guix-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [PATCH v2 0/4] Add GCC cross compiler for arm-none-eabi.


From: Theodoros Foradis
Subject: Re: [PATCH v2 0/4] Add GCC cross compiler for arm-none-eabi.
Date: Sun, 23 Oct 2016 20:01:30 +0300
User-agent: Roundcube Webmail/1.0.6

On 2016-10-16 13:01, Ricardo Wurmus wrote:
Ricardo Wurmus <address@hidden> writes:

Theodoros Foradis <address@hidden> writes:

Hi,

I see that the original patches were applied in master. Did you check if
cross-binutils with 2.25.1 source works for your use case?

Yes, I’m using the regular cross-binutils.  Thanks for the hint!

Would you like me to reformat my patches for gcc-6.2.0, so that they can be applied on current master, or is there some other issue blocking that?

No, it’s good. I’m going to apply your patches shortly. (Just haven’t
found the time to do so.)

I’m sorry for the delay, but a variant of your patches is now in master
as of 569f60164920a36e6597fe25e9373f97f89e8860.

Since my patches had changed I needed to modify your patches a little to
ensure they can be applied.  I also made a few other changes:

* I kept the newlib packages as variables, not procedures, using
  package inheritance instead to override the value of xgcc.
* The toolchain generator procedure’s name (rather than the version)
  includes “-nano” when appropriate.  The version of the toolchain
  package is just the version of xgcc.
* Added the patches to the list of patches in “gnu/local.mk”

I successfully built the cross-compiler based on GCC 6 after these
changes, so everything should work as intended.

Thanks again for your contributions!

~~ Ricardo

Thanks for modifying the patches. I have tested the resulting toolchains
and they work fine.

The one issue I have observed is that newlib (and newlib-nano), is not
provided as a substitute, when installed with a toolchain containing
xgcc-6. I think that only the original newlib derivations are being
built on hydra (the ones using xgcc-4.9 as their input). How could we
get around that, so that they don't have to be built locally every time?
--
Theodoros Foradis



reply via email to

[Prev in Thread] Current Thread [Next in Thread]