Currently working on...

Users who are viewing this thread

Status
Not open for further replies.
whydott said:
gsanders said:
whydott said:
Just want to know, will you fix unlinked village that cause undefeatable faction in this  patch? or is it something not save compatible thing, then you decide to defer it?

  make a list of every unlinked village you think you can find and I'll see what can be done about attacking them correctly to some castle.
I found Thambuldor and Eytherudren of kingdom Nord. also Emakadi of Swadia in reformist video

EDIT : Sarranid Shadowmont

  Emakadi  is already bound to Valhasnosnar_Castle of the elfs.  Possibly the elf lord defected to swadia then left swadians?

---
  Shadowmont was listed as a necromancer village but never assigned to any castle
  Thambuldor is listed as a Dwarf village but never assigned to any castle
  Eytherudren is listed as an elf village and never given to a castle

  I'll have to see if I can run some script to look for unassigned villages. There's probably all sorts of them.  These are Guspav era issues.
  I suppose once I have a list I need to find out who owns that castle and update both the village and castle notes, as each one changes after changing the assignment for  slot_village_bound_center    at that village.  It looks like its a one way assignment; the village is marked as belonging to a castle, but the castle doesn't get any new change.

  Thus for the situation you report where a village is showing as being in a faction but all the castles are taken, I need to check if the lord of each castle is exiled and if so re-set ownership of the castle (and attached villages) back to the faction that owns the castle.
---
  Its not this, the test will have to be longer but I'll miss helping Reformist if I delay further.  Reformist seems pushed by his desire to pump out updates but he doesnt much wait for me to catch up, so I have to get him what I can now.  I'll take time to resolve this isolated village issue more carefully tomorrow instead of in the last few minutes and miss a window AND maybe not catch all the root causes for lack of time.

 
gsanders said:
whydott said:
gsanders said:
whydott said:
Just want to know, will you fix unlinked village that cause undefeatable faction in this  patch? or is it something not save compatible thing, then you decide to defer it?

  make a list of every unlinked village you think you can find and I'll see what can be done about attacking them correctly to some castle.
I found Thambuldor and Eytherudren of kingdom Nord. also Emakadi of Swadia in reformist video

EDIT : Sarranid Shadowmont

  Emakadi  is already bound to Valhasnosnar_Castle of the elfs.  Possibly the elf lord defected to swadia then left swadians?

---
  Shadowmont was listed as a necromancer village but never assigned to any castle
  Thambuldor is listed as a Dwarf village but never assigned to any castle
  Eytherudren is listed as an elf village and never given to a castle

  I'll have to see if I can run some script to look for unassigned villages. There's probably all sorts of them.  These are Guspav era issues.
  I suppose once I have a list I need to find out who owns that castle and update both the village and castle notes, as each one changes after changing the assignment for  slot_village_bound_center    at that village.  It looks like its a one way assignment; the village is marked as belonging to a castle, but the castle doesn't get any new change.

  Thus for the situation you report where a village is showing as being in a faction but all the castles are taken, I need to check if the lord of each castle is exiled and if so re-set ownership of the castle (and attached villages) back to the faction that owns the castle.
---
  Its not this, the test will have to be longer but I'll miss helping Reformist if I delay further.  Reformist seems pushed by his desire to pump out updates but he doesnt much wait for me to catch up, so I have to get him what I can now.  I'll take time to resolve this isolated village issue more carefully tomorrow instead of in the last few minutes and miss a window AND maybe not catch all the root causes for lack of time.

no problem buddy :fruity: no problem
also Vaegirs Ulburban
 
patch v150 Nhot patch

1) Fixed Ibdeles castle.  this problem came from an OSP add-in; it actually isn't Guspav's problem after all, once I tracked down when and where the exterior castle scene for castle 21 came from.  The fix is just copying Native's castle 21 exterior over the one at the mod.

2) increased mana regeneration rate for mages by approximately +50%.

3)a) removed black bolt from secondary spell selections for NPCs.  There seems to be a systemic issue with the entire fear/recovery cycle in sieges.
    ^^^ this was already done.  I'll have to check if any of Reformists' troops have explicit black bolts in inventory; perhaps some necromancers or liches.
  b) Removed black bolts from Mage defectors, mage deserters, and all Necromancer kingdom lords & King.  This is for both new starts and existing save games by the same script methods used successfully before.

