Re: Please vote on --fallback-encoding behavior for filenames

Michael Haggerty wrote:
> Option 3: On the other hand, filenames *are* certainly
more critical
> than authors and log names, so maybe we need a separate
option for
> filenames (e.g., --fallback-filename-encoding).  A
bonus of this
> approach is that it would more gracefully handle the
case of a
> repository that uses a different encoding for filenames
than for
> author/log messages.  (Though if this is a common
problem, then strictly
> speaking we need a --filename-encoding option too.)

I vote to the option 3, in favor of having a separated
--filename-encoding option.

A widely used windows client for CVS in Japan causes
difference of
encodings between filenames (CP932, windows default) and log
messages
(EUC-JP, Unix server encoding). A separate
--filename-encoding will
make sense to some Japanese users. But I think the fallback
option for
filenames won't make sense.

Should it be named as "--path-encoding"? Because
it is used for
directory names too.
-- 
k_satoda

------------------------------------------------------------
---------
To unsubscribe, e-mail: dev-unsubscribecvs2svn.tigris.org
For additional commands, e-mail: dev-helpcvs2svn.tigris.org

Kazutoshi Satoda wrote:
> Michael Haggerty wrote:
>> Option 3: On the other hand, filenames *are*
certainly more critical
>> than authors and log names, so maybe we need a
separate option for
>> filenames (e.g., --fallback-filename-encoding).  A
bonus of this
>> approach is that it would more gracefully handle
the case of a
>> repository that uses a different encoding for
filenames than for
>> author/log messages.  (Though if this is a common
problem, then strictly
>> speaking we need a --filename-encoding option
too.)
> 
> I vote to the option 3, in favor of having a separated
> --filename-encoding option.
> 
> A widely used windows client for CVS in Japan causes
difference of
> encodings between filenames (CP932, windows default)
and log messages
> (EUC-JP, Unix server encoding). A separate
--filename-encoding will
> make sense to some Japanese users. But I think the
fallback option for
> filenames won't make sense.
> 
> Should it be named as "--path-encoding"?
Because it is used for
> directory names too.

To avoid command-line confusion, I have implemented more
flexible
encoding options, but only if cvs2svn is started using the
--options
file method.  If you use an --options file, you can specify
independently:

- a list of encodings for log messages and author names

- an optional fallback encoding for log messages and author
names

- a list of encodings for filenames

- an optional fallback encoding for filenames.

Michael

------------------------------------------------------------
---------
To unsubscribe, e-mail: dev-unsubscribecvs2svn.tigris.org
For additional commands, e-mail: dev-helpcvs2svn.tigris.org