woodbri at swoodbridge.com
Fri Oct 17 11:11:46 EDT 2008
Jeff McKenna wrote:
> I'd like to follow up on Daniel's comment: shouldn't we remove the
> '--ignore-missing-data' compile switch completely and have this
> triggered only by the CONFIG parameter? Definitely. As I understand
> it, if I forget to set that switch for MS4W users they would still need
> somehow re-compile MapServer for this functionality...which we are
> really trying to avoid here, I thought.
No don't drop the switch. It is a major regression for anyone that uses
the switch on a regular basis. If you regularly build mapserver using
the switch, and I do, then you would have to edit every mapfile that you
ever built to add the CONFIG parameter. This will be insane especially
if you have lots of files and you have to go back to clients and fix it
there when they upgrade. Also you have to maintain two copies of every
mapfile, with and without the CONFIG if you are supporting sites with
multiple versions of mapserver running.
It is not a big deal to leave the compile time switch there. You can
document that the switch as depreciated and the people should use the
CONFIG parameter. In fact the compile switch is only changing the
default value of the CONFIG parameter.
> So my vote is to drop that required switch.
> Daniel Morissette wrote:
>> Sorry if I'm late to the game, still working through emails after
>> FOSS4G and another 2 days conference last week.
>> I agree with Assefa that LOG should be dependent on DEBUG [1-5] and
>> MS_ERRORFILE being set and this needs to be documented. If we decide
>> that DEBUG [1-5] is not required for LOG to work then that would be an
>> exception and needs to be documented, together with the required
>> setting of MS_ERRORFILE.
>> If I had written that RFC I would have been tempted of dropping the
>> --ignore-missing data configure option completely to cleanup and
>> simplify the code, but I can live with it.
>> +1 otherwise.
>> Yewondwossen Assefa wrote:
>>> +1. The only comment I would have would be that the documentation on
>>> this feature should specify that the map or layer need to be in debug
>>> mode for any logging to happen. This is I think the current behavior.
>>> Paul Ramsey wrote:
>>>> I've updated the RFC on ignoring missing data to reflect FrankW's
>>>> comments. It is scaled back a bit now, just a global option, handled
>>>> in a CONFIG, rather than a new mapfile tag.
>>>> Absent violent disagreement, I'd like to start on this one soon.
> mapserver-dev mailing list
> mapserver-dev at lists.osgeo.org
More information about the mapserver-dev