4)a) the NPC cleric/paladin timed procedures  "common_battle_turn_undead"  and  "common_battle_rebuke_undead" now check for player and if so skip the auto-cast.  Secondly they adjust the procedure to accommodate mounts; normally spells that need the caster position (anything with a radius check or area effect) get garbage when trying to get the position of a mounted caster; now the correct position is given mount or caster, depending on if mounted or not.
  b) likewise  "common_battle_flame_strike",  "npc_timed_cleric_spells" {Heal, Bless, Flame Strike, Curse, Dispel magic}
  c) likewise "pc_cleric_spells" {Heal, Harm, Flame Strike, Fire Storm, Fire Storm, Turn Undead, Bless, Curse, Bravery, Terror}
        Terror might crash the game, especially at a siege,, but in this one case its not something auto-cast. 
        You the player has to choose it deliberately, so its on you.
  d) likewise "common_battle_healing_spell"
  e) likewise within "presentation_prayer_book"  {positive energy burst, negative energy burst, fire burst, cauterize, dispel magic}
        also {protection from fire, protection from cold}, not that either currently cares about range anyway.
        removed restriction on turn undead while mounted for player
        summoned cleric weapons will auto-remove previously summoned cleric weapons, thus clearing depleted throwing weapons

5) disabled mage auto cast of "dispel magic", as this delayed response to fear/slow/acid seemed to trigger crashes.
    Better only a few very stable, very reliable spells than a full spellbook of flakey spells, even if it seems to trim Guspav's wide selections heavily.

6) slight increase in speed for Hellish bolt (at itemkinds.txt)

7) Adjusted inventory for Rhodok troops of all types so they stop cycling weapons looking for the best weapon when out of range of the enemy.
    This PBOD behavior wastes time better spent swinging at enemies or closing distance while in formation.  The "synchronized ballet" aspect is because they each have more weapons than necessary in equipped weapons slots.  This over supply of weapons is pervasive in all of Guspav's troops, but for today I'll tweak Rhodoks and see how they change.  The idea of assigning 5-6 weapons to each troop is to make them seem more motley and less uniform, but it leads to specific syndromes in PBOD.  Some troops really do deserve to be motley, like zombies or orcs, but some troops being uniform make cultural sense elsewhere.  We'll start small today.

8 ) radius of effect increased for turn undead (from 4 to 7 meters) and rebuke undead (from 4 to 8 meters).

9) sieges now decide whether the player is attacking or defending before manipulating reinforcement waves, so that the player has even handling regardless of which side they are on in a siege.
  "common_siege_defender_reinforcement_check", "common_siege_attacker_reinforcement_check"
and directly at "castle_attack_defender_walls_sally", "castle_attack_walls_belfry", "castle_attack_walls_ladder"

---
10) Adjusted starting assignment of towns, castles and villages for new games to have less Guspav micro-management of assignments and more automatic assignments of villages to centers, in order to reduce the number of orphaned villages seen currently in future fresh started games.
This involves changing the sequence in which villages are bound explicitly to castles so that automatic binding occurs AFTER the first wave of explicit bindings are done.  Before now this was backwards and in large part responsible for the syndrome seen where some villages belong to factions but no bound center exists, so that when all the faction centers (castles, towns) that can be sieged are gone the faction still owns some orphan villages. 

  At the moment this impacts fresh games only, but I've decided "N" will be the new base for the mod at Moddb etc.  I won't support any patch earlier than N, since apparently it needs a freaking act of Congress to get any Moddb'ers to bother to update.  OK I'll motivate them.  Perhaps they're afraid of their computers and patching is completely beyond their skills.  I could believe that.

11) Reverted after battle loot/party removal to M-5 for now; this re-opens vulnerability to freeze after battle but it does close two other potential after battle issues/freezes created only with M-6.

12) reform of morale report and large party morale.
  script_get_player_party_morale_values    simple_trigger # 19
  added bonuses for status: Monarch, Marshall, Policy Serfdom, Policy Quality (which will impact actual party size penalty shown in report)
  base values for food in report already reflect +50% bonus;
  leadership skill reduces the overall morale penalty from large parties
  the break down of how much each of the above contributes to changes is NOT explicitly printed in morale report, only the total effect.  Deal with it.
  base morale set to 60 instead of 50 for party size 0.  Don't think you can suddenly stop feeding them.

Some tutorial seems necessary (all of these added 6 months ago):
  You have been given tea, juice, drinking water, ale, wine, feywine as valid drinks. Each counts as a separate type of food for morale purposes.
  You have been given smoked venison, raisins, dried fruit as new foods that never go rotten.
  Separately, you have been given venison and fish as fresh foods harvestable from the land within Rigale system; but preserve these or they'll rot.
    Butter, chicken, pork, beef, venison, fish, fruit, grapes can all rot.  Some of these can be preserved; look in camp:actions menu.
    Grapes and raisins are different foods for morale purposes; as are fish and smoked fish, dried meat and (raw) beef; sausages and (raw) pork;
  dried fruit and fruit.  Once you understand that food/drink for morale purposes counts once only each unique category of food (regardless of modifier, except rotten, which skips being counted), then you can plan your logistics accordingly.  If you aren't competent to feed your troops, you're too stupid to lead an army.  Sad fact.

13) set "$g_dplmc_lord_recycling" = DPLMC_LORD_RECYCLING_ENABLE  to prevent problems with towns becoming corrupted as lords are exiled.

