OK - I'll look into updating cupid. WHat priority does this have
compared to the STC work and getting cupid to produce votables?
David
2009/1/10 Tim Jenness <[log in to unmask]>:
> On Jan 5, 2009, at 11:45 PM, David Berry wrote:
>
>>
>> Welll, why not retain QUIET, NORM and VERB as they are at the moment,
>> and add levels DEBUG1, DEBUG2, DEBUG3, etc (maybe up as far as
>> DEBUG20?). This voids unadorned integers.
>>
>
> Done.
>
> You now have QUIET NORM VERB DEBUG DEBUG1 .. DEBUG20
>
> From the msgIfget side of things you also have MSG__NONE and MSG__ALL (these
> allow you to specify no messages whatsoever and all possible messages -
> noting that it makes no sense for msgOutif to allow these so it doesn't).
> msg_filter=none will now hide even QUIET messages for absolute silence. ALL
> allows you not to know that there are really 21 DEBUG levels.
>
> msgIfget can also understand integers as well as strings so ilevel parsing
> can now be done with msgIfget. ILEVEL=0 and MSG_FILTER=NONE are identical
> (as is ILEVEL=NONE of course if you use msgIfget).
>
> It might be easiest to port cupid first. CUPID seems to use the following
> tests for display, equivalent to output filtering levels for msgOutif of
>
> > 0 ie 1 MSG__NORM ?
> > 1 2 MSG__VERB ?
> > 2 3 MSG__DEBUG ?
> > 3 4 MSG__DEBUG1 ?
> > 5 6 MSG__DEBUG3 ?
>
> CUPID default is ILEVEL=1 which is equivalent to the >0 clause so there is a
> mismatch between MSG__NORM and ILEVEL=1 though (since MERS uses 1 for
> QUIET). I reserve 0 for absolutely no messages because QUIET is meant to be
> for minimal messages but not zero messages. Maybe it won't matter and we can
> just change all the >0 clauses to MSG__NORM. In most cases calling msgIfget
> and then removing ilevel from the API, replacing it with msgOutif and
> msgBlankif should be sufficient. Should clean up the code as well.
>
> One routine actually checks for ilevel in a range and does not do something
> unless it is 2 or 3. This is not possible in MERS.
>
> --
> Tim Jenness
> Joint Astronomy Centre
>
--
Note my change of e-mail address. Please send e-mail to
[log in to unmask] from now on.
|