Completion tests fail with patches to 11415

Completion tests fail with patches to 11415

[Bart Schaefer]

*** /tmp/zsh.ztst.out.26678     Tue May 16 09:36:31 2000
--- /tmp/zsh.ztst.tout.26678    Tue May 16 09:36:33 2000
***************
*** 1 ****
! line: {tst -x }{}
--- 1,2 ----
! line: {tst -}{}
! MESSAGE:{arg}
Test ../../zsh-3.1.6/Test/53completion.ztst failed: output differs from expected as shown above for:
 code='compdef _tst tst; _tst () { _arguments "-x" ":arg:" }'
 comptest -c "$code" $'tst -\t'
Was testing: _arguments
../../zsh-3.1.6/Test/53completion.ztst: test failed.

-- 
Bart Schaefer                                 Brass Lantern Enterprises
http://www.well.com/user/barts              http://www.brasslantern.com

Re: Completion tests fail with patches to 11415

[Sven Wischnowsky]

Bart Schaefer wrote:

> On May 17,  8:52am, Sven Wischnowsky wrote:
> }
> } } Bart Schaefer wrote:
> } 
> } > ! line: {tst -x }{}
> } > --- 1,2 ----
> } > ! line: {tst -}{}
> } 
> } That's caused by 11406. Humm.
> } 
> } What are we supposed to do now. With respect to the patch, this
> } behaviour is correct (and quite reasonable, I think). Should we only
> } update the test? Should we make _arguments not try default-option
> } completion only when it tried a *non-empty* action?
> } 
> } I think I prefer the former.
> 
> Updating the test is fine with me.  Is the test right after it (which
> previously was expected to produce identical output) also correct with
> respect to the intended post-11406 behavior?

Yes, because there is no argument specification that was `executed'.

Bye
 Sven


--
Sven Wischnowsky                         wischnow@informatik.hu-berlin.de

Re: Completion tests fail with patches to 11415

[Bart Schaefer]

On May 17,  8:52am, Sven Wischnowsky wrote:
}
} } Bart Schaefer wrote:
} 
} > ! line: {tst -x }{}
} > --- 1,2 ----
} > ! line: {tst -}{}
} 
} That's caused by 11406. Humm.
} 
} What are we supposed to do now. With respect to the patch, this
} behaviour is correct (and quite reasonable, I think). Should we only
} update the test? Should we make _arguments not try default-option
} completion only when it tried a *non-empty* action?
} 
} I think I prefer the former.

Updating the test is fine with me.  Is the test right after it (which
previously was expected to produce identical output) also correct with
respect to the intended post-11406 behavior?

Index: Test/53completion.ztst
===================================================================
@@ -121,7 +121,8 @@
  code='compdef _tst tst; _tst () { _arguments "-x" ":arg:" }'
  comptest -c "$code" $'tst -\t'
 0:_arguments
->line: {tst -x }{}
+>line: {tst -}{}
+>MESSAGE:{arg}
 
  code='compdef _tst tst; _tst () { _arguments "-x:arg:" }'
  comptest -c "$code" $'tst -x\t'

-- 
Bart Schaefer                                 Brass Lantern Enterprises
http://www.well.com/user/barts              http://www.brasslantern.com

Re: Completion tests fail with patches to 11415

[Sven Wischnowsky]

Bart Schaefer wrote:

> *** /tmp/zsh.ztst.out.26678     Tue May 16 09:36:31 2000
> --- /tmp/zsh.ztst.tout.26678    Tue May 16 09:36:33 2000
> ***************
> *** 1 ****
> ! line: {tst -x }{}
> --- 1,2 ----
> ! line: {tst -}{}
> ! MESSAGE:{arg}
> Test ../../zsh-3.1.6/Test/53completion.ztst failed: output differs from expected as shown above for:
>  code='compdef _tst tst; _tst () { _arguments "-x" ":arg:" }'
>  comptest -c "$code" $'tst -\t'
> Was testing: _arguments
> ../../zsh-3.1.6/Test/53completion.ztst: test failed.

That's caused by 11406. Humm.

What are we supposed to do now. With respect to the patch, this
behaviour is correct (and quite reasonable, I think). Should we only
update the test? Should we make _arguments not try default-option
completion only when it tried a *non-empty* action?

I think I prefer the former.

Bye
 Sven


--
Sven Wischnowsky                         wischnow@informatik.hu-berlin.de