test-ob-sh/session

test-ob-sh/session

[Achim Gratz]

I'm currently seeing a problem with this test (only on Win7, both
NTEmacs and Cygwin).  The test just hangs and never finishes.  All my
attempts to have it fail in the same way in an interactive sessions were
fruitless.  However, it seems that the first time a session is
initiated, it does not start at all.  The second time it then works as
expected.  Why Emacs hangs when it gets there in batch mode I can't
see...

Incidentally, sh tests will run even when not explicitly configured,
which seems odd since I can't find any code that loads ob-sh either
directly or through autoloads and thus the guard code at the top of the
test file should leave those tests undefined.


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

Factory and User Sound Singles for Waldorf Blofeld:
http://Synth.Stromeko.net/Downloads.html#WaldorfSounds

Re: test-ob-sh/session

[Eric Schulte]

Achim Gratz <Stromeko@nexgo.de> writes:

> I'm currently seeing a problem with this test (only on Win7, both
> NTEmacs and Cygwin).  The test just hangs and never finishes.  All my
> attempts to have it fail in the same way in an interactive sessions were
> fruitless.  However, it seems that the first time a session is
> initiated, it does not start at all.  The second time it then works as
> expected.  Why Emacs hangs when it gets there in batch mode I can't
> see...
>
> Incidentally, sh tests will run even when not explicitly configured,
> which seems odd since I can't find any code that loads ob-sh either
> directly or through autoloads and thus the guard code at the top of the
> test file should leave those tests undefined.
>

There are a number of tests which execute shell code blocks in
test-ob.el.  This file has no guards, so it will be run on every system.
I guess at the time I wrote these tests I assumed that every system
would support both emacs-lisp and sh code blocks.

At some point either

1. all tests in test-ob.el which run sh code blocks should be placed
   behind conditional guards so they are only loaded when shell support
   is present

2. all tests in test-ob.el which run sh code blocks should be switched
   to running emacs-lisp code blocks

Best,

-- 
Eric Schulte
http://cs.unm.edu/~eschulte

Re: test-ob-sh/session

[Achim Gratz]

Achim Gratz writes:
> I'm currently seeing a problem with this test (only on Win7, both
> NTEmacs and Cygwin).  The test just hangs and never finishes.  All my
> attempts to have it fail in the same way in an interactive sessions were
> fruitless.  However, it seems that the first time a session is
> initiated, it does not start at all.  The second time it then works as
> expected.  Why Emacs hangs when it gets there in batch mode I can't
> see...

Due to some other problems with the Cygwin snapshot I was using, I've
rolled back to an older snapshot and that problem goes away for the
Cygwin version of Emacs (but not NTemacs).  Since the test is relatively
recent, I can't say if session handling always was a problem or not.

However then there's a different problem in another test (inline shell
scripts) where it seems to skip over the second invocation or extracts
the output from the third invocation into the second result.  That has
been happening sporadically before and is also not reproduceable in
interactive testing.

> Incidentally, sh tests will run even when not explicitly configured,
> which seems odd since I can't find any code that loads ob-sh either
> directly or through autoloads and thus the guard code at the top of the
> test file should leave those tests undefined.

All my attempts so far produced a nil result on (featurep 'ob-sh), yet
clearly the three tests are configured and run even when I'm not loading
ob-sh...


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

SD adaptation for Waldorf rackAttack V1.04R1:
http://Synth.Stromeko.net/Downloads.html#WaldorfSDada

Re: test-ob-sh/session

[Achim Gratz]

Eric Schulte writes:
> There are a number of tests which execute shell code blocks in
> test-ob.el.  This file has no guards, so it will be run on every
> system.

I know, I'm not (currently) talking about these.  What puzzles me is
that the three tests in test-ob-sh get defined and run even though the
feature ob-sh is not available.  It may be that this gets somehow
subverted by having the other tests unguarded, but I can't see how.

> I guess at the time I wrote these tests I assumed that every system
> would support both emacs-lisp and sh code blocks.

Specifically a shell that is more or less POSIX compatible... even
though even tcsh passes all tests.  :-)

> At some point either
>
> 1. all tests in test-ob.el which run sh code blocks should be placed
>    behind conditional guards so they are only loaded when shell support
>    is present

That would be preferrable for tests which really depend on a shell being
available.

> 2. all tests in test-ob.el which run sh code blocks should be switched
>    to running emacs-lisp code blocks

And this would be preferred for tests which should always be run without
making any assumptions on the environment.



Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

SD adaptations for Waldorf Q V3.00R3 and Q+ V3.54R2:
http://Synth.Stromeko.net/Downloads.html#WaldorfSDada

Re: test-ob-sh/session

[Achim Gratz]

Achim Gratz writes:
> I'm currently seeing a problem with this test (only on Win7, both
> NTEmacs and Cygwin).  The test just hangs and never finishes.  All my
> attempts to have it fail in the same way in an interactive sessions were
> fruitless.  However, it seems that the first time a session is
> initiated, it does not start at all.  The second time it then works as
> expected.  Why Emacs hangs when it gets there in batch mode I can't
> see...

Emacs23 on Linux doesn't hang but fails the test anyway.


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

Wavetables for the Terratec KOMPLEXER:
http://Synth.Stromeko.net/Downloads.html#KomplexerWaves