Re: GWorkspace issues on MinGW

On 2006-08-19 00:24:33 +0300 Nicolas Roard <[email protected]> wrote:

> On 8/18/06, Enrico Sersale <[email protected]> wrote:
>> On 2006-08-18 14:01:02 +0300 Nicolas Roard <[email protected]> wrote:
>>> I'm not sure GWorkspace is actually even supposed to work on
>>> Windows... it probably expects lots of linuxism.
>> 
>> GWorkspace has *NO* linuxisms; besides linux, it runs on darwin, *BSD, 
>> solaris, etc...
> 
> I just meant linuxism as in != Windows ...
> I should have s/linuxism/unixism :-)

Well, it has become about a synonym...

>> Here I see a problem in the "C:\DesktopNewFolder" path; there is not the 
>> path separator that should be between "C:\Desktop" and "NewFolder".
>> Usually I use -stringByAppendingPathComponent: and, in the few cases when I 
>> must append a path separator by hand, I use a function that calls 
>> NSFileManager's -stringWithFileSystemRepresentation:length:; this should 
>> give correct paths also on windows.
>> Anyway, not having windows, I can't test anything and it's very probable 
>> that gworkspace doesn't run on it, but not due of these basical problems 
>> (path handling, etc...).
> 
> It's perhaps also a problem with NSFileManager itself when used on
> Windows, not specifically GWorkspace.

No. You are right; the issues are really in GWorkspace; all the app is based on the assumption that there is only one 'root', that is, "/" on unix. I was not aware that on windows you have a path tree for each mountable "volume"; this makes the things quite complex because, to represent this in a browser, you must add the concept of a "root of the other roots" and write methods to handle this abstraction...