similar to: [LLVMdev] How to read v3.3 dbg metadata using v3.4 LLVM

Thanks Eric, but could you give me a little bit more hints or pointers please? I looked into DebugInfo.h, but I'm still not sure how to start. It sounds like I'd have to somehow manually extract metadata nodes from an instruction. Thanks, JS On Wed, Jan 22, 2014 at 10:14 PM, Eric Christopher <echristo at gmail.com>wrote: > This is likely going to be difficult if possible. I

Török Edwin wrote: > [snip] >>> Something like this (you can of course cache TheMetadata and MDDbgKind) >>> >>> llvm::MetadataContext *TheMetadata = M->getContext().getMetadata(); >>> MDDbgKind = TheMetadata->getMDKind("dbg"); >>> if (MDDbgKind) { >>> if (MDNode *Dbg = TheMetadata->getMD(MDDbgKind, I)) { >>>

On 03/16/2010 05:30 PM, John Criswell wrote: > Török Edwin wrote: >> [snip] >>>> Something like this (you can of course cache TheMetadata and MDDbgKind) >>>> >>>> llvm::MetadataContext *TheMetadata = M->getContext().getMetadata(); >>>> MDDbgKind = TheMetadata->getMDKind("dbg"); >>>> if (MDDbgKind) {

Török Edwin wrote: > [snip] > > Ah, the method got moved to the instruction itself! > > dbgKind = Context->getMDKindID("dbg"); > if (MDNode *Dbg = I->getMetadata(dbgKind)) { > ... > Thanks! This appears to work. I also have code that looks up debug information for GlobalVariables and regular LLVM Value *'s. For the former, I think I can look up

On 03/16/2010 06:52 PM, John Criswell wrote: > Török Edwin wrote: >> [snip] >> >> Ah, the method got moved to the instruction itself! >> >> dbgKind = Context->getMDKindID("dbg"); >> if (MDNode *Dbg = I->getMetadata(dbgKind)) { >> ... >> > > Thanks! This appears to work. Ok. > > I also have code that looks up

Devang Patel wrote: > Hi John, > > On Wed, Nov 4, 2009 at 12:04 PM, John Criswell <criswell at uiuc.edu> wrote: > >> Dear All, >> >> 1) I recall reading somewhere that a few optimizations in LLVM 2.6 strip >> away debug information when such information interferes with >> optimization. Is this correct, >> > > Yes. > >

We have added our own proprietary target to LLVM and up until recently it was configured only for this target. I now build it to include additional targets and this all works fine, I just need to specify the associated triple when compiling and the correct code-generator is selected. But I would like to have it default to the triple for our processor when no triple is specified, and while I can

Before I undertake this upgrade I thought I would see if anyone has any advice on how to do this on a production system. Maybe someone has already "fought this dragon". Current config Gluster V3.0.0. This has been in production for over 16 months: 2 servers with 8 x 2TB hard drives (bricks) replicated svr1:vol1 <-> srv2:vol1 -> rbrick1 svr1:vol2 <-> srv2:vol2 ->

Sorry, I have been away for a while. has anyone fixed CD2 fotr CentOS v3.3 ? -- Luis

Török Edwin wrote: > On 03/16/2010 05:00 PM, John Criswell wrote: > >> Dear LLVMers, >> >> I'm updating some code to use the new LLVM 2.7 API. One piece of this >> code uses the findStopPoint() function to find the source filename and >> line number information of an instruction. >> >> What is the best way to do this under LLVM 2.7 now that

Hi, getMetadata(“dbg”) returns NULL for such a small program as follows: int main(){ char a[10], b[2]; if(a[0] != 0 && a[1] != 0){ strcpy(a, b); } return 0; } The compiling command is: llvm-gcc --emit-llvm -g -c ./src.c I also tried the solutions I found through Google: http://stackoverflow.com/questions/14943447/llvm-line-number-of-an-instruction

Hi, I am creating a function and ran into the problem of selecting a column from a dataset. It seems as though the $ function (as in data$columnname) does not apply in the function. In simplified version: This works: testf2<-function(data,columnnumber){print(data[,columnnumber])} But this doesn't: testf<-function(data,column){print(data$column)} Even though the first solution works,

Dear LLVMers, I'm updating some code to use the new LLVM 2.7 API. One piece of this code uses the findStopPoint() function to find the source filename and line number information of an instruction. What is the best way to do this under LLVM 2.7 now that the stop point intrinsic has been removed? It appears that the debug information is attached as metadata, but what is the easiest way

Martin hi. It seems like you have in fact understood the offered solution, and its disadvantages. The thing is, that as far as I understand Tblgen (and I think I do :-) ), The idea behind it is that records only have state (i.e., fields), and not a behavior (i.e., methods) - so dynamic evaluation can only be achieved by built-in functions. (e.g., !if(), !foreach() etc...). Maybe you can try

On 03/16/2010 05:00 PM, John Criswell wrote: > Dear LLVMers, > > I'm updating some code to use the new LLVM 2.7 API. One piece of this > code uses the findStopPoint() function to find the source filename and > line number information of an instruction. > > What is the best way to do this under LLVM 2.7 now that the stop point > intrinsic has been removed? It

On 03/16/2010 05:21 PM, John Criswell wrote: > Török Edwin wrote: >> On 03/16/2010 05:00 PM, John Criswell wrote: >> >>> Dear LLVMers, >>> >>> I'm updating some code to use the new LLVM 2.7 API. One piece of >>> this code uses the findStopPoint() function to find the source >>> filename and line number information of an

Hello Everybody, I have pretty basic question abut REXML. I want to put the String "John doe" in titles array, if attribute name is equal to "93". Please see the code below. I could not find how to do it.. any help would be very appriciated. Here is the XML i am trying to parse <entrydata columnnumber="2" name="93"> <text>John

Hello everyone, I've got issues getting Zend Guard Loader v3.3 running. Tried several installations, doesn't work for any. Once I install Zend Guard Loader my PHP CLI binary starts to segfault *at exit*. Same is true for apache processes with mod_php enabled: Works until you SIGHUP your apache processes. Doing that your webserver quits with a segfault. Can anyone reproduce that

Hi, My pass scans call instructions for standard C library calls. For some libc functions, however, LLVM uses intrinsics instead. For example, I see that my memcpy calls are replaced by the llvm.memcpy.* intrinsics. This is not a problem because I can simply look for llvm.memcpy calls when scanning for memcpy calls. The problem arises when LLVM implicitly inserts llvm.memcpy intrinsics into my

Thanks Elior, Sorry for the delay - emergencies never seem to go away in compiler development ;-) That sounds like an interesting approach, a kind-of pre-processor for TD files. But will this still not result in a constant, although externally provided, or have I misunderstood? What I would like to do is compute a different value depending on which '-target-cpu <cpu>' option was