[mkgmap-dev] dead end check patch v2

Sun Mar 1 16:49:25 GMT 2020

Hi Mike,

I don't understand why you don't add a rule to the points file containing something like
# disable dead-end-check for ways with end nodes having special tags
(amenity=parking_entrance | railway=car_shuttle | railway=rolling_highway)  {add fixme=true}

You did not yet say why this doesn't work for you.

I would also agree to add this in the default style.
I'd also agree to change the java code to handle the tag mkgmap:dead-end-check=no in addition to the fixme tags so that we could write
(amenity=parking_entrance | railway=car_shuttle | railway=rolling_highway)  {add mkgmap:dead-end-check=no}

BTW: I was about to commit your patch v2 with a few changes, but it suffers from the same problem mentioned before: If the style doesn't use the tag railway anywhere you might not find railway=car_shuttle though this is very unlikely.

Gerd

________________________________________
Von: Mike Baggaley <mike at tvage.co.uk>
Gesendet: Sonntag, 1. März 2020 17:30
An: 'Gerd Petermann'; 'Development list for mkgmap'
Betreff: RE: [mkgmap-dev] dead end check patch v2

Hi Gerd, there are several issues. Firstly, omitting ways from the dead end
check because they have fixme tags is not something I want to do. Whilst
this may be one way of reducing the number of reported warnings, and some
users may want to do this, I do not. As it is undocumented behaviour, other
users will probably not expect it to do this either. For my report, I only
want to omit ways that are marked with valid reasons for a dead end (I am
currently using amenity=parking_entrance, railway=car_shuttle and
railway=rolling_highway, but there may be other tags I have not come across
yet). I do not want to delete existing fixme tags because I want to see the
value of any relevant fixme tag in the report.

The current method of using fixme to control whether or not a way with
certain tags on an end node is omitted from dead end checks is obscure and
requires prospective users to have to ask whether it can be done and if so
how. I have offered two simple solutions with appropriate documentation that
will reduce the likelihood of such questions having to be asked. The first,
using a style file is more flexible as it allows the power of the style to
use combinations of values (e.g. amenity=parking_entrance & highway=service
{set mkgmap:dead-end-check=no}). This method does require anyone currently
using dead end checks to review their style file to determine which tags
they want to be omitted from the check (if they are not using the supplied
default style). The second option can only use the existence of a tag on its
own to omit a way (e.g. --dead-ends=amenity=parking_entrance), but does not
require users to make any changes if they want to continue to omit the same
ways as they do currently.

I do not understand why you are so resistant to making a change that
improves the way mkgmap works, has no effect on the actual map and whose
only impact for existing users is a possible slight change in the handling
of reporting of dead ends. In the patches, I also corrected a typo in one
other message and fixed a double space in another. These changes can equally
be said to be not backwards compatible, but I don't believe anyone would say
that this should prevent them from being done.

Regards,
Mike
-----Original Message-----
From: Gerd Petermann [mailto:gpetermann_muenchen at hotmail.com]
Sent: 27 February 2020 11:04
To: 'Development list for mkgmap' <mkgmap-dev at lists.mkgmap.org.uk>
Subject: Re: [mkgmap-dev] dead end check patch v2

Hi Mike,

I fear I don't understand the problem yet. It seems you don't like the idea
that a tag like fixme=* set in the points rules changes the behaviour of the
dead-end-check?

Gerd

________________________________________
Von: mkgmap-dev <mkgmap-dev-bounces at lists.mkgmap.org.uk> im Auftrag von Mike
Baggaley <mike at tvage.co.uk>
Gesendet: Donnerstag, 27. Februar 2020 00:17
An: 'Development list for mkgmap'
Betreff: [mkgmap-dev] dead end check patch v2

Hi Gerd,

Please find attached the alternate patch for improving the dead end check,
which is fully backwards compatible. This adds a new command line option --
dead-ends which lists the tags and optional values that are to be considered
to be valid dead ends when found on the node at the end of a way. The
default is --dead-ends=fixme,FIXME.

Regards,
Mike