[LLVMdev] Announcing: LLVM 2.8 Release Schedule

Török Edwin edwintorok at gmail.com
Wed Aug 4 06:01:17 PDT 2010


On Tue, 3 Aug 2010 23:57:05 -0700
Bill Wendling <wendling at apple.com> wrote:

> Good news, everybody!
> 
> It's that time of year again. We are going to release LLVM 2.8! I'm
> taking over for Tanya to give her a much needed break. I can only
> hope to perform as well as she has. This is my first time as release
> manager, so bear with me if things don't go smoothly.
> 
> This message is mostly to give you an idea of the schedule we're
> planning on. It's an aggressive schedule, but one that I think is
> doable:
> 
> 8/6 - Branch

Hi Bill,

I think llvm.org should be updated with the new schedule too.

> 8/9 - Pre-release1 out (binaries and tar balls), testing begins
> 8/16 - Pre-release1 testing ends
> 8/22 - Pre-release2 testing begins
> 8/29 - Pre-release2 testing ends
> 8/30 or 9/6 - Release!
> 
> The August 6th branch is coming up soon.

I ported LibClamAV from LLVM 2.7+custom patches to LLVM 2.8
(vanilla).
Not many API calls needed changes, which is nice.
There is one regression that I found so far: PR7712.
No memory leaks in our make check, which is nice again.

I link with the .a files, didn't test the new shared library (as
Debian's package doesn't built it at all).

I have added my comments to ReleaseNotes.html, here is a summary of
what I had to change:
1. the iterator cast that was already documented
2. the CallInst getOperator() that was already documented (switched to
using CallSite).
3. llvm.mem{cpy,move,set}.* intrinsics take 1 additional param.
I used UpgradeIntrinsic* to make them work, since
Intrinsic::getDeclaration was changed in an incompatible way too.

Are there any other intrinsics that changed?

4. SetCurrentDebugLocation takes a DebugLoc now
5. VISIBILITY_HIDDEN is gone
6. SMDiagnostic takes different parameters
7. APIs renamed:
llvm_report_error -> report_fatal_error
llvm_install_error_handler -> install_fatal_error_handler
llvm::DwarfExceptionHandling -> llvm::JITExceptionHandling
8. disable calls to memory_barrier. this is not correct for SMP but at
least its not crashing.

I'll probably have more to add to this list once I try to port our
bytecode compiler.

> Leading up to that date,
> please make sure that the tree is stable – no regressions, etc. If
> you have a major feature which must make it into 2.8, please check
> them in by this time; the sooner the better.

Only one for me is the patch to fix PR6758, since without it PPC JIT
doesn't work for long-running applications at all.
I have a new patch for it that I use in ClamAV, will try if it passes
LLVM's make check (and PR6902) and commit it.

> If you aren't certain
> whether your feature is necessary for 2.8, please ask the appropriate
> code owner. If you are planning major changes which aren't needed for
> 2.8, please hold them off until after we branch.
> 
> TL;DR version: Don't break things until after August 6th. :-)
> 
> I will send out more information about the release later this week.
> 
> Previous platform testers have all ready been contacted about the
> schedule. If you would like to volunteer for testing please contact
> me and let me know.
> 
> As mentioned before the schedule is aggressive. The August 30th
> deadline is only an estimate and we are currently thinking of pushing
> it back a week.

P.S. Would have been nice to know about this schedule more than 2 days
ahead.

Best regards,
--Edwin




More information about the llvm-dev mailing list