Bytecompiler test fails

Bytecompiler test fails

[Fabrice Popineau]

Hi,

When I run `make check', I get a couple of failures (Win64)
but I worry a little bit about this one:

Testing ../../../emacs/test/automated/bytecomp-tests.el
Running 10 tests (2015-06-02 15:34:24+0200)
Warning: `/' called with 0 args, but requires 2 or more
Warning: / called with 0 arguments, but requires 1+
Warning: `/' called with 0 args, but requires 2 or more
Warning: / called with 0 arguments, but requires 1+
Warning: `/' called with 1 arg, but requires 2 or more
Warning: `/' called with 1 arg, but requires 2 or more
   passed   1/10  bytecomp-tests

In my--test11:
c:/Users/Fabrice/AppData/Local/Temp/test-
bytecomp17448gDb.el:2:83:Warning:
    macro `my--test11' defined too late

In my--test12:
c:/Users/Fabrice/AppData/Local/Temp/test-
bytecomp17448gDb.el:2:139:Warning:
    macro `my--test12' defined too late
Loading c:/Users/Fabrice/AppData/Local/Temp/test-bytecomp17448KSY.elc...
Test bytecomp-tests--warnings backtrace:
  (if (unwind-protect (setq value-79227 (apply fn-79225 args-79226)) (
  (let (form-description-79229) (if (unwind-protect (setq value-79227
  (let ((value-79227 (quote ert-form-evaluation-aborted-79228))) (let
  (let ((fn-79225 (function re-search-forward)) (args-79226 (list "my-
  (save-current-buffer (set-buffer (get-buffer-create "*Compile-Log*")
  (lambda nil (save-current-buffer (set-buffer (get-buffer-create "*Co
  ert--run-test-internal([cl-struct-ert--test-execution-info [cl-struc
  ert-run-test([cl-struct-ert-test bytecomp-tests--warnings nil (lambd
  ert-run-or-rerun-test([cl-struct-ert--stats t [[cl-struct-ert-test b
  ert-run-tests(t #[385 "\306\307\"\203G\211\211G\310U\203\211@\20
  ert-run-tests-batch(nil)
  ert-run-tests-batch-and-exit()
  command-line-1(("-L" ";../../../emacs/test/automated" "-l" "ert" "-l
  command-line()
  normal-top-level()
Test bytecomp-tests--warnings condition:
    (ert-test-failed
     ((should
       (re-search-forward "my--test11:\n.*macro" nil t))
      :form
      (re-search-forward "my--test11:\n.*macro" nil t)
      :value nil))
   FAILED   2/10  bytecomp-tests--warnings
Loading c:/Users/Fabrice/AppData/Local/Temp/test-bytecomp17448xwq.elc...
   passed   3/10  test-byte-comp-macro-expand-lexical-override
Loading c:/Users/Fabrice/AppData/Local/Temp/test-bytecomp174489OG.elc...
   passed   4/10  test-byte-comp-macro-expansion
Loading c:/Users/Fabrice/AppData/Local/Temp/test-bytecomp17448ktY.elc...
   passed   5/10  test-byte-comp-macro-expansion-eval-and-compile
Loading c:/Users/Fabrice/AppData/Local/Temp/test-bytecomp17448LMr.elc...
   passed   6/10  test-byte-comp-macro-expansion-eval-when-compile
   passed   7/10  test-eager-load-macro-expand-lexical-override
   passed   8/10  test-eager-load-macro-expansion
   passed   9/10  test-eager-load-macro-expansion-eval-and-compile
   passed  10/10  test-eager-load-macro-expansion-eval-when-compile

Is this to be expected? I pulled the git master this morning.
Thanks,

Fabrice

Re: Bytecompiler test fails

[Eli Zaretskii]

