search

FreddieChopin / bleeding-edge-toolchain

All-in-one script to build bleeding-edge-toolchain for ARM microcontrollers
71 stars 24 forks source link

remove duplicate lines and fix brackets in gcc-6.3.0/gcc/config/arm/t-baremetal #3

Closed dobryj closed 7 years ago

dobryj commented 7 years ago

Automatic build test will probably fail. Previous bug in the bracketing cause, that many multilibs was not compiled previously and it not take a time. Not compiled part was all lines in gcc-6.3.0/gcc/config/arm/t-baremetal following this ifneq (,$(filter armv7 armv7-r armv7-a,$(with_multilib_list))) Build after fix need more time to compile this libraries.

FreddieChopin commented 7 years ago

BTW - the multilib patch will most likely be gone when GCC 7.x is released, as now the multilib configuration is part of the mainline (finally!). So if you want to have your big-endian config upstream, then please try to get your patch into GCC too.

https://gcc.gnu.org/viewcvs/gcc/trunk/gcc/config/arm/t-rmprofile?revision=243015&view=markup&sortby=date

FreddieChopin commented 7 years ago

I'm building the toolchain locally now. As I wrote in the earlier pull request, timeouts in travis-ci are a thing that cannot be solved right now and I'm not sure it will ever be, as it would require a complex change in the script to support multistage builds, which could restart where the previous stage finished. This has almost no value for end user, so I'm reluctant to implement that /;

dobryj commented 7 years ago

Dne 26.12.2016 v 08:22 Freddie Chopin napsal(a):

BTW - the multilib patch will most likely be gone when GCC 7.x is released, as now the multilib configuration is part of the mainline (finally!). So if you want to have your big-endian config upstream, then please try to get your patch into GCC too.

https://gcc.gnu.org/viewcvs/gcc/trunk/gcc/config/arm/t-rmprofile?revision=243015&view=markup&sortby=date

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/FreddieChopin/bleeding-edge-toolchain/pull/3#issuecomment-269175762, or mute the thread https://github.com/notifications/unsubscribe-auth/AC662myQar6HFOnotON1zPyPHol1csFJks5rL2sigaJpZM4LVfIj.

I Am trying to add big-endian patch to mainstream 4 years without success. It have few reasons. First is that previous version of this patch I made in my job and my employer didn't have signed agreement of acceptance of GPL for my work. Although I try it, because GCC had exception for patches <15 lines. Without success. It is little bit complicated, therefore it is my private activity now (and I have agreement for this from my employer).

Second reason is that guys from ARM didn't have HW to made tests after compilation. Yes, it is not strong reason, but prepare complete testing platform for this take some time. And GCC ARM guys are overloaded. We are try to make support for testing platform, but we are overloaded too. We also try to found someone for BE compiler tests as external manpower (we are ready to pay it), but it is impossible found anybody in current job market.

I will try it next year again. You are right, ARM multilib support will be on the main GCC line, not on ARM specific patches. And I hope, that it will improve my chances to push this patch in mainline.