On 8/28/2012 14:08 Rony G. Flatscher said:
> On 28.08.2012 19:32, Mark Miesfeld wrote:
>> On Tue, Aug 28, 2012 at 10:19 AM, Chip Davis <[email protected]> wrote:
>>> On 8/28/2012 08:46 Rony G. Flatscher said:
>>>> Just did that and saw that you yourself had not done that vote up.
>>> I had not bumped the RFE because I wanted to initiate a discussion
>> Actually I think Rony meant Michael had not bumped up his own RFE, but
>> that's not that relevant.
>>
>>> about the implementation of it, first. Specifically, Michael
>>> suggested a flag to indicate that every file should be returned.
>>>
>>> My point was that "return every file" should be the default behavior,
>>> and that the interface could use flags, RE's, wildcards, arguments, or
>>> whatever, to indicate something less than every file was desired.
>>>
>>> I know the "backwards compatibility" argument will be raised, but the
>> But what about "backwards compatibility?"
>>
>> I agree with you that the reasonable default for the "S" flags would
>> have been all files. But it wasn't implemented that way on Unix-like
>> systems. It's been that way for a long time, probably forever. If we
>> suddenly changed the behavior, programs like the ones Michael has
>> already written and has been using would no longer produce the same
>> results.
>>
>> Is that okay to do just because the original default wasn't reasonable?
>>
>> That's the question I'm interested in.
> This is an interesting question!
>
> What, if one just regards the current behaviour on Unix systems to be a (long
> standing) bug, rather
> than a RFE?
> :)
As one who frequently raises the backward-compatibility issue, I fully
recognize the issue with changing SysFileTree. That's why I was
reticent to vote for this RFE. But it is broken now; fixing it will
probably require more effort than writing a new routine.
Which means the best solution is a new routine. Call it "SysFileList"
if you will. I'm with Rony in that all of the various implementations
of SysFileTree that mimicked the behavior of a particular platform's
interactive "list files" command ('ls', 'listfile', 'dir', et al.)
were doomed to cross-platform failure from the start.
It's time we implemented a truly cross-platform file list utility,
where the default is to return all files.
IMHO, of course.
-Chip-
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Oorexx-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/oorexx-devel
