[GRASS5] [bug #1201] (grass) can't remove base map from reclassed map

[Glynn Clements]

Markus Neteler via RT wrote:

> I also consider this to be a bug which sometimes appears.
> On my machine 4-5 times, see also:
> 
> http://intevation.de/rt/webrt?serial_num=837&display=History
> http://intevation.de/rt/webrt?serial_num=863&display=History
> 
> I don't think that it is a user error because it is pretty
> hard to use g.remove in a wrong way.

Note that this particular bug report refers to vector maps. Also, the
fact that the map is called "tmp.recl" suggests that the reclass is
being (mis)managed automatically.

The most likely reason is that some program is creating the reclass
map automatically, then deleting the files for the resulting reclass
map without removing the entry from the reclassed_to file. Presumably
this is because the program in question predates the existence of the
reclassed_to file.

Viable solutions include:

1. Finding and fixing any programs which delete reclass maps. However,
we should also add a "force" switch to g.remove to override the "is
base map" check, to allow for:

a) the possibility that we miss a case, and

b) the situation where a user modifies the database by hand (e.g. with
"rm", "mv" etc).

2. Removing the reclassed_to file altogether. The concept itself is
inherently flawed, and was clearly implemented without sufficient
consideration of the implications.

-- 
Glynn Clements <glynn.clements at virgin.net>