[lldb-dev] functionalities/object-file causing test problems with subsequent tests?

[Jason Molenda]

I take that back.  I think this new regression is fallout from Greg's platform changes yesterday, maybe r217963.  Yeah, that's the culrpit -if I roll r217963 back, the problem goes away.

Greg, this is yours.

(lldb) file functionalities/object-file/bin/hello-unknown-kalimba_arch4-kcc-36
Current executable set to 'functionalities/object-file/bin/hello-unknown-kalimba_arch4-kcc-36' (kalimba4).
(lldb) pla stat
  Platform: kalimba
 Connected: no
(lldb) file /bin/ls
Current executable set to '/bin/ls' (x86_64).
(lldb) r
error: the platform is not currently connected
(lldb) 

> On Sep 18, 2014, at 7:24 PM, Jason Molenda <jmolenda at apple.com> wrote:
> 
> I tried to run the testsuite today, was seeing more than 400 failures.  Running the individual test files seemed to work OK.  It looks like if I run functionalities/object-file, the following test cases will fail (at least on my Mac system here at work).  Anyone else seeing this?
> 
> Command invoked: ./dotest.py functionalities/paths functionalities/platform functionalities/plugins/commands functionalities/process_attach functionalities/process_launch functionalities/recursion functionalities/thread/state
> compilers=['clang']
> 
> Configuration: arch=x86_64 compiler=clang
> ----------------------------------------------------------------------
> Collected 30 tests
> 
> .=success F=fail E=error s=skipped x=expected-fail u=unexpected-success
> 
> .......s........xx..xxxxxx....
> ----------------------------------------------------------------------
> Ran 30 tests in 69.297s
> 
> 
> 
> and then we add in object-file to the list of tests to run:
> 
> Command invoked: ./dotest.py functionalities/object-file functionalities/paths functionalities/platform functionalities/plugins/commands functionalities/process_attach functionalities/process_launch functionalities/recursion functionalities/thread/state
> compilers=['clang']
> 
> Configuration: arch=x86_64 compiler=clang
> ----------------------------------------------------------------------
> Collected 31 tests
> 
> .=success F=fail E=error s=skipped x=expected-fail u=unexpected-success
> 
> .....FFFsF...FFFFxxFFxxxxxxFFFF
> ----------------------------------------------------------------------
> Ran 31 tests in 55.945s
> 
> FAILED (failures=14, skipped=1, expected failures=8)
> 
> 
> I thought maybe it was ending with a kalimba target sitting around or something - but that's not it, I added a target delete in the function that loads & tests files.  Not sure what's going wrong here yet but wanted to mention it in case other people are seeing this.