bug#62420: 29.0.60; fns-tests-collate-strings fails on Cygwin

bug#62420: 29.0.60; fns-tests-collate-strings fails on Cygwin

[Ken Brown]

The test in the subject fails on Cygwin as follows:

Test fns-tests-collate-strings condition:
     (ert-test-failed
      ((should-error
        (string-collate-equalp "xyzzy" "xyzzy" "en_DE.UTF-8"))
       :form
       (string-collate-equalp "xyzzy" "xyzzy" "en_DE.UTF-8")
       :value t :fail-reason "did not signal an error"))

The reason is that en_DE.UTF-8 is actually a valid locale on Cygwin.
[Cygwin gets its locale information from Windows.  The latter supports
the RFC 5646 locale "en-DE", which is called "English (Germany)" in the
"Region" settings.]

The following trivial patch fixes the problem:

--- a/test/src/fns-tests.el
+++ b/test/src/fns-tests.el
@@ -254,7 +254,7 @@ fns-tests-collate-strings
    (should (string-collate-equalp "xyzzy" "XYZZY" nil t))
     ;; Locale must be valid.
-  (should-error (string-collate-equalp "xyzzy" "xyzzy" "en_DE.UTF-8")))
+  (should-error (string-collate-equalp "xyzzy" "xyzzy" "en_XY.UTF-8")))
   ;; There must be a check for valid codepoints.  (Check not 
implemented yet)
  ;  (should-error


Is this OK for the emacs-29 branch?

[Note: The test will still fail on current Cygwin after this patch; but
that's because of a bug in Cygwin, which has been fixed in the
development sources for Cygwin 3.5.]

Ken

bug#62420: 29.0.60; fns-tests-collate-strings fails on Cygwin

[Eli Zaretskii]

> Date: Fri, 24 Mar 2023 10:52:45 -0400
> From: Ken Brown <kbrown@cornell.edu>
> 
> The reason is that en_DE.UTF-8 is actually a valid locale on Cygwin.
> [Cygwin gets its locale information from Windows.  The latter supports
> the RFC 5646 locale "en-DE", which is called "English (Germany)" in the
> "Region" settings.]
> 
> The following trivial patch fixes the problem:
> 
> --- a/test/src/fns-tests.el
> +++ b/test/src/fns-tests.el
> @@ -254,7 +254,7 @@ fns-tests-collate-strings
>     (should (string-collate-equalp "xyzzy" "XYZZY" nil t))
>      ;; Locale must be valid.
> -  (should-error (string-collate-equalp "xyzzy" "xyzzy" "en_DE.UTF-8")))
> +  (should-error (string-collate-equalp "xyzzy" "xyzzy" "en_XY.UTF-8")))
>    ;; There must be a check for valid codepoints.  (Check not 
> implemented yet)
>   ;  (should-error
> 
> 
> Is this OK for the emacs-29 branch?

Yes, thanks.

> [Note: The test will still fail on current Cygwin after this patch; but
> that's because of a bug in Cygwin, which has been fixed in the
> development sources for Cygwin 3.5.]

Maybe mention this in comments?  Bonus points for making it an
expected failure for Cygwin < 3.5, assuming the version of Cygwin
could be obtained by the test code.

bug#62420: 29.0.60; fns-tests-collate-strings fails on Cygwin

[Ken Brown]

On 3/25/2023 8:24 AM, Eli Zaretskii wrote:
>> Date: Fri, 24 Mar 2023 10:52:45 -0400
>> From: Ken Brown <kbrown@cornell.edu>
>> Is this OK for the emacs-29 branch?
> 
> Yes, thanks.
> 
>> [Note: The test will still fail on current Cygwin after this patch; but
>> that's because of a bug in Cygwin, which has been fixed in the
>> development sources for Cygwin 3.5.]
> 
> Maybe mention this in comments?

Done.

>  Bonus points for making it an
> expected failure for Cygwin < 3.5, assuming the version of Cygwin
> could be obtained by the test code.

I don't know how to get the version in Lisp.  I only know how to get it 
in C, using uname.  If someone can tell me how to get it in Lisp, I'll 
try to earn the bonus points.

Closing.

