Related articles |
---|
A small LLVM tutorial: debugging information vincent@famillebelliard.fr (vincent_belliard) (2012-05-25) |
Re: A small LLVM tutorial: debugging information echristo@gmail.com (Eric Christopher) (2012-05-29) |
Re: A small LLVM tutorial: debugging information vincent@famillebelliard.fr (Vincent Belliard) (2012-05-31) |
From: | Vincent Belliard <vincent@famillebelliard.fr> |
Newsgroups: | comp.compilers |
Date: | Thu, 31 May 2012 00:10:16 -0700 (PDT) |
Organization: | Compilers Central |
References: | 12-05-029 |
Keywords: | LLVM |
Posted-Date: | 01 Jun 2012 01:26:29 EDT |
Yes, of course, using the LLVM builder is much easier (or, at least,
quicker to start). In my case, I needed to do all with my language
Entity so I couldn't use the builders. I did my own which are simpler
than the LLVM ones as they cover only what I needed.
This small tutorial is still useful if you use the builders. May some
topics won't be necessary but it's always good to understand deeply
what you are using.
Nevertheless, in the next version, I will add your comment.
vincent
PS: I made a very small language close to C but close to LLVM. It's very
useful to write small libraries functions with a deep control over the
generated code without writting directly in LLVM.
Here is an example of a string compare with my zero padded strings:
i1 @equals_string_contents(%_self: %string_content*, %_cmp: %string_content*)
{
var size_ref: i64 = %_self->size ;
var size_cmp: i64 = %_cmp->size ;
if (size_ref ne size_cmp) return 0 ;
var count: i64 = size_ref >> 3 ;
var ref: i64* = <i64*>%_self->data ;
var cmp: i64* = <i64*>%_cmp->data ;
loop
{
phi phi_count: i64 = count | next_count ;
phi phi_ref: i64* = ref | next_ref ;
phi phi_cmp: i64* = cmp | next_cmp ;
if (phi_count eq 0) return 1 ;
if (*phi_ref ne *phi_cmp) return 0 ;
var next_count: i64 = phi_count - 1 ;
var next_ref: i64* = phi_ref + 1 ;
var next_cmp: i64* = phi_cmp + 1 ;
}
}
Return to the
comp.compilers page.
Search the
comp.compilers archives again.