Christophe de Dinechin
2010-Feb-03 17:26 UTC
[LLVMdev] Changes in FunctionPassManager constructor
On 3 févr. 2010, at 17:54, Eli Friedman wrote:> On Wed, Feb 3, 2010 at 8:34 AM, Christophe de Dinechin > <christophe at taodyne.com> wrote: >> What guarantees does LLVM try to provide regarding source code compatibility? > > None.And this is a good thing because... ? Christophe
David Vandevoorde
2010-Feb-03 17:39 UTC
[LLVMdev] Changes in FunctionPassManager constructor
On Feb 3, 2010, at 12:26 PM, Christophe de Dinechin wrote:> On 3 févr. 2010, at 17:54, Eli Friedman wrote: > >> On Wed, Feb 3, 2010 at 8:34 AM, Christophe de Dinechin >> <christophe at taodyne.com> wrote: >>> What guarantees does LLVM try to provide regarding source code compatibility? >> >> None. > > And this is a good thing because... ?It avoids having to deal with backward compatibility cruft? It's a pain for client code, but I suspect that for the long-term quality of the code base it's a useful decision. Daveed
Christophe de Dinechin
2010-Feb-03 18:03 UTC
[LLVMdev] Changes in FunctionPassManager constructor
Salut David, On 3 févr. 2010, at 18:39, David Vandevoorde wrote:> It avoids having to deal with backward compatibility cruft?You can do that without presenting a self-rewriting web-site as your "tutorial", can't you?> > It's a pain for client code, but I suspect that for the long-term quality of the code base it's a useful decision.I have nothing against fixing LLVM. I have something against being unable to do any kind of regression testing because I need to massively change the test along with the library. That can't really help quality. Thanks, Christophe
Reasonably Related Threads
- [LLVMdev] Changes in FunctionPassManager constructor
- [LLVMdev] Changes in FunctionPassManager constructor
- [LLVMdev] Changes in FunctionPassManager constructor
- [LLVMdev] Changes in FunctionPassManager constructor
- [LLVMdev] Changes in FunctionPassManager constructor