Build regression against Styhead

[prevas-emkan]

Hi,

After commit 1e3228f ("imx-atf: Adjust for compiler virtual renames in oe-core"), the master branch is no longer compatible with Styhead. In meta-poky, commits

• 6b1e65b7d6ce ("base: Switch virtual/cross-XXX to be under recipe specific providers")
• 90e0a0f7f453 ("classes/recipes: Switch virtual/XXX-gcc to virtual/cross-cc (and c++/binutils)")

are in master and not styhead and thus meta-freescale's master branch is no longer compatible with styhead.

[angolini]

I'm confused why it's a regression and not normal development.
Would you mind preparing a PR for fixing the compatibility?

[Villemoes]

There are two problems. One on master, which could be fixed by a PR dropping the claim of styhead compatibility.

The other is the lack of a styhead branch; that cannot be created by a PR, so someone with push access should create a styhead branch based on the last commit which was actually compatible with styhead, d615bad . That branch should probably then also gain a commit dropping the claimed walnascar compat, but that again can't be done as a PR until the styhead branch actually exists.

[prevas-emkan]

I'm confused why it's a regression and not normal development.

I mean a regression in the sense that this layer claims to be compatible with styhead via LAYERSERIES_COMPAT, which is no longer true after commit 1e3228f1da57 ("imx-atf: Adjust for compiler virtual renames in oe-core").

Would you mind preparing a PR for fixing the compatibility?

I can create a PR to drop styhead from LAYERSERIES_COMPAT, but then none of the current branches works with styhead. Someone should create a styhead branch from the latest compatible commit with styhead, d615badf93ec, which I don't have access to.

[angolini]

https://github.com/Freescale/meta-freescale/tree/styhead

I only created the branch. Could you, please test it?

I don't have time right now to work on this, but I can come back to this at the end of the week (as branching a new version takes more than only creating this branch)

[Villemoes]

We will automatically test this tonight, but we do have a nightly build from about a week ago when master was exactly at that commit which succeeded, so now that we can point our nightly builds at the styhead branch rather than master, it should work again. I'll report back tomorrow.

[Villemoes]

As expected, the nightly build succeeded. I've created two follow-up PRs (#2085 , #2086 ) adjusting the compatibility claimed on styhead and master.

This issue can be closed, but I'm not allowed to use the "Close with comment" button.