14)  updated the chain of calls that leads to caravan creation:
a)  triggers:  tripled the chance to create a caravan & placed a display message when a caravan is requested to be made
            (but before knowing if there are too many already for that faction).

b)  "script_create_kingdom_party_if_below_limit"  added a display message when a kingdom caravan has been selected to be made
c)  "script_cf_create_kingdom_party"
d)  party_templates "pt_kingdom_caravan_party" has been tripled in strength to a peak of 135 mid tier troops
        (including some journeyman mages and several clay golems) to better survive roving bandits

e)  "script_cf_select_random_town_with_faction"  changed to use list_01 to return a random town from a given faction, if any still exist
f)  "script_cf_select_random_town_at_peace_with_faction_in_trade_route"    changed to list_01
        this had not been called, giving trade caravans garbage destinations; and, using list system means returned parameter is reg1 not reg0.
g) all of the existing kingdom caravans are glitched and need a one time removal by script as part of the "M-7" scripted update process

example showing boosted caravan for originating at faction capital:
TJaWl.jpg

it would look possible, but actually this caravan will NOT grow weekly:
  the weekly necromancer check filters this party because it is not led by a hero (implicitly, party stack 0 is not a hero)
Cy8m4.jpg

15)  auto-load patch M-7's script for inventory changes

Status:  NexusMods  and Google Drive

 
gsanders said:
patch v150 Nhot patch
10) Adjusted starting assignment of towns, castles and villages for new games to have less Guspav micro-management of assignments and more automatic assignments of villages to centers, in order to reduce the number of orphaned villages seen currently in future fresh started games.
This involves changing the sequence in which villages are bound explicitly to castles so that automatic binding occurs AFTER the first wave of explicit bindings are done.  Before now this was backwards and in large part responsible for the syndrome seen where some villages belong to factions but no bound center exists, so that when all the faction centers (castles, towns) that can be sieged are gone the faction still owns some orphan villages. 

So... Is this supposed  to give the player 17 villages? Because when I start new games I begin with 17 villages >.>

A3EF9B790E69A6ADD4362C15861BD0E633B98DFC
 
mrpeters said:
gsanders said:
patch v150 Nhot patch
10) Adjusted starting assignment of towns, castles and villages for new games to have less Guspav micro-management of assignments and more automatic assignments of villages to centers, in order to reduce the number of orphaned villages seen currently in future fresh started games.
This involves changing the sequence in which villages are bound explicitly to castles so that automatic binding occurs AFTER the first wave of explicit bindings are done.  Before now this was backwards and in large part responsible for the syndrome seen where some villages belong to factions but no bound center exists, so that when all the faction centers (castles, towns) that can be sieged are gone the faction still owns some orphan villages. 

So... Is this supposed  to give the player 17 villages? Because when I start new games I begin with 17 villages >.>

A3EF9B790E69A6ADD4362C15861BD0E633B98DFC

  No, I'll look at that today.  Probably the issue is player is troop 0, and uninitialized fields are always 0.  I'll write a -1 in the town slot "slot_town_lord" early in the initialization.  You'll have to restart your 1 day old game though.  I expect N2 tonight to have a few other cleanups as well.

 
  patch "v150 N2":    hot patch

simple fixes:
1)  villages are marked as owned by troop -1  early in the start game so that player does not inherit 17 villages, as player is troop 0.
    This fixes an issue seen only with new starts in patch 150N.  Games started only last night under patch N need to be restarted, otherwise its unfair to existing players who really do own villages to suddenly "hot patch" their villages away.

2)  when patch N2 is applied, any parties & centers that belong to a faction that is NOT
      {player_faction, player_supporters_faction, necromancers, orcs, drow, and finally  also not equal to $players_kingdom  [which implicitly holds vassal or mercenary allegiance when not (-1) or (player_faction)]}
      AND contain skeletons, zombies, shadows, and all of the necromancy faction troops (dark mages, necromancers, liches, and all the mummy types)
      have those troops removed.

this skips bandit and free undead parties, anything the player contributed to his own faction or as a vassal or mercenary to the faction (s)he serves, all the necromancer faction, and the "natural allies" to necromancers.  Mageocracy although neutral to Drow is actually opposed to Dark mages and Necromancers, so they also are correctly pruned of undead.

    This fairly slow process should be a one time correction, with possibly monthly follow-ups, made necessary mainly by unpatched games extending back to v150B when briefly there were many undead spawned.  Even though it doesn't happen to me, I seem to be the rare exception. 

  EDIT:  testing on MY PC showed this needed between 1.5 and 2.5 seconds with the typical 1100 parties; but the same PC needs 1 minute 40 seconds to load the mod from Warband's first screen.  Most likely, your PC needs more.  I'm not sure HOW much more, but probably it could be worse.  I should probably set this to run every 30 days as well...

3) undid all of patch N's changes to starting village creation.  There are too many issues opened up by that layer of N's changes.
  Players who did NOT have an on-going savegame yesterday need to restart (that is, ONLY need to restart new games made since last night, otherwise you're fine). 
  EDIT: tests correct now.

