Re: [PATCH, v3] Potential solution to librt issue.

[Paolo Carlini <paolo dot carlini at oracle dot com>]

Hi,

> With attached patch I get the same on testcase 20_util/clocks/1.cc

Chris, probably I will not be able to help much during the next day or so.

I think the problem is in the way the new if_language_spec_function works / if-language is called.

I think those spec functions are called passing a *single* C string, then automatically split to a variable number of arguments by the driver, which are finally passed to the vararg implementation. See gcc/gcc.c/version_compare_spec_function and its calls in eg, gcc/config/darwin.h (as version-compare).

If that analysis is correct, maybe we can have a working prototype by just removing the comma in the call and then working with concat inside if_language_spec_function on argv[1], argv[2]... (argv[0] is the language).

Paolo.