Re: [DynInst_API:] commit: further fixes for version definitions

[Bill Williams <bill@xxxxxxxxxxx>]

On 08/25/2015 06:43 PM, Josh Stone wrote:
> On 08/25/2015 10:49 AM, Bill Williams wrote:
> > 9.0.2 has another issue with out of source builds: version.h was not
> > properly in the include path or the install path. Fixed.
> >
> > USERS: I am not tagging 9.0.3 for another 24 hours, minimum. Please send
> > any bug reports on git-head to the list.
> >
> > DEVELOPERS: DO NOT commit anything that is not specifically a bug fix
> > for a bug posted to the list to master until 9.0.3 is tagged.
> I'm doing some rpmdiffs, and noticed a few things.
>
> symlite is no longer part of the build - is that intentional? In 8.2,
> both symlite and symtabAPI would build.  I think this is probably fine.
Intentional, and I should probably clean the configuration option for 
it. The entire point of optimizing symtab was to deprecate symlite. 
Matt, I trust that we're okay to do so?
> dynC lost its soname. The attached patch makes it use the new
> dyninst_library() to get it back.
Thanks. Will sanity check, but assuming there's no smoke, it should be 
fine to push.
> In the testsuite, I'm not seeing any proccontrol pic mutatees anymore.
> i.e. there's no pc*_pic_none. I don't know what happened there. On the
> other hand, a few pc_*.stat_* mutatees showed up new.
That's deliberate; Matt and I discussed this in the context of adding 
another dimension to the proccontrol mutatees and the overly large 
current set of dimensions. Our belief is that there shouldn't be 
anything proccontrol-specific that depends on PIC/PIE vs. not--we're not 
doing any sophisticated parsing, we care not about loads and stores, etc.
> That's all I've got. Sorry I didn't test this so detailed earlier, but
> we'll get release polishing behind you soon enough...

--
--bw

Bill Williams
Paradyn Project
bill@xxxxxxxxxxx