Statusuploaded to NexusMods and GoogleDrive

 
calendar for the next two days:
patch  v150 N3hot patch

1)  Player agent is excluded as a target of Charm Person and Chaos spells.
    This is needed in order to prevent an issue with reinforcement waves, as each side's numbers of live agents spawned and active at a battle are added up by comparing their relation to the player, which flips once charmed.  Thus it breaks reinforcements as they are currently implemented.

2) a) Reinforcements of enemies worked for 1-2 enemy lords but fails to count the agents correctly when 3 or more lords are attacking (and perhaps, defending).

    b) the timing of defenders and attacker reinforcement checks causes bubbles of bandwidth heavy use, so the game pauses noticeably every few seconds at a large siege.  This is unnecessary and could be better reduced by staggering the defender and attacker checks to not overlap.

    c) further, the attacker and defender reinforcement checks are currently performed twice, mainly because undead sides had a raw battle disadvantage imposed by giving the non-undead side credit for each paladin or cleric in their force as counting for two agents instead of one, as a rough estimate of the damage inflicted on undead for purposes of calculating true odds in a fight, but only when 50% or more of the undead force actually is undead.  Now that undead armies are trimmed the adjust is no longer needed, which in turn reduces pressure for a second wave of reinforcement checks which used a different method for determining reinforcements.

    d)  As the battle commences, the reinforcement check first determines which side the player represents: attacker or defender, and updates this stored value.
Now as before the attackers and defenders present and not (dead/in a rag doll state) are tallied to understand if reinforcements need clipping or should be green lighted to continue for this round of checks, with a better scheme for deciding how many of each exist regardless of how many distinct lord parties are involved, as before only 2 parties per side were watched/available for use as the limiting number for that side's reinforcement check. This allowed 1000+ troops to slip into certain large sieges, which is disaster for the average player PC.

  Now that reinforcements are (finally) stable, the main issues for open battles is lack of LODs defined for every last troop model, for which is a much more tedious fix (a full week of non-coding OpenBRF tasks??) isrequired.  It could be done by a community volunteer... not that I'm nagging or pushing anyone into it...

3)  The version number displayed at the 150 series was not really wanting the series displayed as  0.150    but really 1.50, to eliminate confusion with Phantasy Calradia's .723 patch, which some people insist on applying over the 150 series, with predictable failure to load. They then blame me. Its never they did something wrong. I simply write incompetent software, a theory they are anxious to spread by social media.  Now maybe I am a little fast and loose with patches, but they're turnips for putting a patch made in 2014 over a base made in 2018.  Nevermind this, I could help by moving the display to be 1.500 instead of 0.150

4) charged staves are modified to filter the false added attacks shown in the battle message feed.
a)  now, if a spell hits but the agent doing the casting of the spell still has an equipped charged staff in their hands there will not be a message
given or wizard staff damage placed on top of the spell damage.
b)  the test for which type of charged staff is corrected so that 6 false positives are not given with a cascade of distinct damage given; also the wielder is not charged 6 times the mana points, with no clipping of effects when the caster goes to negative mana.  The test for available mana was for a single use and instead up to 5 times that amount was pulled then quickly cancelled elsewhere, leading to charged staves being over powered by what essentially was an exploit.  That is closed.

5) Golems now have some reforms:
  a) at agent spawn time, various golems are given adjusts on damage output, ranging from 133% to 300% of normal damage.
  b) at agent spawn time, various golems are given added strength, as it was difficult to encode strength beyond 24 previously at the troops module.  This in turn impacts encumbrance, which impacts movement speed.
  d) these adjusts should also apply to Ogres and the Ogre upgrades
  c) in order to protect  a),  golems need to be exempt from Cleric/Paladin bless and curse effects, otherwise their damage would be reset back to normal (100%) after 30 seconds; likewise dispel magic, haste, slow, and chaos.

  e) finally, recruitment of golems is adjusted so the player that for example could purchase a higher tier golem might want to purchase instead some multiple number of lesser golems.  Economy of scale production at the Ridas Mageocracy has already made bulk sales of clay golems to royal traders everywhere, and this economy of scale might transfer to a wholesale pricing for quantity purchase to select lords and players who are well liked within the Mageocracy.  As your relation increases with Ridas Mageocracy, more options are shown.

Progress so far:  Reinforcements are exactly correct, after far longer than I would have liked.
  I didn't understand how the parameter at        (store_normalized_team_count,":num_attackers", 1),
  really worked, so using it for feedback later did not do as I thought I asked.  Here's the tip:
    (lt,":num_attackers",7),    # GS "6" here seems to mean 6 waves so next wave starts when there is  (Battle size - # friendles/enemies standing now) space available to fit  1/6th of the defenders/attackers waiting to reinforce.  By setting this number higher, the wave can fit sooner, but will be smaller.
 
    so that later,
    (add_reinforcements_to_entry,4,7),    # <-- this is adding reinforcements to 1/7th of the attacking army's remaining force, IF it fits, else skip reinforcement
                                                            #  This is further changed by your battle size setting
    There are two distinct cycles per side instead of one and their periods are set to not overlap but to allow acceleration of reinforcements if either side starts taking huge losses.  I should also have one use much smaller waves but have a longer period so if one side is taking few losses it can still get reinforcements, while sides taking heavy losses can have a stream of troops moving from spawn point to where they are needed constantly.  This would be a good progress.  At the moment both cycles per side still use value 7, but one should be perhaps 9 not 7, and slightly over 2x longer (or maybe slightly less than 1/2 the period, for the small reinforcements) period than the other loop.  OK, tomorrow.

