This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Targets that don't support exception handling
- To: egcs at cygnus dot com, nick dot burrett at btinternet dot com
- Subject: Re: Targets that don't support exception handling
- From: mrs at wrs dot com (Mike Stump)
- Date: Mon, 2 Mar 1998 17:06:28 -0800
> Date: Sat, 28 Feb 1998 22:30:54 +0000
> From: Nick Burrett <nick.burrett@btinternet.com>
> 1. gcc/egcs assumes that the stack can be unwinded with very few
> instructions. RISC OS can use both a non-contiguous and a contiguous
> stack and requires a function with 30 instructions to unwind the stack.
> It is impossible to inline this function.
I don't understand this point. I'm wondering if you're just confused?
> I was hoping to release a version of egcs for RISC OS users that
> provided a C++ compiler but it seems that egcs now only supports
> targets that support exception handling.
A complete port means that you have gotten EH working, yes. However,
you can do an incomplete job and still have a useful tool. Let us
know if this isn't the case. Offhand, I would say you didn't even try
and use the sjlj exceptions. If I am wrong, let us know that didn't
`just work' for you.
> I removed the exceptions support from libgcc
Not the right way to go, but I bet you knew that.
> and have managed to get libio compiled but I've just stumbled across
> exceptions in libstdc++
It's like downhill skiiing, stop trying to go uphill, you won't make
it.