M::get_base/index() always returns an R64

[eschkufz]

This isn't quite right if you have the m.addr_or() bit set. (This turns the registers in a memory operand into 32-bit registers.)

I'm not sure how to fix this given the changes that berkeley has made. Maybe turn this into two methods:

R32 get_base32() { assert if addr_or is false ... }
R64 get_base64() { assert if addr_or is true ... }

[bchurchill]

The ideal solution would be to return something of static type R (but has a tagged dynamic type of R32 or R64) and let the client check the size if it cares. This is low priority, right? I'll probably get to it next time I mess with all the x64asm types, which might not be soon, if that's ok.

[eschkufz]

That's totally fine. Very low priority.

On Fri, Dec 19, 2014 at 4:07 PM, Berkeley Churchill <
[email protected]> wrote:

The ideal solution would be to return something of static type R (but has
a tagged dynamic type of R32 or R64) and let the client check the size if
it cares. This is low priority, right? I'll probably get to it next time I
mess with all the x64asm types, which might not be soon, if that's ok.

—
Reply to this email directly or view it on GitHub
#121 (comment).

[eschkufz]

I'm right here, so I'm going to check this off real quick.

[bchurchill]

This will require tweaks to the State and SymState classes (at the least) in the get_addr() functions. Should simplify things a little, I think.

[eschkufz]

And some stuff in analysis.

I'll leave my changes here but not commit them yet. You can take this issue back over and edit stoke as necessary whenever you or someone else wants to.

[stefanheule]

You mean "not merge them"? Or why would you not commit them?

[eschkufz]

Yeah yeah. You know what I mean.

On Mon, Apr 6, 2015 at 6:35 PM, Stefan Heule [email protected]
wrote:

You mean "not merge them"? Or why would you not commit them?

—
Reply to this email directly or view it on GitHub
#121 (comment).