  I use a third throttle to make sure the battle does not slowly creep larger for either side when battles are very large (500+), counting only standing agents and not also counting rag dolls like usually get counted.  This keeps a fairly uniform number of agents per side, according to the raw battle advantage, which normally handicaps one side or the other.

6) some minor changes for Necromancers:
  a) the weekly stars aligning event is now tied to the setting of  "$g_cleric_spells_on"
      which is toggled at Phantasy Settings: Cleric spells enable.    It normally is Enabled, and the disable is for debuging crashes.  This debug checking now extends to Necromancer functionality as well.  You're welcome.
      If cleric spells are turned off, this also is turned off, except trim which occurs anyway.

  b) #free undead parties convert some prisoners  ln 2867  the timing is changed from 47 to 97 hours
      the interest here is to see if it has any impact on how long prisoners remain with the player party, as this will silently strip prisoners from the party they are at.  Although the party template has to match certain values, it is possible some party template in the list was blank and applies to player party as well.

  c) Necromancer faction reinforcements for first three months were tied to who owns a specific location, which is dangerous, as it might change hands several times and place undead at various regional capitals. This mechanic is removed completely for necromancers only.

  d) s_t 127 'raise dead enemies as undead'  is now likewise tied to status of  "$g_cleric_spells_on"


Status:  removed

  player charged staff doesn't seem to be charging at spawn but that is a separate issue...
EDIT:  charged staff never initializes because the agent never finishes initializing on agent spawn. 

    This causes many different symptoms; but the fix is relatively quick so an N4 will launch today with all of N3 and one or two further fixes.

patch v150 N4:
  1) fixes the issue with failing to initialize agents spawning in battle, resolving issues with
      haste/slow/dispel magic, dodge, magic mana recovery, and charged staves.

  2) slight tweak to reinforcement waves to better cover corner conditions where only small reinforcements were needed, and to move the second cycle of checks for reinforcements to a longer period to use less CPU bandwidth in large battles.

  3) disabled NPC magic user haste and dispel magic during sieges, as it was at N2 and below.

Status:       uploading to NexusMods
         
 
  Last night I considered how to replace the currently not working Diplomacy supplied recruiters with something that uses different mechanics to achieve the same goal.  I came up with a presentation driven system where a recruiter uses the caravan trade routes for that town to select cities to recruit from, and the recruiter option is only active at towns, not castles.  It still needs considerable work, probably 2 days reshaping Jrider's faction presentation into something that can instead display the 15 trade route destinations and whether they are open or closed in terms of war status or not. 

  From there, a list or pseudo-list has to be made where the fairly complex structure of {destination town, originating town (as the player might have more than 1), number of recruits, type of recruit (male/female/mixed?), and time remaining before they are auto-added to garrison} needs storing.
There needs to be some daily simple_trigger timer that updates the time remaining once a day for all pending recruiting activities and when the days left hits 0 to add to the garrison the correct faction troop in the correct numbers then remove from the recruiter list the 5 entries for that single order.

  The selection interface to choose the recruitment targets needs building, the distance to each town needs calculating, that should impact the cost per recruit and number of days needed to receive troops and thus recruiting needs a presentation to show all of these variables at a glance, so that orders can be placed.  I would say each town could support an order of 30 people, with costs per person somewhere between 3 and 5 times normal to cover overhead in safely transporting them to your town from their origin.  Limiting the origin towns to the established trade nodes for that town I think makes it more interesting, plus every town has trade routes already recorded, making it considerably easier to find towns.  The towns are usually fairly far away.  Putting a high premium on the convenience of the service gives players an incentive to travel to nearby villages themselves while allowing 1-2 weeks delay for their recruitment orders when they are too busy otherwise. Making them wait 1-2 weeks keeps them from relying solely on tier 1 recruiting.  And maybe there should be a potential for tier 2 recruits?  Or specifying male only or female only for a little extra cash?  And should there be discounted refugee recruitment??

    I mention this to show that instead of thinking "oh, recruiting is broken.  Yet another time the mod lets us down, other mods have working recruiters".  Maybe you're not looking at it correctly.  I can either chase my tail making the mod INTO a clone of every other mod, or I can say "OK, heres a layer that isn't working like every other mod.  Maybe it wasn't meant to"... and ACT instead of only REACT.

-----
  started on this but it is much more difficult than some other things I've done so it needs several days.
  probably at the 60% mark now.  I have the raw information needed but getting it to display neatly will be a real grind.
 
