OpenEuphoria

1. Bug in dir() leads to Unicode and win_dir()

• Posted by Greg Haberek <ghaberek at gmail.com> Jul 14, 2005
• 495 views

While working on my Backup Utility for work, I found a bug in
Euphoria's dir() routine. It does not support extended ASCII
characters and just returns a question mark '?' in place of unknown
characters.

I therefore have written a function called win_dir() which uses
Unicode strings to full support all available characters, as well as
up to 32,767 character path names, instead of just 255. It only works
with Windows obviously, since it calls FindFirstFileW, FindNextFileW,
and FindClose in kernel32.dll.

Seeing as how I forced myself to learn how Unicode strings work, I
also wrote a Unicode Conversion Library that can convert plain ASCII
strings (or sequences) to Unicode values, as well as two very useful
routines, allocate_unicode() and peek_unicode().

Both libraries, windir.zip and unicode.zip have just been uploaded to
the Archive. I hope others find these as useful as I do, especially
Derek, who IMHO should implement Unicode in Win32Lib so that it will
be more compatible with the Windows environment.

~Greg

3. Re: Bug in dir() leads to Unicode and win_dir()

• Posted by ags <eu at 531pi.co.nz> Jul 15, 2005
• 452 views

Greg Haberek wrote:
> 
> While working on my Backup Utility for work, I found a bug in
> Euphoria's dir() routine. It does not support extended ASCII
> characters and just returns a question mark '?' in place of unknown
> characters.

Hi. While this is great news and probably explain a crash I am having at
work, I have encountered a problem with this, ie:

junk = win_dir(folname)

the t_race shows after this line:
folname={70F,58:,92\,100d,97a,116t,97a,92\,109m,117u,115s,105i,99c}
junk={{{109m,117u,115s,105i,99c},{100d},0,105i,7,7,0,8,23}}

Which is not right, since there is no 'music' subfolder of "F:\data\music".

I'm running WinXP SP2 and the F: Drive is NTFS.

Also, while I'm here, when this didn't work I was attempting to add this 
to the include file:

global function walk_win_dir(sequence path_name, integer your_function,
                               integer scan_subdirs)
    object nothing

    my_dir = routine_id("win_dir") -- use win_dir() please
                                   -- my_dir is global atom in file.e
    nothing = walk_dir(path_name, your_function, scan_subdirs)
    my_dir = -2 	           -- constant DEFAULT=-2 in file.e
    return nothing
end function

I don't know if this can be incorporated?

Gary

5. Re: Bug in dir() leads to Unicode and win_dir()

• Posted by Greg Haberek <ghaberek at gmail.com> Jul 15, 2005
• 463 views

> But anyway it is just Windows feature, not dir() bug,
> I think. Correct me please, if I'm wrong.

All of Windows 2000/XP uses the Unicode (W) versions of functions. All
ANSI (A)versions are just backward compatible. When converting from
Unicode to ANSI, any unknown characters are replaced by a '?' as the
default character. I just learned this recently. So in reality, this
isn't even a bug, it's a Microsoft "feature" but either way, it has
caused me some issues. I tried creating file names with characters
from 129-255 and all I got back for some of them was a '?' character.

Thus the need to for win_dir() with Unicode support. Besides, who
doesn't want 32,767 character file names? :)

~Greg

7. Re: Bug in dir() leads to Unicode and win_dir()

• Posted by Greg Haberek <ghaberek at gmail.com> Jul 16, 2005
• 617 views
• Last edited Jul 17, 2005

> }}}
<eucode>
> junk = win_dir(folname)
>=20
> the t_race shows after this line:
> folname={70F,58:,92\,100d,97a,116t,97a,92\,109m,117u,115s,105i,99c}
> junk={{{109m,117u,115s,105i,99c},{100d},0,105i,7,7,0,8,23}}
> </eucode>
{{{

>
> Which is not right, since there is no 'music' subfolder of "F:\data\music=
".
>
> I'm running WinXP SP2 and the F: Drive is NTFS.

I forgot to mention this. I'll add it to the include file, here's why:

Let's say "F:\data\music" is a folder. You want all the files inside
that folder. You must specify win_dir( "F:\data\music\" ) with the
leading backslash. If you don't, win_dir() will return the information
*for that folder*.

This information may be found here:
http://msdn.microsoft.com/library/default.asp?url=/library/en-us/fileio/f=
s/findfirstfile.asp

Specifically:
<quote>
To examine a directory that is not a root directory, use the path to
that directory, without a trailing backslash. For example, an argument
of "C:\windows" returns information about the directory "C:\windows",
not about a directory or file in "C:\windows". An attempt to open a
search with a trailing backslash always fails.
</quote>

What Microsoft is saying, is that to obtain information about a
specific directory, specify "F:\data\music" to get information about
the files in a directory, specify "F:\data\music\*" and that
"F:\data\music\" is invalid.

Right now, win_dir() check for a trailing backslash and if it occurs,
adds a * to get all files. If you don't specify a trailing backslash,
then I have no way of knowing if you specified a specific file or a
folder.


> Also, while I'm here, when this didn't work I was attempting to add this
> to the include file:
>
> I don't know if this can be incorporated?

I just do this, it is much easier:

my_dir = routine_id("win_dir")
    junk = walk_dir( path, my_entry_routine, 1 )

9. Re: Bug in dir() leads to Unicode and win_dir()

• Posted by ags <eu at 531pi.co.nz> Jul 20, 2005
• 452 views

Greg Haberek wrote:
> 
> > Do you think it's worth trying to find the cause of this and fix it in
> > win_dir() or to implement a new walk_win_dir() ?
> 
> I think I may need to tweak walk_win_dir(). Also I'm trying to prepend
> win_ to Euphoria routines, so it will be called win_walk_dir()
> instead. I think in order to better support the dir()/win_dir()
> compatibility, I'm going to have it check to see if the path is a
> directory, and if so, look *in* that directory for files. This will
> allow me to make win_walk_dir() like this:

That's cool, I just didn't want to change win_dir() itself, and in fact
had trouble doing that, hence the hacked walk_win_dir().

No problems about the name though :)

Gary