Bovine M2TW Checker

[bovi]

'sundry error' still complains about the 'non 1' value in EDB event counters, guess that will stay?

It will stay until I make a change on it, yes! I am somewhat confident I'll need to program according to that tests on event counters are >= the given value.

-- "Translation default includes name text for name code xyz, but this name code is not used by any faction for neither surname, male first name nor female first name." Does it mean that names listed in text\names are not listed in descr_names? If so then it creates a ton of unnecessary messages (4K+ in BG) as it has no impact on the game.

Correct. However, I disagree the test is unnecessary. There are a lot of other checks that also list stuff that is unused - because being unused can be a mistake rather than an intended omission. I cannot think of any reason why you would intentionally add text for names that are not in use, but most likely these cases are simply leftovers that should be removed.

-- "Faction turks defines surname of rum in descr_names.txt. Translation default does not provide any display text for this name."
{of_Rum}of Rum - it's an oddity of the game that substitutes spaces in descr_names listing with an underscore in text\names, totally works, eg false positive
-- "descr_strat.txt 796: Map character uses surname de, which is not defined for faction france." Refers to this: "character Guillemot de Lyon". A domino effect from the preceding issue, tons of them listed given the sheer number of those entries in descr_names.

I need to fix this, thanks.

'file references errors': the lookup for unit pics might need 'AcceptedMissingUnitPics' lists to avoid miles of error messages, get them here: Attachment 368150

Thanks! I suppose those would rather be "vanillaUnitPics" than accepted missing ones though.

Looks like the text look up for historic events seems to work, it didn't pick up the EDB event counter entry (other then using 3) from the earlier testing - must have really misunderstood what you tried to explain in the part I quoted in the previous post.

I believe so, yes. Not sure how I can rephrase sucessfully, but let's try.

When checking the current value, it is only flagged as a historic event counter if that reference ends in "_accepted" or "_declined". When setting the value of event counters, it gets flagged as a historic event counter depending on which command is used to set it.

Regular, non-historic event counters are not checked for text being present.

[Gigantus]

it is only flagged as a historic event counter if that reference ends in "_accepted" or "_declined".

I think that's what threw me: the accepted\declined entries are regular event counters created by the yes\no historic_event, you actually have to create a line with the historic_event command if you want to make use of that wording.

historic_event EC_REFORMATION true factions { f_pontos, f_hayasdan, } ; Invitation to invest in venture

   while I_EventCounter EC_REFORMATION_accepted = 0 ;;; while-loop to pause the script until a decision is made
      and I_EventCounter EC_REFORMATION_declined = 0
   end_while

      if I_EventCounter EC_REFORMATION_accepted = 1  ; if the player accepts

Correct. However, I disagree the test is unnecessary. There are a lot of other checks that also list stuff that is unused - because being unused can be a mistake rather than an intended omission. I cannot think of any reason why you would intentionally add text for names that are not in use, but most likely these cases are simply leftovers that should be removed.

You normally wouldn't add to text\names but rather remove from descr_names - it's reasonably common as no one bothers to remove the relevant counterparts from text\names, especially since the descr_names entry may exist for multiple factions but only gets removed for one. I get roughly 4000 entries from vanilla (BG) according to the stratmap log. Although I expect about half of it points back to the double barrel\underscore name issue:

"Translation default includes name text for name code bin_haji, but this name code..."
"Faction moors defines surname bin haji in descr_names.txt. Translation default does not ..."

[diealot]

Hi Bovi I am trying to run the Bovine Checker on Insularis Draco Latest build but according to Urbgen the Mod Developer most of the errors in the edct.error.log and the unit.error.log are false report and is non existent. Could you run the tool on the latest build of the Mod and figure this thing out and explain it to us. Thank you very much Bovi.
https://drive.google.com/file/d/1Qnh...usp=share_link

[Gigantus]

The most common error in the edct log relates to hidden traits and their descriptions.
Similar entries can be found in the unit log, eg text for units not yet implemented.

