tramp-test42-auto-load failure

tramp-test42-auto-load failure

[Juanma Barranquero]

tramp-test42-auto-load fails on Windows because remote directories are
not being tested, for lack of a suitable default. So
'tramp-test-temporary-file-directory' is set to "NUL", and
(file-remote-p "NUL") = nil.

In fact, the test used to have

   (skip-unless (tramp--test-enabled))

until commit ad99550610542f8333dc964b3bef2b9e6ad7b0f6, where it was removed.

I've sort of "fixed" it by making the test expect a failure when
'tramp--test-enabled' returns nil, because I didn't want to second
guess that commit's author, Michael Albinus.

Yet I think reverting at least part of that change would be a better
fix, because now it is "expected to fail" on Windows, while the truth
is that the test is not failing; it should simply be skipped.

Re: tramp-test42-auto-load failure

[Michael Albinus]

Juanma Barranquero <lekktu@gmail.com> writes:

Hi Juanma,

> Yet I think reverting at least part of that change would be a better
> fix, because now it is "expected to fail" on Windows, while the truth
> is that the test is not failing; it should simply be skipped.

I will check what's up once I could find an MS Windows machine. This
will take some days; next week I'm not in office.

Best regards, Michael.

Re: tramp-test42-auto-load failure

[Michael Albinus]

Michael Albinus <michael.albinus@gmx.de> writes:

Hi Juanma,

>> Yet I think reverting at least part of that change would be a better
>> fix, because now it is "expected to fail" on Windows, while the truth
>> is that the test is not failing; it should simply be skipped.
>
> I will check what's up once I could find an MS Windows machine. This
> will take some days; next week I'm not in office.

I've checked the history. Looks, like the removal of the skip was an
accident. In master, it is already reintroduced. So I've done in emacs-26.

Best regards, Michael.

Re: tramp-test42-auto-load failure

[Juanma Barranquero]

On Sat, Jun 15, 2019 at 9:40 PM Michael Albinus <michael.albinus@gmx.de> wrote:

> I've checked the history. Looks, like the removal of the skip was an
> accident. In master, it is already reintroduced. So I've done in emacs-26.

Great, thanks.