Re: [cc65] Filenames for POSIX functins on CBM targets

Hi,

> Why shouldn't they? I expect if I set _curunit to 1, then the program
> written for the floppy should also work for the tape (if it does not do
> any fancy disk specific stuff).
>
> So, they should be cross-target, yes.

1. The program won't build on other taegets as there's no _curunit.

2. The other targets likely don't have tape drives to be accessed via open().

> If this is not the case, it will not be possible to open some files
> written some 20 years ago.

Something I personally could live well with.

> Given that even on one platform only (CBM), it is not possible to find
> "disallowed" characters to use a separators, I can only imagine how hard
> it is to integrate the other platforms, too.

Again - Im _NOT_ looking for a single path syntax for all targets! I'm
only looking for _SOME_ path syntax on every target. But the CBMs have
none at all. Therefore there's no need at all to find such a char.

> BTW: You are aware that even Unix/Windows have some differences? While
> open("2:hallo") will happily work on Unix, it will not work on Windows -
> that is, it will work on Windows with NTFS, but the result will be
> different from what most people would expect. With FAT, it will not
> work. With ReFS
> (https://blogs.msdn.com/b/b8/archive/2012/01/16/building-the-next-generation-file-system-for-windows-refs.aspx?Redirected=true),
> I am not sure, as it will not support alternate data streams.

I hope that it is a joke to ask me if I'm aware that there are path
syntax differences between OSes…

> You, on the other hand, seem to expect the cc65 RTL to be an OS in
> itself, which neglects the nifty details of each platform (which file
> names are allowed, which are not?). Am I right?

No, you are wrong. There's still this misunderstanding. I explicitky
want to stay clear from any type of emulation layer.

Classic MacOS and *IX and DOS and Windows and Apple II ProDOS and <you
name it> all have rheir own path syntax - and that's fine.

The only thing I want to say is that you need _some_ path syntax to
write meaningful POSIX programs. So if a target has none one needs to
be invented.

I.e. the same would be true for Apple DOS 3.3 ...

> Having said this: fopen() is a different beast to me. [...]


I'm pretty this is no viable option.

Regards,
Oliver
----------------------------------------------------------------------
To unsubscribe from the list send mail to majordomo@musoftware.de with
the string "unsubscribe cc65" in the body(!) of the mail.