You would need to be familiar with the mod to know which are caused by shortcuts taken\incomplete implementation and which are genuine - the tool cannot do that for you.

[Belovèse]

Soooooooo... My apologies for being so late on this, but I finally did some tests on this edb event_counter condition, and it seems Gigantus hunch was right: it is in fact "event_counter name >= value"

I did some tests in a customized bare geomod to be in a clean state, you can see the results here in the readme and event download the sandbox and try for yourselves: https://gitlab.com/Belovese/modder_s...ref_type=heads

to summarise I had this code in EDB for wooden_walls and results of the different values I gave the event testCounter gave the expected results:

; should be available only for value 2
recruit_pool "Knights Hospitaller"  1   0.7   6  2  requires factions { all, } and event_counter testCounter 2 and not event_counter testCounter 3	
; should be available for values 1 to 4 
recruitment_slots 2 requires factions { all, } and event_counter testCounter 1 and not event_counter testCounter 5

[bovi]

Thanks for getting back to this. I was actually going to go ahead without your verification at some point, but did not find the time. Anyway, now it's only the time standing in the way. Apparently a year has whooshed by already!

[bovi]

The most common error in the edct log relates to hidden traits and their descriptions.
Similar entries can be found in the unit log, eg text for units not yet implemented.

You would need to be familiar with the mod to know which are caused by shortcuts taken\incomplete implementation and which are genuine - the tool cannot do that for you.

Do note that there is a switch you can set in the configuration file to remove the checks for hidden traits and their descriptions:
HIDDEN_TRAIT_DESCRIPTIONS_SUPPRESSED=YES

[Belovèse]

Nice to see you Bovi, time has been running fast indeed!

I'm wondering: could you easily add more flags in the config file to fine tune what the checker does? Like turning on/off the check for missing ui cards, etc...
Or maybe filtering by severity (error, warn, info ..) But I guess it is longer to do if you didn't bake it in at the start?

Also would it be easy or not to integrate custom campaigns ? I'm working on a mod that has 3 of those, it is a bit tedious to check them [emoji28]

But all those are just ideas, no pressure, I'm already very grateful for your checker as it is today!

[bovi]

It's not hard to implement such switches, it's basically to put in a if-disabled-don't-do-your-thing in the code that performs the check in question. I feel however that pretty much all of the checks point out stuff that should be addressed, except the one that already has a switch (which probably only needs addressing in EB2). Missing UI cards are among those I feel should be addressed, really. If any were to be superfluous, it would probably be those pointing out excess files, as those don't really cause any direct problems. They could still indicate missing references that was supposed to use those files.

I don't feel like going through all of them and adding finegrained control, but if you identify particular reports which are needlessly cluttering up, I can do that. I should probably also be able to make it so that you can create your own suppression filters in a config file - such as containing the text "Missing UI card". Each report typically follows a predictable pattern of text so should be able to be identified that way. Or along the severity aspect, that you can define your own severity level for given text patterns and the configured log level would suppress any below that severity.

As an aside, I was thinking of adding a UI to it at some point, allowing you to select specific reports and go into an editor for the file and line in question, then rerun the check to see it was resolved after editing. I thought better of it as I will probably never find that time.

------------

Could you elaborate on the custom campaigns thing? I'm not sure what you're lacking - would those be different sets of files inside a general mod then? I think you are most probably having a different problem with the custom campaigns , but if so, know that I have a set of configuration files myself to run on different mods, and you can too, you just need to point the tool to which configuration file to use. It's this line in runCheck.bat:

java -cp "target;%LIBRARIES%" main.BovineM2twCheckKt configuration.txt

configuration_1648.txt
configuration_bare_geomod.txt
configuration_eb2.txt
configuration_insdraco.txt
(...)

The tool will first look in its own folder for the file, and if it doesn't find it there, it will look in the config folder. This way, you won't risk losing your own configuration files when getting a new version of the tool.

