[ruby-core:45887] [ruby-trunk - Bug #6650][Open] Fixing win32ole test errors

[ruby-core:45887] [ruby-trunk - Bug #6650][Open] Fixing win32ole test errors

[bosko (Boško Ivanišević)]

Issue #6650 has been reported by bosko (Boško Ivanišević).

----------------------------------------
Bug #6650: Fixing win32ole test errors
https://bugs.ruby-lang.org/issues/6650

Author: bosko (Boško Ivanišević)
Status: Open
Priority: Normal
Assignee: luislavena (Luis Lavena)
Category: test
Target version: 
ruby -v: ruby 2.0.0dev (2012-06-25 trunk 36180) [i386-mingw32]


Since FileSystemObject is limited only to ANSI and UTF-16LE encoding tests shouldn't try to use other encodings since they will certainly fail. Instead UTF-16 encoding should be used to check whether writing/reading to/from FSO is properly handled. Attached is a patch that fixes this.


-- 
http://bugs.ruby-lang.org/

[ruby-core:45895] [ruby-trunk - Bug #6650][Assigned] Fixing win32ole test errors

[luislavena (Luis Lavena)]

Issue #6650 has been updated by luislavena (Luis Lavena).

Status changed from Open to Assigned
Target version set to 2.0.0


----------------------------------------
Bug #6650: Fixing win32ole test errors
https://bugs.ruby-lang.org/issues/6650#change-27495

Author: bosko (Boško Ivanišević)
Status: Assigned
Priority: Normal
Assignee: luislavena (Luis Lavena)
Category: test
Target version: 2.0.0
ruby -v: ruby 2.0.0dev (2012-06-25 trunk 36180) [i386-mingw32]


Since FileSystemObject is limited only to ANSI and UTF-16LE encoding tests shouldn't try to use other encodings since they will certainly fail. Instead UTF-16 encoding should be used to check whether writing/reading to/from FSO is properly handled. Attached is a patch that fixes this.


-- 
http://bugs.ruby-lang.org/

[ruby-core:45945] [ruby-trunk - Bug #6650] Fixing win32ole test errors

[suke (Masaki Suketa)]

Issue #6650 has been updated by suke (Masaki Suketa).

% Done changed from 100 to 0

I received the following error.

test_s_codepage_changed(TestWin32OLE) [C:/cygwin/home/svn/ruby/test/win32ole/test_win32ole.rb:363]:
<"\xE3\x81\x82"> expected but was
<"\xE3\x81\x81E">.

Finished tests in 53.506664s, 4.1303 tests/s, 12.0359 assertions/s.
221 tests, 644 assertions, 1 failures, 0 errors, 0 skips

ruby -v: ruby 2.0.0dev (2012-06-28 trunk 36246) [i386-mswin32_100]
----------------------------------------
Bug #6650: Fixing win32ole test errors
https://bugs.ruby-lang.org/issues/6650#change-27553

Author: bosko (Boško Ivanišević)
Status: Closed
Priority: Normal
Assignee: luislavena (Luis Lavena)
Category: test
Target version: 2.0.0
ruby -v: ruby 2.0.0dev (2012-06-25 trunk 36180) [i386-mingw32]


Since FileSystemObject is limited only to ANSI and UTF-16LE encoding tests shouldn't try to use other encodings since they will certainly fail. Instead UTF-16 encoding should be used to check whether writing/reading to/from FSO is properly handled. Attached is a patch that fixes this.


-- 
http://bugs.ruby-lang.org/

[ruby-core:45947] [ruby-trunk - Bug #6650][Feedback] Fixing win32ole test errors

[luislavena (Luis Lavena)]

Issue #6650 has been updated by luislavena (Luis Lavena).

Status changed from Closed to Feedback

suke,

Can you tell us a bit more about WIN32OLE.locale and WIN32OLE.codepage prior running the tests? what is the default locale (chcp)?

