I believe my other patch (that outputs COMPILER to config.ini) has hidden this issue. I was getting it when I specify --rtems-compiler=clang as in
> ./waf bsp_defaults --rtems-bsps=rv64imafdc_medany --rtems-compiler=clang > > config.ini That didn't output COMPILER in config.ini. So during ./waf configure the following line: > value = cp.get(conf.variant, "COMPILER") triggered an exception, and silently fall back to "gcc", and I didn't know what went wrong. I don't think it might happen again now with my other patch, but if any of those lines triggered another exception for whatever reason, it will fall back to GCC. > def get_compiler(conf, cp, variant): > try: > value = cp.get(conf.variant, "COMPILER") > cp.remove_option(conf.variant, "COMPILER") > value = no_unicode(value) > check_compiler(conf, value) > except configparser.NoOptionError: > value="gcc" I hope that clarifies things a little bit. I am OK with this patch not being pushed. On Fri, 10 Apr 2020 at 21:38, Sebastian Huber <sebastian.hu...@embedded-brains.de> wrote: > > On 10/04/2020 22:38, Hesham Almatary wrote: > > Hello Sebastian, > > On Fri, 10 Apr 2020 at 20:23, Sebastian Huber > <sebastian.hu...@embedded-brains.de> wrote: > > On 10/04/2020 12:39, Hesham Almatary wrote: > > Don't default to gcc if the compiler can't be found > > All options should have a default. I should be possible to build a BSP > just with: > > [arch/bsp] > > This patch doesn't change that behaviour. You can still build a BSP > without providing --rtems-compiler at all and it will default to GCC. > However, this patch addresses the case when you provide an > "--rtems-compiler=blah", where "blah" can't be found nor supported. > For example, without this patch, if I pass "--rtems-compiler=clang" > and it's not supported/found, it will fallback default to GCC silently > without giving any errors. I don't think that's a desirable behaviour. > > Could you give me an example command line which shows the undesired behaviour? _______________________________________________ devel mailing list devel@rtems.org http://lists.rtems.org/mailman/listinfo/devel
