[LLVMdev] Upstreaming PNaCl's IR simplification passes

[Joshua Cranmer 🐧]

On 3/6/2014 8:09 PM, Mark Seaborn wrote:
> On 4 March 2014 15:17, Chandler Carruth <chandlerc at google.com 
> <mailto:chandlerc at google.com>> wrote:
>
>     Given that the bitcode is stable, I don't understand why this is
>     important.
>
>
> Is the bitcode format stable now?  I heard talk that LLVM is trying to 
> do this now, but I don't remember seeing an llvmdev thread stating 
> that for sure.  Was there a thread about it that I missed?  I just 
> remember hearing complaints last year that the format was still 
> getting changed. :-)

The addition of attribute groups is the last major change to IR that I 
recall, and the final draft on it that I could find contains this note:
"As usual, LLVM 3.3 will be able to read and auto-upgrade previous 
bitcode and `.ll' files." A naive search for stability brings up a reply 
to the original LLVM IR is a compiler IR message, back in 2011, and 
subsequent references appear to follow back to the train. Given that all 
recent references I recall (note: recall bias may be present) to 
changing the bitcode format have brought up maintaining the ability to 
read old files, I've made a working assumption that the IR is going to 
be stable unless something drastic (like delete an instruction) is done, 
and that such an occurrence is extremely unlikely to happen. Metadata 
and intrinsics may be a different story, though.

-- 
Joshua Cranmer
Thunderbird and DXR developer
Source code archæologist

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20140306/60121971/attachment.html>