Good detective work. There is an 80 character temporary variable (re-used from some other original purpose) which is used to format the combination of the
flags[/i[ and [i]file (i.e. parameter #7, which I believe you are referring to as
fname) parameters of
MX_GETOFD .
I guess I figured that 70 characters or so was more than enough to hold a typical file.ext (since after all, in the AMOS world, we were dealing with a 6.3 limitation). But in the "real world", I suppose we should allow for considerably more, although we start dealing with absurdly large filenames, it's hard to know where to stop. If, say, a 90 character long name is reasonable, what about a 900 character long name?
I should mention that A-Shell is mostly using 260 as the maximum length of a fully-qualified path (directory + filename). (This is more or less the Windows API standard, although technically some filesystems, like NTFS, can handle filenames of nearly unlimited length.) Usually if a path name is in danger of exceeding 260 characters, it would be due to a deeply nested directory rather than a long filename. But perhaps I'm stuck in a 20th century mindset.
I'll be happy to increase the size of this particular buffer from 80 to ... 260(?) in 1222.3, although if you want to comment first, please go ahead.