Re: master 97608c6: Extend Tramp's UTF8 tests

Re: master 97608c6: Extend Tramp's UTF8 tests

[Glenn Morris]

Michael Albinus wrote:

> branch: master
> commit 97608c60b8e0d01d59d51e32a9fdbd2bb3e2da58

>     Extend Tramp's UTF8 tests

Hi - just in case you didn't spot it, these tests now fail on hydra.

https://hydra.nixos.org/build/71555141

Re: master 97608c6: Extend Tramp's UTF8 tests

[Michael Albinus]

Glenn Morris <rgm@gnu.org> writes:

>> branch: master
>> commit 97608c60b8e0d01d59d51e32a9fdbd2bb3e2da58
>
>>     Extend Tramp's UTF8 tests
>
> Hi - just in case you didn't spot it, these tests now fail on hydra.
>
> https://hydra.nixos.org/build/71555141

Ohh, I didn't see. Will check.

Thanks, and best regards, Michael.

Re: master 97608c6: Extend Tramp's UTF8 tests

[Glenn Morris]

PS same failures for me on RHEL 7.4 now that I look.

Re: master 97608c6: Extend Tramp's UTF8 tests

[Michael Albinus]

Glenn Morris <rgm@gnu.org> writes:

Hi Glenn,

> PS same failures for me on RHEL 7.4 now that I look.

For the time being, I have marked the *39* Tramp tests as
:unstable. Since they work for me, I might later ask you to run some
tests for me.

Sorry for the inconvenience.

Best regards, Michael.

Re: master 97608c6: Extend Tramp's UTF8 tests

[Michael Albinus]

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

Hi Glenn,
>
>> PS same failures for me on RHEL 7.4 now that I look.

> For the time being, I have marked the *39* Tramp tests as
> :unstable. Since they work for me, I might later ask you to run some
> tests for me.

I have instrumented tramp-test39-utf8, pushed to master. Could you, pls run

$ make -C test lisp/net/tramp-tests.log SELECTOR='\"tramp-test39-utf8$$\"'

and send me the logfile? It could be rather long, sorry.

Thanks, and best regards, Michael.

Re: master 97608c6: Extend Tramp's UTF8 tests

[Glenn Morris]

Michael Albinus wrote:

> $ make -C test lisp/net/tramp-tests.log SELECTOR='\"tramp-test39-utf8$$\"'
>
> and send me the logfile? It could be rather long, sorry.

Sent off-list.

Re: master 97608c6: Extend Tramp's UTF8 tests

[Michael Albinus]

Glenn Morris <rgm@gnu.org> writes:

Hi Glenn,

>> $ make -C test lisp/net/tramp-tests.log SELECTOR='\"tramp-test39-utf8$$\"'
>>
>> and send me the logfile? It could be rather long, sorry.
>
> Sent off-list.

Could you pls tell me which stat version is running on your host? There
was a bug in coreutils prior version 8.26, see <https://debbugs.gnu.org/23422>.

Best regards, Michael.

Re: master 97608c6: Extend Tramp's UTF8 tests

[Glenn Morris]

Michael Albinus wrote:

> Could you pls tell me which stat version is running on your host? There
> was a bug in coreutils prior version 8.26, see <https://debbugs.gnu.org/23422>.

stat (GNU coreutils) 8.22
stat -c %N 'foo       bar'
  ->  ‘foo\tbar’

Re: master 97608c6: Extend Tramp's UTF8 tests

[Michael Albinus]

Glenn Morris <rgm@gnu.org> writes:

Hi Glenn,

> stat (GNU coreutils) 8.22
> stat -c %N 'foo       bar'
>   ->  ‘foo\tbar’

Thanks. Yes, stat 8.22 works fine with ASCII characters. But more
sophisticated character sets, like the extended tramp-test39-utf8 case
uses, seem to work proper only with stat 8.26 and later, which supports
the QUOTING_STYLE environment variable.

I will install stat 8.22 locally and see, whether the checks in Tramp
could be more precise to detect this situation.

Best regards, Michael.

Re: master 97608c6: Extend Tramp's UTF8 tests

[Michael Albinus]

Glenn Morris <rgm@gnu.org> writes:

Hi Glenn,

>> Could you pls tell me which stat version is running on your host? There
>> was a bug in coreutils prior version 8.26, see <https://debbugs.gnu.org/23422>.
>
> stat (GNU coreutils) 8.22
> stat -c %N 'foo       bar'
>   ->  ‘foo\tbar’

Status update: On my local Fedora machine, I've replaced stat 8.27 by
8.22, temporarily. The test still runs for me with all language
snippets.

stat is not suspicious any longer. I must dig somewhere else.

Best regards, Michael.

Re: master 97608c6: Extend Tramp's UTF8 tests

[Michael Albinus]

Hi Glenn,

I've found an old RHEL5 machine @work, which is able to show the
problem. Looks like the "TaiViet" string makes trouble, perhaps because
it is a composed string. I've taken this out.

Could you, pls, check again?

Thanks, and best regards, Michael.

Re: master 97608c6: Extend Tramp's UTF8 tests

[Glenn Morris]

Michael Albinus wrote:

> I've found an old RHEL5 machine @work, which is able to show the
> problem. Looks like the "TaiViet" string makes trouble, perhaps because
> it is a composed string. I've taken this out.

Yes, it passes now on RHEL7, thanks.