There are some problems with the reinforcements. The defenders get all clumped on top of the wall (not even being able to move and being a perfect target to AoE spells)
 
vrork said:
There are some problems with the reinforcements. The defenders get all clumped on top of the wall (not even being able to move and being a perfect target to AoE spells)

That's a native problem, you'll have to give them directions and tell them were to position themselves to declump them.
 
kazgruta said:
vrork said:
There are some problems with the reinforcements. The defenders get all clumped on top of the wall (not even being able to move and being a perfect target to AoE spells)

That's a native problem, you'll have to give them directions and tell them were to position themselves to declump them.

Hum, but then i think at least a bigger cooldown between the reinforcement waves would be good, especially since the AI (i was just helping the defenders) doesnt know how to do that
 
vrork said:
kazgruta said:
vrork said:
There are some problems with the reinforcements. The defenders get all clumped on top of the wall (not even being able to move and being a perfect target to AoE spells)

That's a native problem, you'll have to give them directions and tell them were to position themselves to declump them.

Hum, but then i think at least a bigger cooldown between the reinforcement waves would be good, especially since the AI (i was just helping the defenders) doesnt know how to do that

The problem may be more the size of people garrisoned. Gsanders added the reinforcements because sieges and battles were taking to long because of small armies sizes. Its much better the way it is now. Just keep smaller armies in your garisons and it will solve your problem.
 
I'm right now examining reinforcements, AGAIN.
  It seemed like it was working then yesterday I tried some tests and it was broken again, so now its on my desk once more.


  =====
  I had a big break through today on reinforcements, so at 13:29 my time I can say it's resolved.  I found exactly what I needed to know to solve it,
after much time burned testing and dumping state of each parameter involved.  It's added to my documentation as to how certain calls actually work.

  Not sure if I should ruin Valentine's day with an update or not, but at least reinforcements work.

  There's plenty of other things that need doing, just not reinforcements, finally.
  When there are not hundreds of troops spawned on either side, they won't crowd up on the walls.  I attacked a 500+ undead garrison and never had more than  100 defenders spawned, and eventually blew through all 500 defenders in one segment without having to span more than one segment.  Frame rate stayed close to 60 FPS the entire time, although I do have better than usual CPU and memory / disk access. My video card is not very expensive:

  CPU  i7-4820K  @ 3.8 Ghz all cores
    RAM  16 GB  DDR-3 1600 CL9  but arranged as 4 discrete memory channels of 4 GB/channel, which is much different than hobbyist rigs
    Chipset  Intel X79
    Disk:  dual  Samsung EVO 850  SSD  boot/compile +  discrete games drive    non-RAID
    Video:  Nvidia GTX 950  / 2 GB video DDR5
    OS:  Win 10 Pro x64

  There are faster rigs for gaming, but I have remarkably little time to play games.  I mainly need fast turnaround time for compile then load the mod, which this gives me  1:20 + 1:40 so just over 3 minutes per pass, with the 1/4 million lines of script forming "Phantasy 2018".
  But it did let me bang out 40 different tries to hammer down this one issue...
 
 
gsanders said:
I'm right now examining reinforcements, AGAIN.
  It seemed like it was working then yesterday I tried some tests and it was broken again, so now its on my desk once more.


  =====
  I had a big break through today on reinforcements, so at 13:29 my time I can say it's resolved.  I found exactly what I needed to know to solve it,
after much time burned testing and dumping state of each parameter involved.  It's added to my documentation as to how certain calls actually work.

  Not sure if I should ruin Valentine's day with an update or not, but at least reinforcements work.

  There's plenty of other things that need doing, just not reinforcements, finally.
  When there are not hundreds of troops spawned on either side, they won't crowd up on the walls.  I attacked a 500+ undead garrison and never had more than  100 defenders spawned, and eventually blew through all 500 defenders in one segment without having to span more than one segment.  Frame rate stayed close to 60 FPS the entire time, although I do have better than usual CPU and memory / disk access. My video card is not very expensive:

  CPU  i7-4820K  @ 3.8 Ghz all cores
    RAM  16 GB  DDR-3 1600 CL9  but arranged as 4 discrete memory channels of 4 GB/channel, which is much different than hobbyist rigs
    Chipset  Intel X79
    Disk:  dual  Samsung EVO 850  SSD  boot/compile +  discrete games drive    non-RAID
    Video:  Nvidia GTX 950  / 2 GB video DDR5
    OS:  Win 10 Pro x64

  There are faster rigs for gaming, but I have remarkably little time to play games.  I mainly need fast turnaround time for compile then load the mod, which this gives me  1:20 + 1:40 so just over 3 minutes per pass, with the 1/4 million lines of script forming "Phantasy 2018".
  But it did let me bang out 40 different tries to hammer down this one issue...

That is great to hear. You have a better processor than me and SSD, a thing i dont have. I still have to get one big enough for my steam lib.
 
vrork said:
That is great to hear. You have a better processor than me and SSD, a thing i dont have. I still have to get one big enough for my steam lib.

  Its not for bragging rights, its to compare MY 60 FPS with whatever you get.  Otherwise I could unlock FPS and get say 95 FPS and its a meaningless number without including specs, resolution, sunspot activity over Earth, etc.