Ken

bug#62420: 29.0.60; fns-tests-collate-strings fails on Cygwin

[Eli Zaretskii]

> Date: Sat, 25 Mar 2023 10:27:51 -0400
> Cc: 62420-done@debbugs.gnu.org
> From: Ken Brown <kbrown@cornell.edu>
> 
> On 3/25/2023 8:24 AM, Eli Zaretskii wrote:
> >> Date: Fri, 24 Mar 2023 10:52:45 -0400
> >> From: Ken Brown <kbrown@cornell.edu>
> >> Is this OK for the emacs-29 branch?
> > 
> > Yes, thanks.
> > 
> >> [Note: The test will still fail on current Cygwin after this patch; but
> >> that's because of a bug in Cygwin, which has been fixed in the
> >> development sources for Cygwin 3.5.]
> > 
> > Maybe mention this in comments?
> 
> Done.

Thanks.

> >  Bonus points for making it an
> > expected failure for Cygwin < 3.5, assuming the version of Cygwin
> > could be obtained by the test code.
> 
> I don't know how to get the version in Lisp.  I only know how to get it 
> in C, using uname.  If someone can tell me how to get it in Lisp, I'll 
> try to earn the bonus points.

What's the value of operating-system-release return in a Cygwin build?

bug#62420: 29.0.60; fns-tests-collate-strings fails on Cygwin

[Corwin Brust]

On Sat, Mar 25, 2023 at 9:46 AM Eli Zaretskii <eliz@gnu.org> wrote:
>
> What's the value of operating-system-release return in a Cygwin build?
>
corwi@Avalon /cygdrive/c/Users/corwi
$ emacs --version; emacs -batch -eval '(message "%s" operating-system-release)'
GNU Emacs 26.3
[...]
3.0.7-338.x86_64

bug#62420: 29.0.60; fns-tests-collate-strings fails on Cygwin

[Ken Brown]

On 3/25/2023 10:46 AM, Eli Zaretskii wrote:
>> Date: Sat, 25 Mar 2023 10:27:51 -0400
>> Cc: 62420-done@debbugs.gnu.org
>> From: Ken Brown <kbrown@cornell.edu>
>> I don't know how to get the version in Lisp.  I only know how to get it
>> in C, using uname.  If someone can tell me how to get it in Lisp, I'll
>> try to earn the bonus points.
> 
> What's the value of operating-system-release return in a Cygwin build?

Perfect!  It's "3.4.6-1.x86_64".  I should be able to use that. 
Interestingly, that variable was made obsolete a couple years ago in 
commit d88e12aa19c4.

Ken

bug#62420: 29.0.60; fns-tests-collate-strings fails on Cygwin

[Eli Zaretskii]

> Date: Sat, 25 Mar 2023 11:14:57 -0400
> Cc: 62420@debbugs.gnu.org
> From: Ken Brown <kbrown@cornell.edu>
> 
> > What's the value of operating-system-release return in a Cygwin build?
> 
> Perfect!  It's "3.4.6-1.x86_64".  I should be able to use that. 
> Interestingly, that variable was made obsolete a couple years ago in 
> commit d88e12aa19c4.

We can un-obsolete it if Cygwin asks nicely ;-)

There's also the method used in report-emacs-bug--os-description,
maybe we could refactor it into a public function.

bug#62420: 29.0.60; fns-tests-collate-strings fails on Cygwin

[Ken Brown]

On 3/25/2023 11:49 AM, Eli Zaretskii wrote:
>> Date: Sat, 25 Mar 2023 11:14:57 -0400
>> Cc: 62420@debbugs.gnu.org
>> From: Ken Brown <kbrown@cornell.edu>
>>
>>> What's the value of operating-system-release return in a Cygwin build?
>>
>> Perfect!  It's "3.4.6-1.x86_64".  I should be able to use that.
>> Interestingly, that variable was made obsolete a couple years ago in
>> commit d88e12aa19c4.
> 
> We can un-obsolete it if Cygwin asks nicely ;-)

Thanks, but that's unnecessary.  It turns out that the test succeeds on 
the currently released Cygwin.  The bug that caused it to fail was 
introduced in the development sources for 3.5 and is now fixed there. 
So there's no known Cygwin release in which it fails.

But I'm glad to know about operating-system-release anyway.  Maybe it 
will be useful some other time.

Ken