Displaying 3 results from an estimated 3 matches for "replaceusewithinstr".
2008 Jun 17
4
[LLVMdev] Transforming ConstantExprs to Instructions
...ntExpr::replaceWithInstr() virtual method, which
translates all of the uses of a constantexpr with their Instruction
equivalents, whenever possible. Each of the subclasses of ConstantExpr can
implement this as appropriate. Or, thinking of it, it's probably better to
have a protected and virtual replaceUseWithInstr method which is called for
each use by replaceWithInstr().
Is this a useful addition? Is this a sane approach?
Gr.
Matthijs
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
UR...
2008 Jun 17
0
[LLVMdev] Transforming ConstantExprs to Instructions
...nstr() virtual method, which
> translates all of the uses of a constantexpr with their Instruction
> equivalents, whenever possible. Each of the subclasses of ConstantExpr can
> implement this as appropriate. Or, thinking of it, it's probably better to
> have a protected and virtual replaceUseWithInstr method which is called for
> each use by replaceWithInstr().
>
> Is this a useful addition? Is this a sane approach?
Is it possible to design the pass to work with both? The general approach
is to make stuff handle "User"s instead of Instructions. It is much more
compile tim...
2008 Jun 17
0
[LLVMdev] Transforming ConstantExprs to Instructions
...nstr() virtual method, which
> translates all of the uses of a constantexpr with their Instruction
> equivalents, whenever possible. Each of the subclasses of ConstantExpr can
> implement this as appropriate. Or, thinking of it, it's probably better to
> have a protected and virtual replaceUseWithInstr method which is called for
> each use by replaceWithInstr().
>
> Is this a useful addition? Is this a sane approach?
I can't imagine any uses for a method like that besides a pass like
yours. (This method is essentially the opposite of
ConstantFoldInstruction, and is therefore usuall...