  In general larger SSDs have more bandwidth than the smallest ones, due to writing across multiple memory channels internally.  So a 128/120 GB SSD performs more poorly than a 256/240 GB SSD.  I found price/performance was good a few years ago (the last time I had money) for 240 GB drives and the 480 were too expensive and 120 were too close in price to 240.  Thus I invested in 240 drives, but the key is write performance.  Most read well but cheap ones write poorly.  It helps to research.
 
So, I found the cause of N4 not making fresh games correctly.

  There was ONE line with a two character typo that clipped the rest of the initialization from happening. The mistake was so small I just kept not seeing it.  Even automated tools showed there was a "change" and I nodded my head saying thats right, thats a new line without SEEING the mistake.

  N5 uploaded to Nexusmods.
-----

  In other news I'm somewhere around 97% done with building a replacement recruiting system.
  obviously distance needs better contrast, but the first priority was to align the columns and make sure each field was driven by the correct data, such as the ex-drow capital giving the correct culture troops, or the red warning for prejudice at the Orc capital to not bleed red into the next field's default color.  When the player changes the number ordered I need to be able to save that to a new list, so that is this afternoon's job, then count down status each day and show the orders already placed so that town is locked for new orders until arrival.

The number available at each town is roughly 30 but adjusted by prejudice reaction to your current party plus minor adjusts by player character class.  Rogues for example get more recruits and a discount on recruits, who may or may not be a little shady...  This is how Orc capital has only 5 recruits available; the standard village algorithm of prejudice scrubbing recruits applies.  War status block recruits from that town.  This was taken at Obsidian Spire, which Blazing Hand now owns, and has been migrated to Mageocracy culture.  Ahmerrad is migrated to Blazing Hand culture.
jP85M.jpg
yyE3k.jpg
BoHnc.jpg
1RpgJ.jpg

Now changing to Ahmerrad,
RbNJg.jpg
-iPjP.jpg
  (except this is an old picture, the "ordered" radio button is gone.)
  we see the effects of setting Culture (via Migration) to Mageocracy.
  so even though we're at war with Ridas mageocracy, mages are available for sheltered recruiting.
  We can also recruit golems but not via this interface.

  I'll remove duplicate trade routes.  That is one of the reasons show trade routes was added to the (cheat menu) hidden Reports: Reference Materials.  Unless enable cheat menu is on, the trade route dump is hidden.  That is not my cruelty, rather it dumps to recent messages and looks like it did nothing, so I didn't want anyone frustrated by a thing that seemed "broken".  It actually does just enough to be useful.  There is much in this mod that isn't very obvious to casual users.

---
  Looks like the only duplicate route is Ahmerrad to Obsidian Tower.  I made a script to remove the second route at each side, as well as prepare a new permanent list to store, track, and update orders placed at the recruiter.  Finally the update adjusts gear for Drow Lolth's chosen to remove drow scythe and substitute a different weapon.

Capturing changes made by the player to select town and number of troops turns out to be needlessly difficult, so after two days of trial and error I decided this was wasting time and made a menu replacement, which also solves the floating dialog box from constable during the presentation.
Instead, this is called from menu system at the same place where town menu recruiting occurs.

TLDR
  Next up I need to drive the permanent list of reinforcements ordered in order to update status for orders already placed for that destination town and to update days remaining each day so that troops can be delivered securely when the countdown hits 0.

  Besides testing the only thing I THINK still needing done is generating a message when troops are delivered.  Otherwise it looks about done,
but I'll test it tomorrow.  Oh, and Murphy's Law:  just when I'm almost done I found where the regular diplomacy recruiter was stuck not moving and how to un-stick him/her/it.  Like, after 5 days making an alternative system...

EDIT: #Gazillion+1:  so the permanent list I chose was getting thrashed by some background task because it was not a dedicated list, did that and now it is starting to smell like soup in my kitchen.  I wanted to watch new orders change status in real time so I disabled the normal limiter that keeps player from going back and placing another order.  The idea here is to see the status change from "ready" and the number available so that once I've picked that town to send me some recruits it is drained and none are available this round, plus I should see them in the next column as underway:
M_FCl.jpg
6TvoJ.jpg

The destination and origin town names were swapped but at least I got my first recruits from the system, a difficult case where I asked for mixed troops (male and female arriving as a set) to test the system.  In a few more game days I should see more arrivals.
8K-2P.jpg

  Here I see the last delivery is stuck -- it delivered fine but I forgot to clear the order, silly me.  The other orders are counting down the days nicely.  I suppose I could also shrink the  "order" status in column 5 and "ready" status  in column 4 while I'm making changes.

