Linux installation

[Bob Warren]

I first met the term "standalone" when I initially made contact with 
Rev. Before this, using Windows only, I had always dealt with "executables".
So now, in order to clarify certain issues connected with the subject of 
installation, I would like to make a series of statements about these 2 
types of binary files, and I would be grateful if you would correct me 
if I am wrong.

A "standalone" means what it says. It stands alone and is completely 
independent. It carries what might have originally been library 
information within itself, and uses this at runtime. It does NOT refer 
to what might be similar library routines already embedded in the 
operating system it later runs on, that could be more or less current or 
"up to date" in comparison. Consequently, if I produce a "standalone" as 
a single file in Rev, everything it needs is bundled with it and it 
should run on anyone's machine by simply copying it (almost anywhere) to 
the HD and double-clicking on it.

My recent little experiment of producing a GUI module in RB has been 
very instructive. It did not run on Mark Wieder's Kubuntu and complained 
about the unavailability of a certain library routine. To me, this seems 
to indicate that RB does not produce a "standalone" GUI program but an 
"executable" one in the traditional Windows manner. At runtime, on a 
different machine, if the libraries found in the operating system are 
not 100% compatible with the libraries present in the system where the 
program was created, it fails. Thus, such RB programs, different to Rev 
standalones, might require a "setup" in the traditional Windows manner, 
even under Linux.

If what I say is correct - AND PLEASE JUMP ON ME IMMEDIATELY IF I AM 
INCORRECT - then the question of "setting up" Rev programs in Linux is 
**normally** a relatively simple one. There is no damn "registry", and 
all you have to do is to copy (perhaps with decompression) the program 
and other files where you like on the HD, and to provide (if you can) 
icons on the desktop and/or in the start menu.

Here are a few more statements that I hope you will put me right on 
where necessary.

There are 8 HD paths that my little module hopes to deduce from the 
operating system. On my Ubuntu Linux, here is an example:

/home/bob/Desktop/            (DesktopFolder)
/home/bob/                (PreferencesFolder)
/usr/                    (SystemFolder)
/tmp/                    (TemporaryFolder)
/home/bob/                (ApplicationSupportFolder)
failed                    (FontsFolder)
failed                    (StartupItemsFolder)
/home/bob/.Trash/            (TrashFolder)

And here is another example on Gentoo (with KDE):

/home/rishi/Desktop/            (DesktopFolder)
/home/rishi/                (PreferencesFolder)
/usr/                    (SystemFolder)
/tmp/                    (TemporaryFolder)
/home/rishi/                (ApplicationSupportFolder)
failed                    (FontsFolder)
failed                    (StartupItemsFolder)
failed             (TrashFolder)

Of these, the first 5 are always the same under all distributions of 
Linux, except for the user name. Therefore, as Jacqueline Landman Gay 
suggested, the use of the global $HOME variable in Rev should be 
sufficient to deduce these paths.

Different distributions of Linux typically vary the paths to the last 3 
categories, and the RB functions for deducing them may or may not be 
successful, or in other words, they are unreliable for general Linux use 
and can only be employed perhaps for a limited set of distros where they 
are known to succeed.

And finally, considering the above, although it would be nice if Rev 
implemented the suite of reportable system path-deducing functions 
through specialFolderPath in Linux, or even extended them to provide 
reliable reports on the Fonts, Startup Items and Trash, there is no 
greatly imperative need for it at this very moment, unless you want to 
install new fonts, get your program executed automatically at OS 
startup, etc.

I AWAIT YOUR CORRECTIONS. Thanks.

Bob