These test fail with WIN32OLE.codepage 1252 and 65001, so I'm wondering why they are not in your case.
----------------------------------------
Bug #6650: Fixing win32ole test errors
https://bugs.ruby-lang.org/issues/6650#change-27555

Author: bosko (Boško Ivanišević)
Status: Feedback
Priority: Normal
Assignee: luislavena (Luis Lavena)
Category: test
Target version: 2.0.0
ruby -v: ruby 2.0.0dev (2012-06-25 trunk 36180) [i386-mingw32]


Since FileSystemObject is limited only to ANSI and UTF-16LE encoding tests shouldn't try to use other encodings since they will certainly fail. Instead UTF-16 encoding should be used to check whether writing/reading to/from FSO is properly handled. Attached is a patch that fixes this.


-- 
http://bugs.ruby-lang.org/

[ruby-core:45952] [ruby-trunk - Bug #6650] Fixing win32ole test errors

[h.shirosaki (Hiroshi Shirosaki)]

Issue #6650 has been updated by h.shirosaki (Hiroshi Shirosaki).

File 0001-Use-UTF-16-String-instead-of-UTF-8.patch added

I had same failure on Win7 with chcp 932(default), 1252 and 65001. I guess it depends on Windows system default encoding.
If I change the string from UTF-8 to UTF-16LE, the test works fine for me.
UTF-16LE or ASCII String might be better. I attached the patch.
----------------------------------------
Bug #6650: Fixing win32ole test errors
https://bugs.ruby-lang.org/issues/6650#change-27560

Author: bosko (Boško Ivanišević)
Status: Feedback
Priority: Normal
Assignee: luislavena (Luis Lavena)
Category: test
Target version: 2.0.0
ruby -v: ruby 2.0.0dev (2012-06-25 trunk 36180) [i386-mingw32]


Since FileSystemObject is limited only to ANSI and UTF-16LE encoding tests shouldn't try to use other encodings since they will certainly fail. Instead UTF-16 encoding should be used to check whether writing/reading to/from FSO is properly handled. Attached is a patch that fixes this.


-- 
http://bugs.ruby-lang.org/

[ruby-core:45977] [ruby-trunk - Bug #6650] Fixing win32ole test errors

[luislavena (Luis Lavena)]

Issue #6650 has been updated by luislavena (Luis Lavena).


h.shirosaki (Hiroshi Shirosaki) wrote:
> I had same failure on Win7 with chcp 932(default), 1252 and 65001. I guess it depends on Windows system default encoding.
> If I change the string from UTF-8 to UTF-16LE, the test works fine for me.
> UTF-16LE or ASCII String might be better. I attached the patch.

Hiroshi, I don't fully follow the rationale in the patch.

You're encoding the string first as UTF-16LE and then force encoding as UTF-16 when sending it to FSO.

AFAIK FSO Only understands ANSI and UTF-16LE.

Also, by doing force_encoding, you're changing test_str to the new encoding and not preserving the original.

I couldn't find WIN32OLE in the list of maintainers, but will be great to hear his comments.
----------------------------------------
Bug #6650: Fixing win32ole test errors
https://bugs.ruby-lang.org/issues/6650#change-27579

Author: bosko (Boško Ivanišević)
Status: Feedback
Priority: Normal
Assignee: luislavena (Luis Lavena)
Category: test
Target version: 2.0.0
ruby -v: ruby 2.0.0dev (2012-06-25 trunk 36180) [i386-mingw32]


Since FileSystemObject is limited only to ANSI and UTF-16LE encoding tests shouldn't try to use other encodings since they will certainly fail. Instead UTF-16 encoding should be used to check whether writing/reading to/from FSO is properly handled. Attached is a patch that fixes this.


-- 
http://bugs.ruby-lang.org/

[ruby-core:45979] [ruby-trunk - Bug #6650] Fixing win32ole test errors

[h.shirosaki (Hiroshi Shirosaki)]

Issue #6650 has been updated by h.shirosaki (Hiroshi Shirosaki).