> From: Fabrice Popineau <fabrice.popineau@gmail.com>
> Date: Tue, 2 Jun 2015 13:49:20 +0000 (UTC)
> 
> In my--test11:
> c:/Users/Fabrice/AppData/Local/Temp/test-
> bytecomp17448gDb.el:2:83:Warning:
>     macro `my--test11' defined too late
> 
> In my--test12:
> c:/Users/Fabrice/AppData/Local/Temp/test-
> bytecomp17448gDb.el:2:139:Warning:
>     macro `my--test12' defined too late
> Loading c:/Users/Fabrice/AppData/Local/Temp/test-bytecomp17448KSY.elc...
> Test bytecomp-tests--warnings backtrace:
>   (if (unwind-protect (setq value-79227 (apply fn-79225 args-79226)) (
>   (let (form-description-79229) (if (unwind-protect (setq value-79227
>   (let ((value-79227 (quote ert-form-evaluation-aborted-79228))) (let
>   (let ((fn-79225 (function re-search-forward)) (args-79226 (list "my-
>   (save-current-buffer (set-buffer (get-buffer-create "*Compile-Log*")
>   (lambda nil (save-current-buffer (set-buffer (get-buffer-create "*Co
>   ert--run-test-internal([cl-struct-ert--test-execution-info [cl-struc
>   ert-run-test([cl-struct-ert-test bytecomp-tests--warnings nil (lambd
>   ert-run-or-rerun-test([cl-struct-ert--stats t [[cl-struct-ert-test b
>   ert-run-tests(t #[385 "\306\307\"\203G\211\211G\310U\203\211@\20
>   ert-run-tests-batch(nil)
>   ert-run-tests-batch-and-exit()
>   command-line-1(("-L" ";../../../emacs/test/automated" "-l" "ert" "-l
>   command-line()
>   normal-top-level()
> Test bytecomp-tests--warnings condition:
>     (ert-test-failed
>      ((should
>        (re-search-forward "my--test11:\n.*macro" nil t))
>       :form
>       (re-search-forward "my--test11:\n.*macro" nil t)
>       :value nil))
>    FAILED   2/10  bytecomp-tests--warnings

Please try again, I think I fixed this.  (The reason was the long name
of temporary-file-directory, something that's typical on MS-Windows,
which caused part of the warning be spilled to the next line, so ".*",
which doesn't match a newline, didn't work.)

Re: Bytecompiler test fails

[Fabrice Popineau]

[-- Attachment #1: Type: text/plain, Size: 1232 bytes --]

2015-06-21 19:30 GMT+02:00 Eli Zaretskii <eliz@gnu.org>:

> > From: Fabrice Popineau <fabrice.popineau@gmail.com>
> > Date: Tue, 2 Jun 2015 13:49:20 +0000 (UTC)
> >
> > In my--test11:
> > c:/Users/Fabrice/AppData/Local/Temp/test-
> > bytecomp17448gDb.el:2:83:Warning:
> >     macro `my--test11' defined too late
> >
> > In my--test12:
> > c:/Users/Fabrice/AppData/Local/Temp/test-
> > bytecomp17448gDb.el:2:139:Warning:
> >     macro `my--test12' defined too late
> > Loading c:/Users/Fabrice/AppData/Local/Temp/test-bytecomp17448KSY.elc...
>
...

> >    FAILED   2/10  bytecomp-tests--warnings
>
> Please try again, I think I fixed this.  (The reason was the long name
> of temporary-file-directory, something that's typical on MS-Windows,
> which caused part of the warning be spilled to the next line, so ".*",
> which doesn't match a newline, didn't work.)
>

You seem to be right: the test now succeeds.

lI still have 3 failures:

SUMMARY OF TEST RESULTS
-----------------------
Files examined: 75
Ran 1467 tests, 1381 results as expected, 8 unexpected, 78 skipped
3 files contained unexpected results:
  vc-tests.log
  package-test.log
  ert-x-tests.log

Do you manage to pass all tests with mingw32?

Best regards,

Fabrice

[-- Attachment #2: Type: text/html, Size: 2438 bytes --]

Re: Bytecompiler test fails

[Eli Zaretskii]

> From: Fabrice Popineau <fabrice.popineau@gmail.com>
> Date: Sun, 21 Jun 2015 23:20:50 +0200
> Cc: Emacs developers <emacs-devel@gnu.org>
> 
> lI still have 3 failures:
> 
> SUMMARY OF TEST RESULTS
> -----------------------
> Files examined: 75
> Ran 1467 tests, 1381 results as expected, 8 unexpected, 78 skipped
> 3 files contained unexpected results:
> vc-tests.log
> package-test.log
> ert-x-tests.log

What's the problem with vc-test?  Do you have RCS installed, and does
it crash?  If so, I'm working on that (the problem seems to be in RCS).

