[llvm-dev] Line number without -Debug ?

[Irini Stavrakantonaki via llvm-dev]

I see your point. But my issue here is just to have this #Line 
information (nothing more)
Since, as you said, there is a way to do it and still not enable debug 
mode ...I would like to try it and see if I can get any benefit

--
irini

On 15/12/15 18:21, mats petersson wrote:
> My point as enabling debug mode is what causes this data to be 
> collected and stored. Since there is more work than "nothing", it is 
> not enabled when doing release builds of the compiler. Debug mode may 
> enable a few other things too, but essentially, it's purpse in life 
> [as far as I understand] is to provide store and allow retrieval of 
> debug information.
>
> And my point was "why do you need to know line-number information", 
> when you are compiling in release? Seems like you need something that 
> you shouldn't need, hence me (still) thinking it's an XY question.
>
> --
> Mats
>
> On 15 December 2015 at 16:06, Irini Stavrakantonaki 
> <istavrak at ics.forth.gr <mailto:istavrak at ics.forth.gr>> wrote:
>
>     Is this "work" heavier than enabling the debug mode?
>
>     Nop. No XY question here. In fact it is just be more useful for me
>     to have the information of line number but still be in release mode.
>     (if this is possible or in other words less "painful" for my pass)
>
>     Thanks
>
>     --irini
>
>
>     On 15/12/15 17:07, mats petersson wrote:
>>     Since there is extra "work" required to store that sort of
>>     infromation, I wouldn't want my compiler to do that unless I have
>>     specifically asked for it.
>>
>>     What is the actual problem you are trying to solve? I feel that
>>     you are asking an XY question (in other words, "I have problem X,
>>     I think solution is Y, so I ask how to do Y", rather than asking
>>     "how do I do X").
>>
>>     --
>>     Mats
>>
>>     On 15 December 2015 at 10:28, Irini Stavrakantonaki via llvm-dev
>>     <llvm-dev at lists.llvm.org <mailto:llvm-dev at lists.llvm.org>> wrote:
>>
>>         Hi Eric and thanks for your answer,
>>
>>         What I mean is that I want to find the number of line on the
>>         IR code of a specific instruction.
>>         I found out that I can use:
>>
>>             DebugLoc Loc = CurrentInstruction->getDebugLoc();
>>                 unsigned Line = Loc.getLine();
>>
>>         But this works only if DEBUGFLAG is True.
>>         I'd like to be able to get the Line number without being on a
>>         'Debug' mode. Is this possible?
>>
>>         Thanks again,
>>
>>         -irini
>>
>>
>>         On 15/12/2015 09:37, Eric Christopher wrote:
>>>         Not sure what you mean, but the answer is likely "no" as
>>>         there's nothing for anything to retrieve. Diagnostics inside
>>>         the compiler can usually get it from certain IR constructs
>>>         that have line information, but I'm not sure if that's what
>>>         you mean.
>>>
>>>         -eric
>>>
>>>         On Mon, Dec 14, 2015 at 5:54 PM Irini Stavrakantonaki via
>>>         llvm-dev <llvm-dev at lists.llvm.org
>>>         <mailto:llvm-dev at lists.llvm.org>> wrote:
>>>
>>>             Hi list,
>>>
>>>             Is it possible to retrieve the source/IR line number
>>>             without being on a debug mode?
>>>
>>>             Thank you in advance,
>>>
>>>             --
>>>             irini
>>>
>>>             _______________________________________________
>>>             LLVM Developers mailing list
>>>             llvm-dev at lists.llvm.org <mailto:llvm-dev at lists.llvm.org>
>>>             http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
>>>
>>
>>         -- 
>>         irini
>>
>>
>>         _______________________________________________
>>         LLVM Developers mailing list
>>         llvm-dev at lists.llvm.org <mailto:llvm-dev at lists.llvm.org>
>>         http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
>>
>>
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20151215/5fa56ca7/attachment.html>