[Belovèse]

Thanks for all the infos, indeed filters and levels are more a small QoL thing, it is easy for me to do more or less the same thing on my own so it's not worth you taking time for it if you don't feel like it

Would you share your config for InsularisDraco with me? This is one of the mods that has numerous custom campaign that I was talking about.

[bovi]

Um, sure. It's been quite some time since I ran it on that mod however, so the newer configuration options are not present, at least REMOVE_UNUSED_FILES and UNUSED_AGENT.

You definitely need to change the DATAFOLDER. Probably also STRAT to reflect the custom campaign in question. My guess would be that you'd want to have one configuration file with STRAT=world/maps/campaign/custom/520AD, another with STRAT=world/maps/campaign/custom/630AD, and more for any other custom campaigns that were not present in the version I downloaded ages ago.

;Set DATAFOLDER to the data folder for your mod, such as C:/Games/M2TW/mods/my_mod_name/data
;C:/Games/M2TW needs to be changed to wherever your M2TW is installed to.
;my_mod_name needs to be changed to whatever your mod folder is called.
DATAFOLDER=../../../../M2TWOtherMods/mods/Insularis-draco/InsularisDraco/data

;set STRAT to the folder containing your campaign script and descr_events.
;Probably this is the default imperial_campaign, but possibly base.
STRAT=world/maps/campaign/imperial_campaign
;STRAT=world/maps/base

;The checker can be configured to require vanilla animations to be present in
;the mod's own animations folder. As this is the case for EB2, this is the
;default setting even if most mods can rely on vanilla animation files being
;available to the mod. Chances are that you want to uncomment the setting
;below for any other mod than EB2.
VANILLA_ANIMATIONS_REQUIRED=NO

;The checker expects all hidden traits to use descriptions "hidden_desc" and 
;"hidden_effects_desc". Any hidden traits with specific descriptions are
;considered to be intended to be shown to the player and assigned as hidden by
;mistake. If your mod wishes to keep trait specific text even for hidden
;traits, uncomment this line below to suppress those reports.
HIDDEN_TRAIT_DESCRIPTIONS_SUPPRESSED=YES

;The checker tries to match all sound file names with the accent it is used for.
;Any file names not matching the defined accent prefixes will be reported. If
;you are not doing voice modding or otherwise don't want these reports, you can
;suppress them by uncommenting the line below.
ACCENT_PREFIXES_SUPPRESSED=YES

;The checker tries to match all referenced sound files with the file system.
;If all your sound files are packed rather than available as individual files,
;you want to suppress reports about missing files by uncommenting the line below.
SOUNDFILE_INTEGRITY_SUPPRESSED=YES

[Belovèse]

Thanks a lot. Swapping the strat campaign paths one by one was already what I was doing, but would it be possible to have the tool automatically check all campaign folders, like it does for the battles? In the current way it generates some false positive because of the unchecked campaigns

[bovi]

But isn't each campaign its own set of files for different "worlds" where the same units and factions apply? There's no intermingling between what is in imperial_campaign and 630AD, is there?

But perhaps you are referring to the unused files having false positives when not all of them are considered in unison, when I think about it. That would be a good point, I may need to do that. Only, I guess I'd have to change how the information is stored to know which of the descr_strat's line 345 it's talking about, as an example.

[Gigantus]

This would be an example (from BG) how the path looks for custom (provincial) campaigns:

- bare_geomod\data\world\maps\campaign\custom\india
- bare_geomod\data\world\maps\campaign\custom\japan

Basically a standard imperial_campaign folder plus the maps\base directory content for a different map.
Note: in this directory the 'description' file actually has a function, unlike in the imperial_campaign folder file, eg it's the info text presented when selecting a campaign from the available list.

[Belovèse]

From memory (not on my computer), the main annoyance with false positives caused by the event counters that are used or declared in scripts

[Gigantus]

While working on the latest release of Rise of three Kingdoms the checker threw up several issues:

The following file hasn't been included in the 'vanillafiles' exception list file (my bad):

sounds/SFX/individual/torch.wav

------------------
Character label related errors (script.error.log), example:

CS 2096: Character name reference to "yuanshu1", which is not a starting character or scripted character.

