Beginner: problems with the win packages

[Alexander Belchenko]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Marcos Chaves пишет:
> On 11/9/06, Alexander Belchenko <bialix at ukr.net> wrote:
>> For standalone bzr.exe in folder where bzr installed there is file
>> start_bzr.bat. You can add command
>>
>> @chcp 1252
>>
>> to the start of this batch file. And then every time open console for
>> bzr session either via Start -> Programs -> Bazaar menu or via Context
>> menu in Explorer. I hope it works for you.
> 
> Hi Alexander,
> 
> thanks for the info. I was thinking about a workaround similar to
> this, just to get bzr to work for me.
> 
>> In long term I think we need explicitly check for codepage 'cp0' and
>> treat them as 'ascii' and/or fallback to bzrlib.user_encoding that seems
>> to have valid value. John, what you think about this patch:
>>
>> === modified file 'bzrlib/osutils.py'
>> - --- bzrlib/osutils.py   2006-11-02 07:29:02 +0000
>> +++ bzrlib/osutils.py   2006-11-09 07:53:01 +0000
>> @@ -372,6 +372,10 @@
>>              mutter('encoding stdout as sys.stdin encoding %r',
>> output_encoding)
>>      else:
>>          mutter('encoding stdout as sys.stdout encoding %r',
>> output_encoding)
>> +    if output_encoding == 'cp0':
>> +        output_encoding = bzrlib.user_encoding
>> +        mutter('cp0 is invalid encoding.'
>> +               ' encoding stdout as bzrlib.user_encoding %r',
>> output_encoding)
>>      return output_encoding
>>
>>
>> @@ -1085,7 +1089,7 @@
>>                           "  Continuing with ascii encoding.\n"
>>                           % (e, os.environ.get('LANG')))
>>
>> - -    if _cached_user_encoding is None:
>> +    if _cached_user_encoding in (None, 'cp0'):
>>          _cached_user_encoding = 'ascii'
>>      return _cached_user_encoding
>>
>>
>> Marcos, can you try this patch on your system? Is it works for you?
>>
> 
> I tried your patch, and it works for me. The second part of the patch
> (if _cached_user_encoding...) did not have any effect - the first part
> solved the problem.

Second part needed because we need to ensure that we always use valid
encoding.

> 
> Your patch is similar to something that I was trying, but I still
> didn't read all the posts about the problems that lead to the use this
> mechanism to identify the best encoding, so I don't know if this is
> appropriate. What I don't like about my approach (below) is that I
> don't think it would be nice to list all the exceptions that people
> find (if codepage == 'cp0'... ops, but then another user has 'cp1'...
> and the other one just talked about 'cp2'...). What do you think?
> 
> --- bzrlib-orig/osutils.py      2006-11-09 12:14:10.000000000 -0200
> +++ bzrlib/osutils.py   2006-11-09 12:14:04.000000000 -0200
> @@ -364,9 +364,9 @@
>     cp1252, but the console is cp437
>     """
>     output_encoding = getattr(sys.stdout, 'encoding', None)
> -    if not output_encoding:
> +    if not output_encoding or output_encoding == 'cp0':
>         input_encoding = getattr(sys.stdin, 'encoding', None)
> -        if not input_encoding:
> +        if not input_encoding or input_encoding == 'cp0':
>             output_encoding = bzrlib.user_encoding
>             mutter('encoding stdout as bzrlib.user_encoding %r',
> output_encoding)
>         else:

My first patch check for 'cp0' at the end of function
get_terminal_encoding() so we don't need to duplicate checks.

But if you talk about 'cp1' or so (that seems impossible) then we can be
*really* paranoid and do following checks:

=== modified file 'bzrlib/osutils.py'
- --- bzrlib/osutils.py   2006-11-02 07:29:02 +0000
+++ bzrlib/osutils.py   2006-11-10 17:12:33 +0000
@@ -25,6 +25,7 @@

 from bzrlib.lazy_import import lazy_import
 lazy_import(globals(), """
+import codecs
 import errno
 from ntpath import (abspath as _nt_abspath,
                     join as _nt_join,
@@ -372,6 +373,18 @@
             mutter('encoding stdout as sys.stdin encoding %r',
output_encoding)
     else:
         mutter('encoding stdout as sys.stdout encoding %r',
output_encoding)
+
+    # check encoding
+    try:
+        codecs.lookup(output_encoding)
+    except LookupError:
+        mutter('%s is invalid encoding.'
+               ' encoding stdout as'
+               ' bzrlib.user_encoding %s', (output_encoding,
+                                            bzrlib.user_encoding)
+              )
+        output_encoding = bzrlib.user_encoding
+
     return output_encoding


@@ -1087,6 +1100,13 @@

     if _cached_user_encoding is None:
         _cached_user_encoding = 'ascii'
+    else:
+        # check encoding
+        try:
+            codecs.lookup(_cached_user_encoding)
+        except LookupError:
+            _cached_user_encoding = 'ascii'
+
     return _cached_user_encoding



But I think it's superfluous.


Alexander
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFFVLWUzYr338mxwCURAgNjAKCCSK2xg1UL5YXPt4dFWXXBDEP5RgCggEvB
EJasJf1U2LBxea/kOMO/T6o=
=FNyr
-----END PGP SIGNATURE-----