>>> b/meta/recipes-core/os-release/os-release.bb
>>>> index 7f3d9cb..88bf8f8 100644
>>>> --- a/meta/recipes-core/os-release/os-release.bb
>>>> +++ b/meta/recipes-core/os-release/os-release.bb
>>>> @@ -12,7 +12,7 @@ do_configure[noexec] = "1"
>>>>  
>>>>  # Other valid fields: BUILD_ID ID_LIKE ANSI_COLOR CPE_NAME
>>>>  #                     HOME_URL SUPPORT_URL BUG_REPORT_URL
>>>> -OS_RELEASE_FIELDS = "ID ID_LIKE NAME VERSION VERSION_ID
>>>> PRETTY_NAME"
>>>> +OS_RELEASE_FIELDS = "ID ID_LIKE NAME VERSION VERSION_ID
>>>> PRETTY_NAME BUILD_ID HOME_URL BUG_REPORT_URL SUPPORT_URL
>>>> CPE_NAME"
>>> I'm not sure I agree with this. It would usually be a policy
>>> decision, e.g. a ditro config which would change this and you'd not
>>> be doing this with a bbappend?
>> Yeah,  its the feedback I was looking for. Isn't OE  core "NoDistro"
>> and I will be wanting to have the CPE_NAME available with out having
>> to define DISTRO. I am jet lagged so I am not seeing how to fix this
>> for meta-secruity.
> This sounds very much like "I want to be a distro but I don't want to
> be a distro".
>
> Changing os-release just because a layer was included is a really bad
> idea. Its distro policy what would be included there, including what
> the cpe name might be.
This is why I think its appropriate to have at least  "CPE_NAME" added.
>
> Whether the nodistro defaults are right is a different question but
> meta-security should not be poking there.

I really didn't want to do another layer for something as simple as
"CPE_NAME" but why not.

thanks for the feedback as always.

Kind regards,

Armin

>
> Cheers,
>
> Richard
>
>
>

Re: [PATCH] os-release: add more variables to list

[richard.purdie]

On Mon, 2018-10-29 at 12:36 -0700, akuster808 wrote:
> On 10/29/18 10:12 AM, richard.purdie@linuxfoundation.org wrote:
> > This sounds very much like "I want to be a distro but I don't want
> > to
> > be a distro".
> > 
> > Changing os-release just because a layer was included is a really
> > bad
> > idea. Its distro policy what would be included there, including
> > what
> > the cpe name might be.
> 
> This is why I think its appropriate to have at least  "CPE_NAME"
> added.
> > 
> > Whether the nodistro defaults are right is a different question but
> > meta-security should not be poking there.
> 
> I really didn't want to do another layer for something as simple as
> "CPE_NAME" but why not.

Does OE have a standard "CPE_NAME" which works for everything/everyone?
Should it have?

I get the feeling I'm missing something here but you're not giving me
much to work with and the original patch didn't just add CPE_NAME...

Cheers,

Richard