similar to: [LLVMdev] LangRef 'struct' patch--preliminary

On 01/11/2010 11:20 AM, Jeffrey Yasskin wrote: > Awesome, thanks! Committed as r93170 with the following change: > > s/local variables/registers/. "Local variable" refers to allocas in > LLVM, rather than %whatever SSA "variables". Excellent. I was not actually happy with that term when I wrote it, but wasn't sure of the standard terminology. It should

Awesome, thanks! Committed as r93170 with the following change: s/local variables/registers/. "Local variable" refers to allocas in LLVM, rather than %whatever SSA "variables". On Mon, Jan 11, 2010 at 1:28 AM, Dustin Laurence <dllaurence at dslextreme.com> wrote: > Here is a patch that cleans up a couple of bugs and makes what I think > are a couple of small

On 02/12/2010 09:51 AM, Chris Lattner wrote: > I think that the point is that you can define your own standard runtime interfaces: > > void *myopen(const char*path) { > return fopen(path, ...); > } Maybe my experience hand-coding LLVM will actually be of some help. What I did for this case is what I think Chris is suggesting--I have a .c file with functions that return

On 01/02/2010 11:24 AM, Bob Wilson wrote: > Yes, that is correct. It is supported in the trunk sources, but it has > not yet been released. Hmm. Would the same also be true of the "inbounds" keyword for GEP? It doesn't seem to be recognized ("expected type"). Dustin

On 01/08/2010 09:17 PM, Nick Lewycky wrote: > Try using 'internal' linkage instead of 'linkonce'. That did it, thanks. --- gemini:~/Projects/LLVM/Tests/Inline(0)$ cat testInline.optdis.ll ; ModuleID = 'testInline.optbc' define i32 @main(i32 %argc, i8** nocapture %argv) nounwind readnone { ret i32 42 } gemini:~/Projects/LLVM/Tests/Inline(0)$ --- > If you're

On 01/11/2010 02:33 PM, Chris Lattner wrote: > Grey boxes are fine with me, OK, until told differently I will assume that the preferred outcome is for all code examples I touch to be put in that format if they aren't already. Dustin

I have yet another question that I believe also stems from deep ignorance of the linkage types. How do you declare a global variable without defining it? The IR ref. clearly indicates that you can do this, but it looks like one of the many "too obvious to mention" things that I struggle with. It's easy with functions, of course: "declare @foo" in the header and

On 01/09/2010 10:00 AM, Samuel Crow wrote: > > Always inline is the closest to a preprocessor macro you can get in > LLVM Assembly since it doesn't have a preprocessor at all. Mine does. :-) > ...LLVM does > aggressive inlining for functions used only once so those instances > don't require specification as alwaysinline. What I'm trying to do is understand the

Thanks everyone, a set of wrapper routines it will be then. Dustin, are the routines you wrote open source or do you know if there is already a project that provides such a portable interface to libc for LLVM? If not, I'll write my own routines, but if there is a way to adopt a common standard or avoid reinventing the wheel I'm all for it. Mike

On Jan 11, 2010, at 11:31 AM, Dustin Laurence wrote: > > If these patches are useful I'll send more, but I should know one > thing. > I notice that the example code in the LangRef is not formatted > consistently; sometimes in a grey box, sometimes just inline. My > guess > is the preferred format changed at some point and older ones are just > not updated yet. I

Here is a patch for LangRef.html that adds a section for 'external' linkage. It probably needs love from someone with more knowledge, but perhaps the patch will motivate that person to improve it. Dustin -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: LangRef.patch URL:

On 01/13/2010 01:52 PM, Duncan Sands wrote: > as I mentioned in another email, unwind is not completely unsupported: > it does work for rethrowing an exception. Good point. Not understanding how languages implement exceptions under the hood, I lose the nuances that should be in a reference document. How's this version? Dustin -------------- next part -------------- An embedded and

Hello Dustin, Always inline is the closest to a preprocessor macro you can get in LLVM Assembly since it doesn't have a preprocessor at all. LLVM does aggressive inlining for functions used only once so those instances don't require specification as alwaysinline. --Sam ----- Original Message ---- > From: Dustin Laurence <dllaurence at dslextreme.com> > Cc: llvmdev at

Here is a small doc patch based on answers from the list and from the links mentioned. For stylistic consistency I've followed the language in the va_arg description for the analogous situation. Dustin -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: LangRef.unwind.patch URL:

On Sat, Jan 2, 2010 at 2:31 PM, Dustin Laurence <dllaurence at dslextreme.com> wrote: > On 01/02/2010 11:24 AM, Bob Wilson wrote: > >> Yes, that is correct.  It is supported in the trunk sources, but it has >> not yet been released. > > Hmm.  Would the same also be true of the "inbounds" keyword for GEP?  It > doesn't seem to be recognized

On Sat, Jan 9, 2010 at 12:57 PM, Dustin Laurence <dllaurence at dslextreme.com> wrote: > I have yet another question that I believe also stems from deep > ignorance of the linkage types.  How do you declare a global variable > without defining it?  The IR ref. clearly indicates that you can do > this, but it looks like one of the many "too obvious to mention" things

Hi Dustin, > Here is a small doc patch based on answers from the list and from the > links mentioned. For stylistic consistency I've followed the language > in the va_arg description for the analogous situation. as I mentioned in another email, unwind is not completely unsupported: it does work for rethrowing an exception. Ciao, Duncan.

On 01/08/2010 02:10 PM, John McCall wrote: > 'llc' is an IR-to-assembly compiler; at -O3 it does some pretty neat > machine-code and object-file optimizations, but it does not apply > high-level optimizations like CSE or inlining. 'opt' is the tool > which does IR-to-IR optimization. A vital clue, but I'm still not getting it: --- gemini:~/Projects/Nil/nil(0)$

Hello Dustin, Alwaysinline is not a hint. It forces something inline that wouldn't have otherwise been as long as the linkage type permits it. (You just ran into a situation where linkage did not permit it.) Personally, I don't see the need for a preprocessor in most circumstances. If you need to do type substitution you can use an opaque type. The only reason for conditional

On 04/06/2010 11:45 AM, Tanya Lattner wrote: > While lack of linux testing of the testsuite is a problem, this is > not why the release is slipping. We need more people to fix bugs. It > can't be the same people fixing the bugs for every release. > > We need more people in the community to help and I can't see any way > of sugar coating this message. I don't know