On Jul 5, 2006, at 19:01, Garrett Rooney wrote:
> On 7/5/06, Michael Zheng (???) <zhengxinxing@gmail.com> wrote:
>> Hi all.
>>
>> I got "Section header expected" when I try to checkout my project
>> from the repository that contains a UTF-8 codec authz file.
>>
>> The authz file contains the BOM("ef bb bf") and if I got rid of
>> the BOM from my authz file but still kept it as UTF-8, I would
>> checkout my project successfully.
>>
>> I have the Chinese directory in my project and try to use Chinese
>> username to auth. So I have to keep the authz file as UTF-8 or svn
>> will fail to recognise the Chinese.
>>
>> Is this a bug?
>
> Well, it would certainly be nice if a BOM didn't throw the config
> parser for a loop like that, so yeah, I'd call it a bug, but there is
> a good work around, since a BOM on a utf-8 file is unnecessary anyway,
> since its endianness is irrelevant.
The file's endianness may not be in question, but its character
encoding still is. Suppose I have a text editor like BBEdit on Mac OS
X, and I have configured it to use ISO-8859-1 for files by default,
since that's the kind of file I usually deal with, but for certain
files, I need to use UTF-8. For those files, I'd like to include a
byte-order marker at the beginning, because BBEdit will recognize
that when opening the file again later and will realize that it's a
UTF-8 file. In the absence of the BOM, BBEdit can't know the file's
character encoding and will fall back to the configured default, and
I would have to manually tell the editor what character encoding to
use for this file -- every time I open it, which gets old fast.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Wed Jul 5 21:02:12 2006