Tim Haines

> The abi break detected seems to be caused by newly inserted enum values, which should be reasonable in this case as we are adding support for a new architecture....

@bigtrak @woodard We've been down this road before. It would be a huge breakage to use the libtool conventions because Dyninst's SONAME includes both the major and minor version. Currently,...

Any motion on this since the last discussion? I'd like to do a point release (probably 12.3) so we can get this out for testing with our partners before the...

> 1. You still think that I should separate out the change to the debugging interface? I'm not sure what is the right way to pull them out of this...

@bigtrak I don't see any more failures than usual: https://bottle.cs.wisc.edu/search?dyninst_branch=PR1283

@kupsch I believe you were looking at this before you founded the stripped parsing issue. Do you have any immediate/incomplete notes here?

@mxz297 @sashanicolas Could one of you look into this?

@wcohen I'm not able to reproduce this using the current HEAD of master and gcc-9.3.0. I see $ ./locrangecheck heapsort-0.5/heapsort unable to find targ1080 # _init [1000,10e0] unable to find...

Well, that's odd. I just ran HEAD of master with gcc-10.1.0 and got the same answer as using gcc-9.3.0. I also tested Dyninst-10.1 using gcc-10.1.0 and got the same answer....

Confirmed. I see the same issue when building heapsort with clang-10.