Michael Meeks wrote:
On Tue, 2007-02-27 at 16:49 +0100, Stephan Bergmann wrote:
can we sort this out somehow? I basically see two options, either
- do not report any out-of-memory conditions and (silently) fail, as all
the other sal/osl/rtl functions do, or
- return it differently, by not using pInfo (could be a return value,
success == 0, failure != 0).
I would be happy with a solution that
I am getting heartily sick of constant API & impl. dithering here. Any
chance someone can create a patch that solves their issues and commit
it ? It is -highly- uncertain to me that the discussion of this issue
will yield any sensible result - certainly, I clearly have no idea how
to predict what will be acceptable, or how to write it. The more
pointless quibbling the average patch gets - the more I have to invest
in personal (and team) dental insurance to deal with the results of the
endless grinding ;-)
I'm sorry you consider this "pointless quibbling." I raised my concerns
with your patch in
<http://porting.openoffice.org/servlets/ReadMsg?list=dev&msgNo=17054>,
and never got a response. I think I made my concerns clear in that
mail, and I don't consider it my job to supply a patch that fixes them.
If nobody wants to listen to those concerns, that's fine with me too;
there's more to life than sal API.
Anyway, I (and I think others too) experienced over and over again that
neglected API design in sal lead to high doses of frustration down the
road, so, yes, I might indeed be considered somewhat picky when it comes
to new sal API. Sorry if that pisses you off.
My only criteria is that I don't want the 'intern' API to force users
to write code for error conditions that are never going to happen (or
have to cast return values to void or whatever silliness):
I propose neither.
-Stephan
[...]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
