[PATCH] selftests: Make run_tests executable

[PATCH] selftests: Make run_tests executable

[Frederic Weisbecker]

Make the main script that launches the tests to
become executable. Otherwise people need to launch
it as an argument to a new bash instance.

Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
---
 0 files changed, 0 insertions(+), 0 deletions(-)
 mode change 100644 => 100755 tools/testing/selftests/run_tests

diff --git a/tools/testing/selftests/run_tests b/tools/testing/selftests/run_tests
old mode 100644
new mode 100755
-- 
1.7.5.4

Re: [PATCH] selftests: Make run_tests executable

[Andrew Morton]

On Mon, 30 Jan 2012 18:04:29 +0100
Frederic Weisbecker <fweisbec@gmail.com> wrote:

> Make the main script that launches the tests to
> become executable. Otherwise people need to launch
> it as an argument to a new bash instance.
> 

That's a feature ;)

The x bit can easily get lost on kernel scripts.  For example, if
someone upgrades their kernel with patch(1) and the patch adds or
renames the script, the x bit gets lost.

If we were to set the x bit for the subset of people who get their
scripts direct from a git repository then those people might write and
distribute tools which fail when used by people who have lost their x
bit.  I've seen this happen at least once before.

Re: [PATCH] selftests: Make run_tests executable

[Frederic Weisbecker]

On Mon, Jan 30, 2012 at 03:51:05PM -0800, Andrew Morton wrote:
> On Mon, 30 Jan 2012 18:04:29 +0100
> Frederic Weisbecker <fweisbec@gmail.com> wrote:
> 
> > Make the main script that launches the tests to
> > become executable. Otherwise people need to launch
> > it as an argument to a new bash instance.
> > 
> 
> That's a feature ;)
> 
> The x bit can easily get lost on kernel scripts.  For example, if
> someone upgrades their kernel with patch(1) and the patch adds or
> renames the script, the x bit gets lost.
> 
> If we were to set the x bit for the subset of people who get their
> scripts direct from a git repository then those people might write and
> distribute tools which fail when used by people who have lost their x
> bit.  I've seen this happen at least once before.
> 

I see. Another way could be to chmod the run_test script from the Makefile.
Would that sound clumsy?

Re: [PATCH] selftests: Make run_tests executable

[Andrew Morton]

On Wed, 1 Feb 2012 16:24:53 +0100
Frederic Weisbecker <fweisbec@gmail.com> wrote:

> On Mon, Jan 30, 2012 at 03:51:05PM -0800, Andrew Morton wrote:
> > On Mon, 30 Jan 2012 18:04:29 +0100
> > Frederic Weisbecker <fweisbec@gmail.com> wrote:
> > 
> > > Make the main script that launches the tests to
> > > become executable. Otherwise people need to launch
> > > it as an argument to a new bash instance.
> > > 
> > 
> > That's a feature ;)
> > 
> > The x bit can easily get lost on kernel scripts.  For example, if
> > someone upgrades their kernel with patch(1) and the patch adds or
> > renames the script, the x bit gets lost.
> > 
> > If we were to set the x bit for the subset of people who get their
> > scripts direct from a git repository then those people might write and
> > distribute tools which fail when used by people who have lost their x
> > bit.  I've seen this happen at least once before.
> > 
> 
> I see. Another way could be to chmod the run_test script from the Makefile.
> Would that sound clumsy?

I suggest we remove the script and make it a makefile target instead. 
"make tests" or whatever.  As a microbonus that would remove the need
to duplicate the TARGETS definition in the two scripts.

Re: [PATCH] selftests: Make run_tests executable

[Frederic Weisbecker]

On Wed, Feb 01, 2012 at 11:54:03AM -0800, Andrew Morton wrote:
> On Wed, 1 Feb 2012 16:24:53 +0100
> Frederic Weisbecker <fweisbec@gmail.com> wrote:
> 
> > On Mon, Jan 30, 2012 at 03:51:05PM -0800, Andrew Morton wrote:
> > > On Mon, 30 Jan 2012 18:04:29 +0100
> > > Frederic Weisbecker <fweisbec@gmail.com> wrote:
> > > 
> > > > Make the main script that launches the tests to
> > > > become executable. Otherwise people need to launch
> > > > it as an argument to a new bash instance.
> > > > 
> > > 
> > > That's a feature ;)
> > > 
> > > The x bit can easily get lost on kernel scripts.  For example, if
> > > someone upgrades their kernel with patch(1) and the patch adds or
> > > renames the script, the x bit gets lost.
> > > 
> > > If we were to set the x bit for the subset of people who get their
> > > scripts direct from a git repository then those people might write and
> > > distribute tools which fail when used by people who have lost their x
> > > bit.  I've seen this happen at least once before.
> > > 
> > 
> > I see. Another way could be to chmod the run_test script from the Makefile.
> > Would that sound clumsy?
> 
> I suggest we remove the script and make it a makefile target instead. 
> "make tests" or whatever.  As a microbonus that would remove the need
> to duplicate the TARGETS definition in the two scripts.

Sounds good. But we still need the TARGET in the makefile to list the
directories. Although I can do a simple ls.