Update for new Pochven region

[ozzyogkush]

In the latest update, CCP split off a new region "Pochven" that encompasses the 27 systems converted to Final Liminality status. These systems no longer have direct Stargate access to their previous connections. We're still seeing those connections come up in Routes. In addition, the systems in Pochven are connected to each other through new standings-based Trig Stargates. Basically 2 ways into each system (one from each direction) in a giant triangle of 3 constellations.

[tcstudios]

I've been looking at putting together a patch for the eve_universe database that puts in all the Pochven changes. The attached sql file is the result of that. A few caveats though:

• applying this currently to a production PF system will likely break routing at least when calculating routes that would previously have gone through one of the Pochven systems. This is because PF tries to get a route from ESI first and CCP is still returning old paths from that endpoint using stargates that this db update removes (causing the PF routing function to throw an exception)
• because of the above, it's been difficult to test if the changes fully address the issue of the Pochven split, so use at your own risk (and make a backup first :) )
• the script changes all of the Pochven systems to a security class of 'T' to distinguish it from other regions. Because it's difficult to test the changes at the moment (because of point 1), it's unclear what all of the side-effects of introducing this class designation has, or if there are any regression issues. Initial testing seems stable, though PF seems to group the Pochven systems as Wormholes. It should be fairly easy to modify the script however to switch that designation to 0.0 (or another already defined class) if necessary.
• I did not attempt to adjust the stations table at all - I don't believe the initial data export used to install PF had any reference to stations in any Pochven systems, but might need to be checked on an install that has already built up a decent cache in the eve_universe db

pochven-patch.sql.zip

[C64Gamer]

Temporary workaround : list the systems in the pochven region and add the 27 to the avoidance list when making the esi call.

[ 30003503, 30004080, 30004079, 30004078, 30004083, 30005192, 30005193, 30005194, 30005195, 30005205, 30005200, 30005199, 30005198, 30005015, 30002761, 30002764, 30002765, 30002768, 30002803, 30002805, 30002791, 30000139, 30000144, 30000142 ]

This should work or?

[rdymade]

Temporary workaround : list the systems in the pochven region and add the 27 to the avoidance list when making the esi call.

[ 30003503, 30004080, 30004079, 30004078, 30004083, 30005192, 30005193, 30005194, 30005195, 30005205, 30005200, 30005199, 30005198, 30005015, 30002761, 30002764, 30002765, 30002768, 30002803, 30002805, 30002791, 30000139, 30000144, 30000142 ]

This should work or?

you missed the 3 main systems Niarja, Archee and Kino respective id's: 30003504, 30002702, 30001372

but yea this should work

[ivveh]

Any code on how to make this avoidance list in the esi call?

[Fumbo]

I pushed the last version of our corp pathfinder with the fixes cherry picked from develop and this pochven issue. What I found is that the ESI call is done with the list of current connections on the map. If the list of connections is too big then it fails on the ESI and fallback to the DB.

So you need both the sql update and the fix in app/Controller/Api/Rest/Route.php to have the route module properly working with pochven.

I'll update here if my corp mates and I report any problems with this fix.

[ivveh]

@Fumbo Hey man, thanks a bunch for this. I'll make sure to let you know if anything doesn't work. Otherwise, assume it does work :D

[Fumbo]

Just to mention that the last commit I did was due to the list above not being in line with exact pochven IDs. This was the only issue so far. It's stable since that commit.

You can safely take my commit 789a5ba

[Tupsi]

@Fumbo Installed your patch on my test system, looks good so far, so thank you for that. I noticed a difference in info section of a system inside Pochven where I was wondering if this is intended or missed. If you look at for instance the system Kuharah on the new system, you will see the box for Sov. being empty while on the old version it shows Triglavian Collective. While I understand that the system is implicitly Triglavian, I was wondering if there is a way to bring that back into the box.

[Tupsi]

The setup page still seems to think that we have the old number of neighbours and displays the wrong total number. Anyone has an idea where this can be adjusted to the new max? Hitting rebuild fixes it until you reload the page, so somewhere is that max number.

[csatlos]

I discovered an unintended side effect of the Pchven SQL patch discussed in this issue. I believe doing an SQL DELETE on records in the system table causes a cascade delete of the child records in the planet table. The SQL script doesn't re-insert the planets again and this causes some script errors when trying to load information about the system on the map. I amended the SQL script to do an UPDATE instead and I believe it achieves the same result but we don't lose the planets.

[Tupsi]

I discovered an unintended side effect of the Pchven SQL patch discussed in this issue. I believe doing an SQL DELETE on records in the system table causes a cascade delete of the child records in the planet table. The SQL script doesn't re-insert the planets again and this causes some script errors when trying to load information about the system on the map. I amended the SQL script to do an UPDATE instead and I believe it achieves the same result but we don't lose the planets.

While trying your change I also found some false data in the whole pochven patch. For instance the systems are not in their new constallations and need to be moved. Also the new Constellatoins have different names then mentioned on evemaps (the all start with Krai... there). If I find the time and make the change, I will reference my fix here. https://github.com/Tupsi/pathfinder/commit/7dee405612681c6436f35ef63a78843e78c2f58d

[neothermic]

Just throwing in as I don't think it's been mentioned, that a more functional problem with this is also the absence of the new pochven entrance/exit wormhole types, and their associated attributes.

[samoneilll]

@tcstudios any chance you might make a similar patch for the new gates that just went live?