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

Re: Bug#728388: ITP: gcc-arm-none-eabi -- GCC cross compiler for ARM Cortex-A/R/M processors

Hi Agustin,

On Fri, 01 Nov 2013 08:01:55 -0300, Agustin Henze <tin@debian.org> wrote:
> On 10/31/2013 07:37 PM, Keith Packard wrote:
> > I've got gcc and binutils patches in the new queue already; a small
> > patch on top of the existing Debian 4.8 sources, so the source package
> > is tiny.
> Hi Keith, those are big news! I took a look in the source package and I can
> see some constraints and disadvantages in comparison with gcc-arm-embedded
> but I think that we can work together (if you want) and improve your
> package.
> Constraints
> * Only C language support. Do you disable c++ for some reason?
> * Link time optimization disabled. Why?
> Disadvantages
> * It's delivered without any prebuilt libc. You can't build a project "out
> of the box".

As I understand it (based on discussions we had at DebConf13), Keith's
intention for his packages was to provide bare-metal support for M0s mainly
(think Altus Metrum's rocket support boards). Hence only C support and no
supplied libc; the latter was a bit of a sticking point since it's nice to
have a libc but it's hard to know which one to package. We weren't sure at
the time how much work was involved in supporting multiple libcs with a
single package set.

> * It's a lot of work maintain a toolchain working well. Therefore maintain
> in a good shape the package applying some patches "by hand" can become in a
> hard work.

This is why it makes sense to build on the excellent work by the gcc
packaging team, and just carry patches required temporarily, with the general
idea that they should be upstreamed eventually. (I was led to believe this is
the intention of at least some people at Linaro with regard to the
gcc-arm-none-eabi toolchain.)

In some cases Matthias wouldn't be against hosting patches in the main
gcc-source packages too, although I can't speak for him as far as
gcc-arm-none-eabi support is concerned.



Attachment: signature.asc
Description: PGP signature

Reply to: