FX fxcoudert@gmail.com
Sat Sep 13 08:44:00 GMT 2014

Hi James,

Thanks for the patch. I can confirm the reasoning, and that it allows the build of trunk GCC to proceed where it previously failed. I had, in fact, independently come up with a very similar patch on my own, before finding yours online :(

The only suggestion which I might make is this one:

> Index: gcc/config/darwin-driver.c
> ===================================================================
> --- gcc/config/darwin-driver.c	(revision 215234)
> +++ gcc/config/darwin-driver.c	(working copy)
> @@ -105,7 +103,7 @@ darwin_default_min_version (unsigned int
>    const unsigned int argc = *decoded_options_count;
>    struct cl_decoded_option *const argv = *decoded_options;
>    unsigned int i;
> -  static char new_flag[sizeof ("10.0.0") + 6];
> +  static char new_flag[sizeof ("10.10.0") + 6];
>    /* If the command-line is empty, just return.  */
>    if (argc <= 1)

I don’t know why the exact size of new_flag should be sizeof(…)+6, but it makes sense to widen it with the new version.

Given the above, could a darwin maintainer please approve this patch (either with or without my addition)?


PS: James, for the sanitizer bits, they should be (IIUC) submitted upstream rather than here. To be honest, I always build without the sanitizer, because I think its integration into GCC can best be described as capricious.

More information about the Gcc-patches mailing list