[Coco] RS-DOS: EOF() but for writing?

[William Astle]

Writing past the end of the file will just make the file bigger unless 
you end up with a "Disk Full" error. Same thing for a PUT past the end 
of the file - it just extends the file if it can. So you have to get clever.

There isn't any handy "tell()" function that gives the current file 
position for sequential files. There also isn't a SEEK command either. 
So that makes things less easy than it would otherwise be.

You can use FREE() to identify if the next write will definitely 
succeed. If FREE() returns a number greater than 0, there's still a free 
granule that can be allocated to extend the file. Presumably this will 
be the case the vast majority of the time and it should guarantee that 
the next write will succeed for either sequential or random files.

If FREE() returns 0, then you have to be more clever.

For sequential files, that means you have to keep count of every byte 
you write to the file. However, since "O" mode files are truncated, you 
don't have to figure out how big the file was when you opened it. But 
you do need to convert every write to a string so you can identify how 
long it is, and include any terminator/delimiters that will be written 
by Disk Basic as well. You have to do this for *every* write whether 
regardless what FREE() returns.

If FREE() returns 0, then you would need to calculate if you next string 
will overflow the current granule. If your current running count is 
zero, you have space since you would have got a DF error opening the 
file otherwise. If not and your current count is an even granule size, 
the disk is full. If not, you calculate the next higher even granule 
size from your running count and calculate the difference. If the thing 
you're about to write is that size or smaller (including any 
terminators, etc.), it fits. Otherwise, the disk is full.

Random files are a bit different. With LOF() and FREE() you can pretty 
much guarantee you won't trigger a DF error with some careful 
calculations involving the record size. Obviously, writing to a record 
<= LOF() will work in this case since it's already allocated and Disk 
Basic files can't have "holes".

On 2023-01-26 13:15, Allen Huffman via Coco wrote:
> I know you can use EOF() to know if an input file is at the end, but is there a way to do the opposite, and know if a write would fail? I suppose this is not easy, since it doesn’t know how much you plan to write next.
> 
> Without an ON ERROR in CoCo 1/2 BASIC, you would just write until the program crashes with a ?DF ERROR (disk full).
> 
> Thoughts?
> 
> --
> Allen Huffman - PO Box 7634 - Urbandale IA 50323 - 515-999-0227 (vmail/TXT only)
> http://www.subethasoftware.com - https://www.facebook.com/subethasoftware
> 
> 
>