> Entirely unneeded in TP/BP, so long as you explicitly
> Close the output file(s) you've opened.
jo> I think we've had this discussion before. When last we discussed it,
jo> I found that a write to a file may require a flush in some cases.
Yes, it's been discussed before. However, the end result of all
discussions was that no one could prove or provide a case where a Flush
prior to Close did anything useful or was ever _needed_. Thus, it's
superfluous code (generated and executed), which I don't feel is
appropriate to be posting in an echo where "instructions" are being
given.
Yes, there are situations where data may not be _committed_ to the
media following a Close, but that's an issue only of the operating
system (Windows, etc.), system caching, or something beyond the scope of
a particular application program. In the case of TP/BP programs, the
RTL _flushes_the_internal_buffers_ when the close is executed, which is
no more than Flush would be doing. Once that's complete, physical
recording on/to the media is a question of the o/s and environment, and
therefore there's nothing gained by performing the (extraneous) Flush
within the program.
... Constants aren't; variables don't.
--- OMX/Blue Wave/DOS v2.20
---------------
* Origin: Mike's Place (1:114/307.0)
|