lmp/bb-config: use variables instead of the content

[quaresmajose]

Let's let bitbake take care of the use of the variables:

LMP_FACTORY_IMAGE
OSTREE_BRANCHNAME
GARAGE_TARGET_VERSION
GARAGE_TARGET_URL
DISTRO_VERSION

All of then have references for other variables:

MACHINE
H_BUILD
LMP_VERSION

BB_ENV_PASSTHROUGH is not needed in this case because all the varaibles needed is already defined in local.conf.
The main reason for this change is to let bitbake track and show what have changed in the cache signatureres.

For example:
When the H_BUILD change, bitbake will show diferences in all variable that is using the H_BUILD somehow and root cause of the cache signature change it's the change of all the afected variables:

H_BUILD
GARAGE_TARGET_VERSION
GARAGE_TARGET_URL
DISTRO_VERSION

With this patch it will show that there are differences in all the variables as in previoes but the root cause of the cache signature change it's just the variable H_BUILD.

[ricardosalveti]

Why changing this line?

[quaresmajose]

changed to 'EOFEOF' to not evaluate the variable when writing conf/local.conf

[ricardosalveti]

Don't we need to change bitbake to allow these env variables to be consumed as well?

[ricardosalveti]

BB_ENV_PASSTHROUGH or similar.

[MrCry0]

just a note: to avoid rewriting a default BB_ENV_PASSTHROUGH value, it would be better to use BB_ENV_PASSTHROUGH_ADDITIONS

[quaresmajose]

BB_ENV_PASSTHROUGH is not needed in this case because all the varaibles needed is already defined in local.conf.
The main reason for this change is to let bitbake track and show what have changed in the cache signatureres.

For example:
When the H_BUILD change, bitbake will show diferences in all variable that is using the H_BUILD somehow and root cause of the cache signature change it's the change of all the afected variables:

H_BUILD
GARAGE_TARGET_VERSION
GARAGE_TARGET_URL
DISTRO_VERSION

With this patch it will show that there are differences in all the variables as in previoes but the root cause of the cache signature change it's just the variable H_BUILD.

[quaresmajose]

The main intention is to facilitate debugging issue in the sstate-cache, there is no functionality changes in this PR and this only affects the bitbake cache signatures.

[MrCry0]

LGTM

[angolini]

BB_ENV_PASSTHROUGH is not needed in this case because all the varaibles needed is already defined in local.conf. The main reason for this change is to let bitbake track and show what have changed in the cache signatureres.

For example: When the H_BUILD change, bitbake will show diferences in all variable that is using the H_BUILD somehow and root cause of the cache signature change it's the change of all the afected variables:

H_BUILD
GARAGE_TARGET_VERSION
GARAGE_TARGET_URL
DISTRO_VERSION

With this patch it will show that there are differences in all the variables as in previous but the root cause of the cache signature change it's just the variable H_BUILD.

Please add this note to the PR description.

[quaresmajose]

BB_ENV_PASSTHROUGH is not needed in this case because all the varaibles needed is already defined in local.conf. The main reason for this change is to let bitbake track and show what have changed in the cache signatureres.
For example: When the H_BUILD change, bitbake will show diferences in all variable that is using the H_BUILD somehow and root cause of the cache signature change it's the change of all the afected variables:

H_BUILD
GARAGE_TARGET_VERSION
GARAGE_TARGET_URL
DISTRO_VERSION

With this patch it will show that there are differences in all the variables as in previous but the root cause of the cache signature change it's just the variable H_BUILD.

Please add this note to the PR description.

done

[ricardosalveti]

LGTM, we can land tomorrow morning to avoid possible issues when we're not around.

[quaresmajose]

LGTM, we can land tomorrow morning to avoid possible issues when we're not around.

Please note I don't have privileges to merge this repo