Re: [PATCH v3] Add regression tests for ctf reading

[Guillermo Martinez <guillermo.e.martinez@oracle.com>]

On Monday, December 13, 2021 10:11:54 AM CST Dodji Seketeli wrote:
> Hello Guillermo,
Hello Dodji

> I ran the tests and found some issues on my system.  I have described
> them below.
> 
> "Guillermo E. Martinez via Libabigail" <libabigail@sourceware.org> a
> écrit:
> 
> [...]
> 
> > 	* tests/data/test-read-ctf/test-enum-many.o.hash.abi: Likewise.
> 
> [...]
> 
> >  tests/data/test-read-ctf/test-enum-many.c     |  10 +
> 
> [...]
> 
> 
> > +++ b/tests/test-read-ctf.cc
> 
> [...]
> 
> 
> > +  {
> > +    "data/test-read-ctf/test-enum-many.o",
> 
> This file is not present in the patch.
> 
> So running the test yields:
> 
> $ ./build/tests/runtestreadctf 
> NOT PROCESSED TYPE 18446744073709551615
> NOT PROCESSED TYPE 18446744073709551615
> NOT PROCESSED TYPE 18446744073709551615
> NOT PROCESSED TYPE 18446744073709551615
> NOT PROCESSED TYPE 18446744073709551615
> NOT PROCESSED TYPE 18446744073709551615
> NOT PROCESSED TYPE 18446744073709551615
> NOT PROCESSED TYPE 18446744073709551615
> NOT PROCESSED TYPE 18446744073709551615
> NOT PROCESSED TYPE 18446744073709551615
> NOT PROCESSED TYPE 18446744073709551615
> NOT PROCESSED TYPE 18446744073709551615
> NOT PROCESSED TYPE 18446744073709551615
> NOT PROCESSED TYPE 18446744073709551615
> NOT PROCESSED TYPE 18446744073709551615
> NOT PROCESSED TYPE 18446744073709551615
Those message are corner cases that i future patches will be fixed. e.g:
   https://sourceware.org/pipermail/libabigail/2021q4/003994.html

> runtestreadctf: /home/dodji/git/libabigail/patches/tests/test-read-ctf.cc:277: virtual void test_task_ctf::perform(): Assertion `__abg_cond__' failed.
> Abandon (core dumped)
> $
> 
> The program aborts because "tests/data/test-read-ctf/test-enum-many.o"
> was not found.
Upss, so sorry! 
> You can detect this kind of problems by running 'make distcheck-fast'
> before submitting the patch.  It's super helpful.
OK, thanks for the advice, I will do so
> 
> Also, please be aware that all new files to the tests/data/* directory
> sub-tree have to be added to the tests/data/Makefile.am so that these
> files are added to the final distribution tarball.  Otherwise, "make
> distcheck-fast" will fail.
Ok I'll keep an eye on this.
> [...]
> 
> diff --git a/tests/data/test-read-dwarf/PR27700/test-PR27700.o b/tests/data/test-read-dwarf/PR27700/test-PR27700.o
> deleted file mode 100644
> index 3f084d21..00000000
> Binary files a/tests/data/test-read-dwarf/PR27700/test-PR27700.o and /dev/null differ
> diff --git a/tests/data/test-read-dwarf/test-PR26568-1.o b/tests/data/test-read-dwarf/test-PR26568-1.o
> 
> [...]
> 
> > +++ b/tests/test-read-dwarf.cc
> 
> [...]
> 
> >    {
> > -    "data/test-read-dwarf/test-PR26568-1.o",
> > +    "data/test-read-common/test-PR26568-1.o",
> 
> This file data/test-read-common/test-PR26568-1.o doesn't exist in the
> patch.
> 
> As a result, running runtestreaddwarf yields:
Upps, sorry again.

> $ build/tests/runtestreaddwarf 
> runtestreaddwarf: /home/dodji/git/libabigail/patches/tests/test-read-dwarf.cc:590: virtual void test_task_dwarf::perform(): Assertion `__abg_cond__' failed.
> Abandon (core dumped)
> $
> 
> It aborts because at least the file
> "tests/data/test-read-common/test-PR26568-1.o" is not present.  This too
> can be detected by running "make distcheck-fast".
OK
> [...]
> 
> I hope this helps.  Thanks.
sure thing!!
> Cheers,

Thanks so much for your comments, I'll send patch v4 soon,
Guillermo