Re: Handling unicode file names on Windows

[Paul K <paulclinger@...>]

> One option on Windows is to use winapi:
> http://stevedonovan.github.com/winapi/api.html#short_path

Thanks to Steve and all others who responded. I did use some of the
suggestions and got most of the code working, but there are still
couple of issues to resolve.

It turned out that I can use wxwidgets methods for reading/writing
file content and these are already Unicode aware (including windows),
so I'm using wxFile methods for that.

Where I still have a problem is dofile. Obviously, dofile
[[some/Unicode/file/name]] is not going to work for the same reason
io.open doesn't work. What I want to do is to use the short name in
this case; this is where I have an issue.

This is what I tried:

> winapi.short_path("d:\\Lua\\下载\\foo.lua")
nil	"The system cannot find the path specified."

> winapi.short_path("d:\\Lua\\下载\\")
nil	"The filename, directory name, or volume label syntax is incorrect."

> winapi.short_path('d:/Lua/下载/foo.lua')
nil	"The system cannot find the path specified."

> winapi.short_path('D:/Lua/下载')
"D:/Lua/FD31~1"

For some reason it doesn't see the path, while it's definitely there.
Also, it creates a *file* in that folder, while I need it to find a
folder with the unicode name.

The correct name is this "dofile[[d:/Lua/D7A9~1/foo.lua]]" (and the
one that doesn't work is this: "dofile[[d:/Lua/下载/foo.lua]]"). Notice
that the shortname returned by short_path is different as it is for a
*file* with the same name, rather than a *folder*.

Also, it would be nice if short_path would *not* create a file by
default. Maybe a second parameter (true/FALSE) that would indicate if
the name should be returned for a file that doesn't exist. I'm not
sure why winapi.short_path('D:/Lua/下载') doesn't return the right value
when this unicode name is a folder and winapi.short_path('D:/Lua/下载/')
doesn't work at all (returns an error). It appears the current logic
is not going to work for directories (per msdn docs
http://msdn.microsoft.com/en-us/library/windows/desktop/aa363858%28v=vs.85%29.aspx):

- An application cannot create a directory by using CreateFile,
therefore only the OPEN_EXISTING value is valid for
dwCreationDisposition for this use case. To create a directory, the
application must call CreateDirectory or CreateDirectoryEx.
- To open a directory using CreateFile, specify the
FILE_FLAG_BACKUP_SEMANTICS flag as part of dwFlagsAndAttributes.
Appropriate security checks still apply when this flag is used without
SE_BACKUP_NAME and SE_RESTORE_NAME privileges.
- When using CreateFile to open a directory during defragmentation of
a FAT or FAT32 file system volume, do not specify the MAXIMUM_ALLOWED
access right. Access to the directory is denied if this is done.
Specify the GENERIC_READ access right instead.

Thank you.

Paul.

On Wed, Aug 15, 2012 at 11:45 PM, steve donovan
<steve.j.donovan@gmail.com> wrote:
> On Thu, Aug 16, 2012 at 8:35 AM, Paul K <paulclinger@yahoo.com> wrote:
>> information). What options do I have for working with unicode file
>> names? I want to be able to read/write files with unicode paths. This
>> is using Lua 5.1.4.
>
> One option on Windows is to use winapi:
>
> http://stevedonovan.github.com/winapi/api.html#short_path
>
> winapi.short_path takes an encoded path (UTF-8 or UTF-16) and returns
> the equivalent short path, which can be opened with io.open.
>
> Sometimes a legacy option works nicely ;)
>
> steve d.
>