> Do you manage to pass all tests with mingw32?

Not all of them, but I think the ones that fail are different.

Re: Bytecompiler test fails

[Fabrice Popineau]

[-- Attachment #1: Type: text/plain, Size: 1034 bytes --]

2015-06-22 4:51 GMT+02:00 Eli Zaretskii <eliz@gnu.org>:

> > From: Fabrice Popineau <fabrice.popineau@gmail.com>
> > Date: Sun, 21 Jun 2015 23:20:50 +0200
> > Cc: Emacs developers <emacs-devel@gnu.org>
> >
> > lI still have 3 failures:
> >
> > SUMMARY OF TEST RESULTS
> > -----------------------
> > Files examined: 75
> > Ran 1467 tests, 1381 results as expected, 8 unexpected, 78 skipped
> > 3 files contained unexpected results:
> > vc-tests.log
> > package-test.log
> > ert-x-tests.log
>
> What's the problem with vc-test?  Do you have RCS installed, and does
> it crash?  If so, I'm working on that (the problem seems to be in RCS).
>
>
This command is expected to fail in test/automated/vs-tests.el:133:

 (shell-command-to-string (format "cvs -Q -d:local:%s co module" tmp-dir))

because tmp-dir is a windows path, aka c:/Users/.../temp .

I have an msys2/rcs installed:

$ pacman -Q rcs
rcs 5.9.4-1
$ pacman -Q cvs
cvs 1.11.23-2

But if I'm right, being an msys2 rcs/cvs, they don't support native win32
path syntax.

Fabrice

[-- Attachment #2: Type: text/html, Size: 1892 bytes --]

Re: Bytecompiler test fails

[Eli Zaretskii]

> From: Fabrice Popineau <fabrice.popineau@gmail.com>
> Date: Mon, 22 Jun 2015 07:21:54 +0200
> Cc: Emacs developers <emacs-devel@gnu.org>
> 
> This command is expected to fail in test/automated/vs-tests.el:133:
> 
> (shell-command-to-string (format "cvs -Q -d:local:%s co module" tmp-dir))
> 
> because tmp-dir is a windows path, aka c:/Users/.../temp .
> 
> I have an msys2/rcs installed:
> 
> $ pacman -Q rcs
> rcs 5.9.4-1
> $ pacman -Q cvs
> cvs 1.11.23-2
> 
> But if I'm right, being an msys2 rcs/cvs, they don't support native win32 path
> syntax.

It depends on how they were ported, but in general, yes, I'd expect
them not to support d:/foo/bar file names.

I don't know how to fix that, given that it could be a native Windows
application, an MSYS application, or a Cygwin application.  The only
idea I have is provide a primitive that would return the type of the
application, and then act accordingly.

Re: Bytecompiler test fails

[Eli Zaretskii]

> Date: Mon, 22 Jun 2015 18:05:07 +0300
> From: Eli Zaretskii <eliz@gnu.org>
> Cc: emacs-devel@gnu.org
> 
> > From: Fabrice Popineau <fabrice.popineau@gmail.com>
> > Date: Mon, 22 Jun 2015 07:21:54 +0200
> > Cc: Emacs developers <emacs-devel@gnu.org>
> > 
> > This command is expected to fail in test/automated/vs-tests.el:133:
> > 
> > (shell-command-to-string (format "cvs -Q -d:local:%s co module" tmp-dir))
> > 
> > because tmp-dir is a windows path, aka c:/Users/.../temp .
> > 
> > I have an msys2/rcs installed:
> > 
> > $ pacman -Q rcs
> > rcs 5.9.4-1
> > $ pacman -Q cvs
> > cvs 1.11.23-2
> > 
> > But if I'm right, being an msys2 rcs/cvs, they don't support native win32 path
> > syntax.
> 
> It depends on how they were ported, but in general, yes, I'd expect
> them not to support d:/foo/bar file names.
> 
> I don't know how to fix that, given that it could be a native Windows
> application, an MSYS application, or a Cygwin application.  The only
> idea I have is provide a primitive that would return the type of the
> application, and then act accordingly.

Done.  Please try running the test suite again.

There was an unrelated issue with vc-svn, which I also fixed, at least
for a native w32 svn executable.