This is the line referenced:

console_command give_trait "YuanShu1" HuangDi-scotland1100 2

And this is the character's line in descr_strat:

character    J-1901-YuanShu, named character, male, leader, age 43, x 142, y 114, portrait A059, label YuanShu1

This line in the script (2095) was not flagged so I am not sure if the issue is connected to the quotation marks?

if I_CharacterExists YuanShu1

------------------
The custom portrait check (stratmap.error.log) does not seem to work correctly either:

This one is in a campaign script spawn:

File "ui/custom_portraits/a397/card_dead.tga" is not referenced in the Custom characters
File "ui/custom_portraits/a397/card_old.tga" is not referenced in the Custom characters
File "ui/custom_portraits/a397/card_young.tga" is not referenced in the Custom characters
---
character    J-3204-XuShu, named character, age 22 , x 106 , y 104, label XuShu_2, portrait A397, battle_model wenren

This one is in descr_strat:

File "ui/custom_portraits/a059/card_dead.tga" is not referenced in the Custom characters
File "ui/custom_portraits/a059/card_old.tga" is not referenced in the Custom characters
File "ui/custom_portraits/a059/card_young.tga" is not referenced in the Custom characters
---
character    J-1901-YuanShu, named character, male, leader, age 43, x 142, y 114, portrait A059, label YuanShu1

[bovi]

I expect that to be an issue with the checker not considering capitalization and the quotemarks, yes. Adding it to the pile.

[Gigantus]

More for the pile:

From sundry.error.log - tokens are valid

Unrecognized token in EDCT 35224: regionislocal. Typo in EDCT or fault in the checker?

Trigger G5detained
    WhenToTest CharacterTurnStart

    Condition I_EventCounter agent_upd = 2
    and RegionIsLocal
    and not CharacterIsLocal
    and AgentType = diplomat
    and not Trait Detained > 0

    Affects Detained  2  Chance  100

Unrecognized token in EDCT 20031: losingmoney. Typo in EDCT or fault in the checker?

Trigger MerchantSkills_add_1
 WhenToTest CharacterTurnEnd

 Condition IsFactionLeader
       and Trait WideTraitSet = 0
       and Trait GoodAdministrator > 0
       and not LosingMoney
       and Attribute Authority > 4
       and Attribute Piety < 5
       and Trait Epicurean < 2

 Affects MerchantSkills 1 Chance 15

[mekkermann]

Hi,

first of all thanks for your great tool it made modding and bugfixing so much easier.

Sadly I did run into an problem with the battle_models file yesterday where the checker would hang and never finished. I didn´t had this before and think it´s thanks to the new unit_models and skins we use here ( maybe the path? ) or we have an problem in the file which the Bovine checker can not digest. I have to bug the battle_models file so the checker notices an syntax error and therefore skips the lines below.

Can you please look up the files if you can make out an compability problem with the bovine checker?
This would help immensly since I cannot mod without the bovine checker.

Here is the link for the files, https://drive.google.com/file/d/1lbG...usp=drive_link

The mod is Roar of Conquest but internal beta version so I would give you a link to the version if it´s important for bugfixing on your side.

Thanks in advance.


Edit:

I found this program https://www.twcenter.net/forums/show...09-Unit-Editor

Which showed me no concerning errors in the battle_models file and also had the option to clear up the file from un-used entries.

As I did this the battle_models file was down to 777 entries from 1786 before and the bovine checker worked again.

So can it be that there is an limit of entries in the battle_models file the Bovine checker can digest?

[Gigantus]

I think it was rather an unknown way of listing the individual entries, eg line arrangement. The unit editor formats it's modelDB output in a way the checker has no problem with - in fact that's what I do if I have come across that particular issue with the checker.

Hopefully bovi can figure out what your issue was and include it in the next release.