version-info --include-history UnicodeDecodeError (518609)

[Robert Collins]

On Fri, 2010-04-09 at 09:51 +0900, Stephen J. Turnbull wrote:
> Robert Collins writes:
> 
>  > version info has a number of different formatters available; some may
>  > make sense only as files, some could encode to the local encoding. I
>  > don't think thats ideal; perhaps we could make the formatter 'know' that
>  > it is writing to stdout and error unless stdout is a pipe (and thus
>  > presumably being redirected). That seems fragile though.
> 
> Nope.  This is TOOWTDI.  Inside of Bazaar, everything should be
> explicitly Unicode (if "text", ie, intended to be human-readable) or
> str (if humans never read it without the intermediation of some
> application).  Stuff in str format probably should be encapsulated in
> some class, but that's beyond the scope of this post.  Unicode is
> *the* encapsulation for text.  End of story.

As already said in this thread, RIO isn't 'text', its a defined series
of bytes.

> This implies that text *must* be converted on exit from Bazaar. 

yes

>  All
> exits must be watched, just as all entrances must be watched. 

yes

>  stdout
> and stderr are nondeterministic, but the heuristic "convert to
> top-level locale encoding if a TTY, else to UTF-8" should work well in
> practice.  (Except maybe on Windows, where the widechar Unicode API
> might be more appropriate for non-TTYs.)

That is the heuristic I was talking about.

-Rob
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20100409/9f9c661b/attachment-0001.pgp