Re: B_STAT_CHANGED

On Thu, 19 May 2005, Axel [iso-8859-15] Dörfler wrote:

> "Andrew Bachmann" <abachmann-CzTbWX96vYR8UrSeD/[email protected]> wrote:
>> "Axel Dörfler" <[email protected]> wrote:
>>> There are two solutions to this problem:
>>> 1) we could add another message field to B_STAT_CHANGED messages,
>>> that
>>> contain information about the intermediate state of the event, or
>>> 2) we introduce another event type like B_STAT_UPDATE to the list.
>> I vote for #2 since apps should already be dealing with numerous stat
>> changed
>> events from the dos filesystem, as previously noted.  I'm thinking
>> that such an
>> event could also be used for doing something when a file was finished
>> downloading or copying like that add-on monitor thing we have.
>
> Right, although I am not sure what you mean with the first part: I
> meant B_STAT_UPDATE as an event for the intermediate update,
> B_STAT_CHANGED as what it does under BeOS. The dos file system will
> have to be ported for Haiku anyway :-)
> But that reminds me that we should use a better name than
> B_STAT_UPDATE... maybe B_INTERMEDIATE_STAT_CHANGED, or
> B_INTERIM_STAT_CHANGED, or B_INTERMEDIATE_STAT_UPDATE, or whatever.

I actually find your other solution cleaner. Why should there be two 
message types for virtually the same thing? I understand the argument 
about the old applications, but are there really some out there, that 
rely on B_STAT_CHANGED being sent only when the file has been closed and 
that aren't open source? At any rate the BeBook doesn't document 
B_STAT_CHANGED to be sent only on file close, so I would consider those 
apps buggy and wouldn't add work-arounds to support them explicitely.

CU, Ingo