  I should restore dialogs from M or earlier so the normal village recruiter is available too, now that I maybe have it working.  It was missing a merge into simple_triggers, and there are a bunch of targets for me to run down now.  I doubt, for example, the training was doing anything, so there probably is a missing simple_trigger to install from original diplomacy that got missed in making Pho_II source, which then was missing when folded into Phantasy Calradia, so that it is borked for Phantasy 2018.  It's getting a little clearer what led to what...
jG55r.jpg


Status:  Update Today ?? (I hope), Fri 23 Feb

 
  patch v150 N6: hot patch

1)  The main change is a new mechanic to recruit troops in place of the weak and seldom successful Diplomacy recruiter, who usually gets eaten by bandits even if he did find a way to move (and now, suddenly, he moves).  This new mechanic is found at the town menu: Recruit Troops menu,
and is used IN PLACE OF your normal town recruiting (so its either one or the other, but not both).  This lets you ask a caravan (abstracted, there is no caravan for you, or enemies, to interact with on map), securely deliver recruits (at a substantially higher cost than if you or a recruiter recruited them personally) from any trade town in the trade route list for your owned town.  Thus this does not service castles, only towns that you personally own.
The mechanics are fairly complex so I'll refer the rules used to a different thread (game mechanics).  This took a FULL WEEK to make!

2) a) Drow Lolth's chosen now traded in weapons that can't be used while mounted for something very nice that can be.
    b) Charged Staves now do the normal damage as well as magical damage, and they make less noise when "firing".  Maybe.
    c) "Obsidian Tower" and "Ahmerad" each placed trade routes to the other... which is a technical mistake, since placing a route ADDS one at BOTH ends of the route, making duplicated routes at both locations.  I made a little script that removes the second route when this patch is auto-applied, meaning after YOU patch to N6, sometime in the next game day the script that removes the second trade route at each of the two towns  (and also the scripts that remove Drow Lolth's chosen's useless weapon and auto-insert/equip the replacement are ran.  So actually you need to let your in progress game continue for about a day before you see all the effects.
    d) freshly created games after patch N6 of course instantly have all the changes, through a little magic on my side.  Its technically much harder to mend games already underway than new games, but both cases are handled.  I have the technology...

3)  yesterday while finishing the replacement recruiter system, I stumbled across a large pocket of Diplomacy simple_trigger timer activated things that never made it into Pho_II source, which then when Guspav's Phantasy Calradia source was merged onto Pho_II, meant pieces of Diplomacy were incomplete.  You've doubtless noticed but were polite enough not to ride me about it.
So anyway, I can't merge all of it without breaking save game, as too many simple triggers need changing, and adding even ONE new one breaks savegame, utterly. But here is what I COULD add, due to a little fault tolerance built into the v150 base (as in, spare "emergency" simple triggers that were declared, but empty, in case they were needed later). The bad news is, I am out of spares, so changes will be more difficult... unless I was going to have a new, incompatible version anyway, (say 6-8 weeks) like say v160  with different target capabilities...

  the good news:
  a) Policy setting actually does something over time now.  s_t 143
  b) scout ai  - scouts know what they're doing, maybe        s_t 139
  c) patrol ai - patrols now can handle being sent to villages to patrol closely there  s_t 138
  d) constable training - finally happens!    s_t 137
  e) process messengers s_t 136
  f)  process gift caravans  s_t 135
  g) recruiter kit s_t 134      this is why recruiters would spawn and never move.

  Specifically there was an updater for exiled lords that is NOT able to be implemented, plus I'm not sure if exiled lords cause issues as they are re-introduced.  Its better not to add everything the last few minutes before a patch anyway. Let's start small...

  probably the simple trigger numbers quoted are 2 less than actual red number in a dump, as what I have noted as #133 certainly was shown as #135 in a error message.  But OK, happy to have them anyway.

  NEARLY ALL of group 3's changes will NOT work until 24 hours have passed in game!  This is because the blank timers I took over were set at 24 hours till next fire but most of these are 1 hour timers (and patrol and scout are 0.2 hour timers; I may need to adjust those to be less often for performance).

  -----
  If I ever feel like I ran out of things to do I'm sure Guspav wouldn't object to me solving the 28 lvl 3 compiler warnings seen by build_module.bat or the 19 warnings issued by W.R.E.C.K. either, but there really are plenty of things that need doing.  I should clean up orphan villages but I don't want to make heavy changes 2 hours before patch.  This is enough.

Testing:
    notice the constable upgrading, and the corrected labels for delivered troops.
r6-3f.jpg

    notice the depleted numbers available for 2 of the towns.  This is taken from a different destination city (Ahmarad)  than the images I showed in the previous post; the depleted available recruits are because another town has recruits in the pipeline, and until THEY clear other towns can only get 1/3rd as many troops from that source town.  This is mainly to reduce taking one town, migrating it to a particular faction (in one case "Mageocracy" and in another case "Blazing Hand"), and then trying to flood all your owned towns with large numbers of special troops.  The normal flow returns when the pipeline is empty of troops originating FROM that town.  In this case neither origin town showing depletion is migrated, but there ARE orders from Narra and Shin
eV5fW.jpg


Statusuploaded to NexusMods.
              https://www.nexusmods.com/mbwarband/mods/6194/?tab=files&issue_id=373519
              just got approved.  It's live
 
Status
Not open for further replies.
Back
Top Bottom