On Wed, Jul 8, 2015 at 6:06 AM, Philip Martin
<philip.martin_at_wandisco.com> wrote:
> Branko Čibej <brane_at_wandisco.com> writes:
>
>> On 07.07.2015 19:58, Zk W wrote:
>>> Hi All
>>>
>>> We have SVN 1.6.
>>> We like to perform a svn merge on a file where its filename has spaces.
>>> eg
>>> Roaming Apple.txt
>>>
>>> Is this syntax correct below to svn merge to a working copy ?
>>> If not, what should it be ?
>>>
>>> svn merge -c 12345 "https://mytext.abb.com/app/path/to/Roaming
>>> Apple.txt" .
>>>
>>> using quotes?
>>>
>>>
>>> Right now, we get error
>>> svn: Unable to parse URL '/svn/app/!svn/bc/12576/path/to/Roaming Apple.txt
>>> using
>>> svn merge -c 12345 https://mytext.abb.com/app/path/to/Roaming Apple.txt .
>>>
>>> Any help is appreciated.
>>
>>
>> You have to properly escape spaces (and other special characters) in an
>> URL. See, for example,
>>
>> https://en.wikipedia.org/wiki/Percent-encoding
>>
>> In your case, the URL would be
>>
>> https://mytext.abb.com/app/path/to/Roaming%20Apple.txt
>
> It might be a bit more complicated than that.
>
> The command line client has code to auto-escape command line URLs.
> svn_client_args_to_target_array() calls svn_opt__arg_canonicalize_url.
> The client should automatically convert spaces to %20.
>
> The error message shows an ra_neon (or ra_serf?) error while failing to
> parse a bc URL from the server. I think that means the client has
> correctly escaped the space and that the error is occurring later. I
> suppose manually escaping the space might be sufficient to make it work,
> but it is possible that it will not and that a client/server upgrade is
> required.
This is also why sane pre-commit blocks often block the use of
whitespace, punctuation, and non-flat-text-ASCII characters in any
file or directory names. It won't save you at this moment, but it can
help prevent repititions of the problem in future projects. Files with
names like "Nico's Big Project(1883)[see notes]Look How Cool!!!!.html"
tend to choke the phlegm out of any scripting language that has to
deal with them. And note that I didn't even *try* for Unicode....
Received on 2015-07-08 13:54:58 CEST