Skip to contentSkip to conversation

Package inclusion

DISCUSSION Carlos Sánchez de La Lama 2 participants 5 messages Oct 15, 2015 – Oct 16, 2015

lore
Share & tools
lore
#0 lore raw

Hi all,

I have some doubts on how package inclusion in images is determined. For example, when building core-image-base, I would say that both

kernel-module-uvesafb
kernel-module-input

are recommended (through RRECOMENDS in packagegroup-base).

However, -uvesafb makes it to the final image, whereas -input does not. I double-checked on the running system under QEMU *and* reading the image manifest at tmp/deploy/images/qemux86/core-image-base-qemux86.manifest

I must be missing something. Can someone give me a pointer?

Thanks a lot, BR

Carlos

Signature · 6 lines
-Debemos anhelar no sólo la verdad, sino también la virtud -sentenció
Pitágoras-. El conocimiento que se obtiene mediante oro, y no a través de un
merecimiento virtuoso, puede apartarnos del camino recto y ser nefasto para
nosotros y para nuestro entorno.

Marcos Chicot, "El asesinato de Pitágoras" (2013)
#1 lore raw
Quoted 11 lines · On Thursday 15 October 2015 18:03:13 Carlos Sánchez de La Lama wrote:

Hi all,

I have some doubts on how package inclusion in images is determined. For example, when building core-image-base, I would say that both

kernel-module-uvesafb
kernel-module-input

are recommended (through RRECOMENDS in packagegroup-base).

However, -uvesafb makes it to the final image, whereas -input does not.

I don't see a reference to kernel-module-uvesafb in packagegroup-base.

Quoted 5 lines

I double-checked on the running system under QEMU *and* reading the image manifest at tmp/deploy/images/qemux86/core-image-base-qemux86.manifest

I must be missing something. Can someone give me a pointer?

If you would like to analyse the package dependency tree for an image one way that works is to use buildhistory:

http://www.yoctoproject.org/docs/current/ref-manual/ref-manual.html#maintaining-build-output-quality

Among other things this will give you a .dot dependency graph file (for which I can recommend xdot.py as a viewer).

Cheers,
Paul

Signature · 2 lines
Paul Eggleton
Intel Open Source Technology Centre
#2 lore raw

Hi Paul,

Quoted 9 lines
Quoted 7 lines

I have some doubts on how package inclusion in images is determined. For example, when building core-image-base, I would say that both

kernel-module-uvesafb
kernel-module-input

are recommended (through RRECOMENDS in packagegroup-base).

I don't see a reference to kernel-module-uvesafb in packagegroup-base.

You are right, my bad. I was examining several cases and I used the wrong one when writing the mail. I should have said

kernel-module-nls-utf8 (which goes into the image) kernel-module-input (which does not make it to the image)

and both appear in RRECOMENDS_packagegroup-base.

Quoted 2 lines

If you would like to analyse the package dependency tree for an image one way that works is to use buildhistory:

I checked the buildhistory (nice report, BTW), and I see

"packagegroup-base" -> "kernel-module-nls-utf8" [style=dotted];

but not a trace of kernel-module-input. I do not understand why -input gets a differnt handling than -nls-utf8. My guess is it gets excluded somehow, but I have not found why/where.

Thanks!

Carlos

Signature · 3 lines
'La vie est une longue chute, Marcus. Le plus important est de savoir tomber.'

Joël Dicker, "La Vérité sur l'affaire Harry Quebert" (2012)
#3 lore raw
Quoted 23 lines · On Friday 16 October 2015 09:38:56 Carlos Sánchez de La Lama wrote:

Hi Paul,

Quoted 9 lines
Quoted 7 lines

I have some doubts on how package inclusion in images is determined. For example, when building core-image-base, I would say that both

kernel-module-uvesafb
kernel-module-input

are recommended (through RRECOMENDS in packagegroup-base).

I don't see a reference to kernel-module-uvesafb in packagegroup-base.

You are right, my bad. I was examining several cases and I used the wrong one when writing the mail. I should have said

kernel-module-nls-utf8 (which goes into the image) kernel-module-input (which does not make it to the image)

and both appear in RRECOMENDS_packagegroup-base.

Quoted 3 lines

If you would like to analyse the package dependency tree for an image one way
that works is to use buildhistory:

Quoted 7 lines

I checked the buildhistory (nice report, BTW), and I see

"packagegroup-base" -> "kernel-module-nls-utf8" [style=dotted];

but not a trace of kernel-module-input. I do not understand why -input gets a differnt handling than -nls-utf8. My guess is it gets excluded somehow, but I have not found why/where.

You definitely have the kernel-module-input package created - right? Because it's an RRECOMMENDS, that package can actually not be produced by the kernel (i.e. because the corresponding config option is 'y' or 'n' instead of 'm').

Another much less likely possibility is that you have BAD_RECOMMENDATIONS set to include kernel-module-input. bitbake -e | grep ^BAD_RECOMMENDATIONS will tell you if that is the case.

Cheers,
Paul

Signature · 2 lines
Paul Eggleton
Intel Open Source Technology Centre
#4 lore raw

Hi Paul,

Quoted 8 lines
Quoted 4 lines

kernel-module-nls-utf8 (which goes into the image) kernel-module-input (which does not make it to the image)

and both appear in RRECOMENDS_packagegroup-base.

You definitely have the kernel-module-input package created - right? Because it's an RRECOMMENDS, that package can actually not be produced by the kernel (i.e. because the corresponding config option is 'y' or 'n' instead of 'm').

You got it. The package is not created.

I should have checked that before.

Thanks for your help, BR

Carlos

Signature · 6 lines
-Debemos anhelar no sólo la verdad, sino también la virtud -sentenció
Pitágoras-. El conocimiento que se obtiene mediante oro, y no a través de un
merecimiento virtuoso, puede apartarnos del camino recto y ser nefasto para
nosotros y para nuestro entorno.

Marcos Chicot, "El asesinato de Pitágoras" (2013)