Re: POSIX Issue: To_Posix_String() generates incorrect data

[Mark Lorenzen <mark.lorenzen@ofir.dk>]

"Thierry S." <tpschneider1@shaw.ca> writes:

> Hi,
>
> I am running RH9 (2.4.20-8), with GNAT 3.15p and Florist (compatible version
> downloaded from ACT-Europe/Libre) and I am having all sorts of problems with
> the POSIX library.
>
> Would anybody have any idea of what is wrong?
> I must have setup it up incorrectly, but the INSTALL and README files are
> very sparse of information when it comes to solving issues.
>
> I am using a small little program and I implemented the following test:
> ...
> My_String : consatnt string := "/home/tschn/file";
> The_Posix_String : constant Posix.Posix_String := Posix.To_Posix_String
> (My_String);
> To_Posix_And_Back_String : constant string :=
> Posix.To_String(The_Posix_String);
>
> ... and To_Posix_And_Back_String  is not equal to My_String (it is all
> wrong)!!!!
>
> I then implemented the 2 routines myself, and it was very easy and working.
> So how come the Florist library gets it wrong?  For sure, I cannot use my
> version of the routines as the Posix instances are called all over my
> application, and I don't want to modify it.  On top of that, I am sure that
> if I wold do this replacement, my trouble would not be over, and the Posix
> interface would not work properly anyways.
>
> Again, the issue of function To_Posix_String() is only an illustration of my
> problems, the source of them must be somewhere else, like of the ./configure
> which may not deal with my version of Linux RH9 well.
>
> Thank you for your help.
> Thierry

I am unable to recreate the problem on Fedora Core 1 using florist
3.15p and two different Ada compilers - the one coming with Fedora
Core 1 (gcc 3.3.2 20031022) and gcc 3.4.0 20040121
(prerelease). Everything works as expected.

Please post a little test program (that can compile!).

- Mark Lorenzen