All of lore.kernel.org
 help / color / mirror / Atom feed
* SYSROOTS_EXCLUDEDEPS_SYSROOT changing more hashes than I expect
@ 2021-07-30 15:02 Jate Sujjavanich
  2021-07-30 15:18 ` [OE-core] " Richard Purdie
  0 siblings, 1 reply; 5+ messages in thread
From: Jate Sujjavanich @ 2021-07-30 15:02 UTC (permalink / raw)
  To: openembedded-core

[-- Attachment #1: Type: text/plain, Size: 560 bytes --]

I added meta-java to my bblayers.conf file, and the next build invalidated much of my sstate cache. This included such recipes as perl, which I thought would be unrelated to that layer. Through some digging, it appears that the addition to SYSROOTS_EXCLUDEDEPS_SYSROOT in meta-java affects that recipe.

I wonder if this variable should be added to something like BB_HASHBASE_WHITELIST. I'd think that the value in the exclusion should only affect the signature if it matches the current recipe. Is something more needed in a hash calculation?

- Jate S.

[-- Attachment #2: Type: text/html, Size: 576 bytes --]

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [OE-core] SYSROOTS_EXCLUDEDEPS_SYSROOT changing more hashes than I expect
  2021-07-30 15:02 SYSROOTS_EXCLUDEDEPS_SYSROOT changing more hashes than I expect Jate Sujjavanich
@ 2021-07-30 15:18 ` Richard Purdie
  2021-07-30 16:04   ` Jate Sujjavanich
  0 siblings, 1 reply; 5+ messages in thread
From: Richard Purdie @ 2021-07-30 15:18 UTC (permalink / raw)
  To: Jate Sujjavanich, openembedded-core

On Fri, 2021-07-30 at 08:02 -0700, Jate Sujjavanich wrote:
> I added meta-java to my bblayers.conf file, and the next build invalidated 
> much of my sstate cache. This included such recipes as perl, which I thought 
> would be unrelated to that layer. Through some digging, it appears that the 
> addition to SYSROOTS_EXCLUDEDEPS_SYSROOT in meta-java affects that recipe.
> 
> I wonder if this variable should be added to something like BB_HASHBASE_WHITELIST. 
> I'd think that the value in the exclusion should only affect the signature if 
> it matches the current recipe. Is something more needed in a hash calculation?

It definitely should not be in BB_HASHBASE_WHITELIST. Changes in output should
affect the hash and changing SYSROOTS_EXCLUDEDEPS_SYSROOT does potentially change
the output.

This is a bug in meta-java where it should only be setting this for it's own
recipes and not every recipe in the system.

Cheers,

Richard


^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: SYSROOTS_EXCLUDEDEPS_SYSROOT changing more hashes than I expect
  2021-07-30 15:18 ` [OE-core] " Richard Purdie
@ 2021-07-30 16:04   ` Jate Sujjavanich
  2021-07-30 16:57     ` [OE-core] " Richard Purdie
  0 siblings, 1 reply; 5+ messages in thread
From: Jate Sujjavanich @ 2021-07-30 16:04 UTC (permalink / raw)
  To: openembedded-core

[-- Attachment #1: Type: text/plain, Size: 548 bytes --]

The ones in meta/conf/layer.conf are much more global. This one in meta-java/conf/layer.conf certainly does not apply to perl and other irrelevant recipes.

SSTATE_EXCLUDEDEPS_SYSROOT += "\
   openjdk-8-native->(ant-native|attr-native|coreutils-native|icedtea7-native|libxslt-native|make-native|openssl-native|zip-native|unzip-native) \
"
Do you have a recommendation on how to apply this strictly within meta-java? The only way I can think of is to move the EXCLUDEDEPS to the specific recipe(s). Are there any other good ways?

- Jate S.

[-- Attachment #2: Type: text/html, Size: 600 bytes --]

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [OE-core] SYSROOTS_EXCLUDEDEPS_SYSROOT changing more hashes than I expect
  2021-07-30 16:04   ` Jate Sujjavanich
@ 2021-07-30 16:57     ` Richard Purdie
  2021-08-02 13:22       ` Jate Sujjavanich
  0 siblings, 1 reply; 5+ messages in thread
From: Richard Purdie @ 2021-07-30 16:57 UTC (permalink / raw)
  To: Jate Sujjavanich, openembedded-core

On Fri, 2021-07-30 at 09:04 -0700, Jate Sujjavanich wrote:
> The ones in meta/conf/layer.conf are much more global. This one in 
> meta-java/conf/layer.conf certainly does not apply to perl and other irrelevant recipes.
> 
> SSTATE_EXCLUDEDEPS_SYSROOT += "\
>     openjdk-8-native->(ant-native|attr-native|coreutils-native|icedtea7-native|libxslt-native|make-
> native|openssl-native|zip-native|unzip-native) \
> "
> 
> Do you have a recommendation on how to apply this strictly within meta-java? 
> The only way I can think of is to move the EXCLUDEDEPS to the specific recipe(s). 
> Are there any other good ways?

Sorry, your original email said "SYSROOT_<something>" and I read that as 
being one of the SYSROOT_* variables. I kind of filtered out the rest. 
SSTATE_* is a very different thing!

I think we need to add:

setscene_depvalid[vardepsexclude] = "SSTATE_EXCLUDEDEPS_SYSROOT"

to sstate.bbclass.

Cheers,

Richard



^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: SYSROOTS_EXCLUDEDEPS_SYSROOT changing more hashes than I expect
  2021-07-30 16:57     ` [OE-core] " Richard Purdie
@ 2021-08-02 13:22       ` Jate Sujjavanich
  0 siblings, 0 replies; 5+ messages in thread
From: Jate Sujjavanich @ 2021-08-02 13:22 UTC (permalink / raw)
  To: openembedded-core

[-- Attachment #1: Type: text/plain, Size: 196 bytes --]

Sorry about the crucial mistyping. I tested this out with a local.conf setting. After building again, adding the meta-java layer resulted in a much cleaner build on dunfell.

Thanks
- Jate S.

[-- Attachment #2: Type: text/html, Size: 208 bytes --]

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2021-08-02 13:22 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-07-30 15:02 SYSROOTS_EXCLUDEDEPS_SYSROOT changing more hashes than I expect Jate Sujjavanich
2021-07-30 15:18 ` [OE-core] " Richard Purdie
2021-07-30 16:04   ` Jate Sujjavanich
2021-07-30 16:57     ` [OE-core] " Richard Purdie
2021-08-02 13:22       ` Jate Sujjavanich

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.