Your comments suggest code changes have been made but I don't see them here yet, possibly you forgot to push them live?

I comment as I fix the problems, but push after final testing. Should be ready later today. Thanks for the detailed review.

Not sure what's wrong, but for some reason the revert of this isn't included yet, it still shows e.g. AT-NOE instead of AT-3

After a lengthy session with ChatGPT, it turned out there were two problems on my side. First, I had push right for upstream, which I have now blocked locally (not that I intentionally used it). Second, though git reported my fork to be in sync, it wasn't and still had the old version of osmose_config.py. I have fixed that and can see the latest version everywhere. Hope that helps.

I tried to run the plugin on a part of Austria but it seems the extracts server is down at the moment

I have tested with three provinces (including Niederösterreich, which has > 500.000 segments). Most errors come from either maxspeed:type without maxspeed or maxspeed:source=zone.

Some false positives:

1. warning 'sign without maxspeed`' for

<tag k="source:maxspeed" v="sign" />
<tag k="maxspeed:forward" v="70" />
<tag k="maxspeed:backward" v="100" />

2. Similar story, variable maxspeeds can be fully covered in maxspeed:(*:)conditionals

source:maxspeed=sign
maxspeed:variable=yes
maxspeed:conditional/maxspeed:forward:conditional/... = 80 @ (X); 120 @ (Y)

Probably it's best to disable 303220 if there's other maxspeed:* tags present besides maxspeed:type. (Obviously, make sure not to trigger class 303222 when skipping 303220)

(Also, it may be better to replace [value] without maxspeed by source:maxspeed/maxspeed:type=[value] without maxspeed, as this improves the readability for cases like sign - first example above - or in case there's an unknown source:maxspeed, such as survey, which would now show survey without maxspeed, which is more confusing than source:maxspeed=survey without maxspeed)

1 warning 'sign without maxspeed`' for

This isn't a false positive. It should read source:maxspeed:forward and source:maxspeed:backward, which are obviously different.

Same for number 2. But these are extreme cases not worth dealing with at the moment.

As mentioned, I've run this with 500K segments - and checked the error log.

I can change the message if that's what it takes.

Edit: for 2., it's maxspeed=* and source:maxspeed=* + whatever variable or conditional values.

Ok, I'd personally prefer to avoid requiring unnecessary tagging (if a single tag can indicate one looked to the front and to the back, why tag it twice... but hopefully those people will then use :both_ways instead. Fortunately maxspeed:both_ways isn't used too much, because it'd cause a lot of false positives)

I'll leave it to the others to decide then :)