Re: [rfc] New test directory gdb.pr gdb.gnats

Re: [rfc] New test directory gdb.pr gdb.gnats

[Michael Elizabeth Chastain]

> With this in mind, I'd like to follow up the suggestion that there be
> a directory ``gdb.pr'' (gdb.gnats?) created that had the potential to
> contain a NNNN.exp test for each PR reported in the PUBLIC bug tracking
> database.

I'm in favor of the new directory.  I'm agnostic about the name.

> Of course I'm not proposing anything silly such as there _has_ to be a
> test for each PR or that PR's can only be closed if there is a test or
> that ..... This is a community group and community groups tend to only
> do what members of the group are actually interested in doing.

As a member of the community, I enjoy writing test scripts for PR's
that I file and PR's that I close.  So if you create the directory I
will contribute test scripts.

Michael Elizabeth Chastain
<chastain@redhat.com>
"love without fear"

Re: [rfc] New test directory gdb.pr gdb.gnats

[Elena Zannoni]

Michael Elizabeth Chastain writes:
 > > With this in mind, I'd like to follow up the suggestion that there be
 > > a directory ``gdb.pr'' (gdb.gnats?) created that had the potential to
 > > contain a NNNN.exp test for each PR reported in the PUBLIC bug tracking
 > > database.
 > 
 > I'm in favor of the new directory.  I'm agnostic about the name.
 > 

I would prefer not to tie the directory name to a particular bug
tracking system, so I am all for gdb.pr(s).

Elena

[rfc] New test directory gdb.pr gdb.gnats

[Andrew Cagney]

Hello,

(someone on gdb-patches suggested this, they don't appear to have
followed it up)

GDB's bug database is slowly accumulating bugs.  If anyone ever fixes
one of these then it would nice to able to add a testcase to confirm
that each specific problem didn't reoccur.  I think it would also be
nice to not clutter gdb.base up with tests that are simply checking that
a single specific bug doesn't reoccur.  With this in mind, I'd like to
follow up the suggestion that there be a directory ``gdb.pr''
(gdb.gnats?) created that had the potential to contain a NNNN.exp test
for each PR reported in the PUBLIC bug tracking database. 

Of course I'm not proposing anything silly such as there _has_ to be a
test for each PR or that PR's can only be closed if there is a test or
that ..... This is a community group and community groups tend to only
do what members of the group are actually interested in doing.

Thinking about it, there is one thing I should probably emphasize.  I
suspect this is a new direction for GDB.  It means that we're
differentiating between tests for specific bugs VS tests confirming
certain generic functionality still works.  It could be argued that
instead of adding specific tests, we should be extending the existing
test framework.  To be honest, I think both should be occurring in
parallel.

Again, comments thoughts, suggestions, especially from the people that
started the original thread on gdb-patches.  Oh, and definitly from the
testsuite maintainers :-)

	Andrew