> You're encoding the string first as UTF-16LE and then force encoding as UTF-16 when sending it to FSO.
> 
> AFAIK FSO Only understands ANSI and UTF-16LE.
> 
> Also, by doing force_encoding, you're changing test_str to the new encoding and not preserving the original.

I did force_encoding("UTF-16") because FSO doesn't allow UTF-16LE encoding and an error occurs. Bosko's patch also uses UTF-16 instead of UTF-16LE. force_encoding doesn't change binary representation of test_str.

----------------------------------------
Bug #6650: Fixing win32ole test errors
https://bugs.ruby-lang.org/issues/6650#change-27581

Author: bosko (Boško Ivanišević)
Status: Feedback
Priority: Normal
Assignee: luislavena (Luis Lavena)
Category: test
Target version: 2.0.0
ruby -v: ruby 2.0.0dev (2012-06-25 trunk 36180) [i386-mingw32]


Since FileSystemObject is limited only to ANSI and UTF-16LE encoding tests shouldn't try to use other encodings since they will certainly fail. Instead UTF-16 encoding should be used to check whether writing/reading to/from FSO is properly handled. Attached is a patch that fixes this.


-- 
http://bugs.ruby-lang.org/

[ruby-core:45987] [ruby-trunk - Bug #6650] Fixing win32ole test errors

[luislavena (Luis Lavena)]

Issue #6650 has been updated by luislavena (Luis Lavena).

Assignee changed from luislavena (Luis Lavena) to suke (Masaki Suketa)

Masaki,

Can you confirm tests pass after applying Hiroshi's patch?

Thank you.
----------------------------------------
Bug #6650: Fixing win32ole test errors
https://bugs.ruby-lang.org/issues/6650#change-27591

Author: bosko (Boško Ivanišević)
Status: Feedback
Priority: Normal
Assignee: suke (Masaki Suketa)
Category: test
Target version: 2.0.0
ruby -v: ruby 2.0.0dev (2012-06-25 trunk 36180) [i386-mingw32]


Since FileSystemObject is limited only to ANSI and UTF-16LE encoding tests shouldn't try to use other encodings since they will certainly fail. Instead UTF-16 encoding should be used to check whether writing/reading to/from FSO is properly handled. Attached is a patch that fixes this.


-- 
http://bugs.ruby-lang.org/

[ruby-core:46146] [ruby-trunk - Bug #6650] Fixing win32ole test errors

[luislavena (Luis Lavena)]

Issue #6650 has been updated by luislavena (Luis Lavena).


Masaki Suketa, 

Can you confirm patch from Hiroshi fix the error for you?

----------------------------------------
Bug #6650: Fixing win32ole test errors
https://bugs.ruby-lang.org/issues/6650#change-27763

Author: bosko (Boško Ivanišević)
Status: Feedback
Priority: Normal
Assignee: suke (Masaki Suketa)
Category: test
Target version: 2.0.0
ruby -v: ruby 2.0.0dev (2012-06-25 trunk 36180) [i386-mingw32]


Since FileSystemObject is limited only to ANSI and UTF-16LE encoding tests shouldn't try to use other encodings since they will certainly fail. Instead UTF-16 encoding should be used to check whether writing/reading to/from FSO is properly handled. Attached is a patch that fixes this.


-- 
http://bugs.ruby-lang.org/

[ruby-core:46175] Re: [ruby-trunk - Bug #6650] Fixing win32ole test errors

[Masaki Suketa]

On Sat, Jun 30, 2012 at 11:24:03PM +0900, luislavena (Luis Lavena) wrote:
> 
> Can you confirm tests pass after applying Hiroshi's patch?
>

Sorry for being too late to reply.
I confirmed Hiroshi's patch works fine.

  Best Regards,
  Masaki Suketa