Bovine M2TW Checker

[bovi]

Great work!
your tool now runs through most parts and found alot of stuff - other tool didn´t

only eda.log and edb.log seems to be blank

Thanks! Good! Not good!

But seriously, blank edct and eda means it has crashed somewhere else. I'll take a look at your logs, hoping there's a screenshot among them.

yes, for the most parts. but i altered by hand because of the officer-torch-bug after AnthoniusII remembered me, that graculs can´t handle it

Manual alterations can quickly accumulate formatting mistakes. I hope this tool can assist you in keeping it clean at least, but it does not check everything under the sun and is prone to crash if you don't follow the proper format. I imagine it is a lot more finicky than the game itself.

My bet: the torch lines (regular and officers) are fixed entries and I am guessing it's vanilla series is fixed as a cross check in the tool - meaning the tool will run into a possible fail situation if you have edited that line.

My tool does not check anything at all about torch values. I don't even know what they mean, so I just discard their segments when reading the file. Although, if the number of torch segments is wrong, it will seriously freak out and crash as it will not find the start of the next entry where it expects it.

[bovi]

Here are (hopefully) all the files you requested, they are all from the same mod (RotK).

Then I will download the entire mod I think. Easier to know I've got all I need then - unless your current files are too different from the released version.

[QuintusSertorius]

This is an excellent tool that has proven itself invaluable in the development of EBII. Our next release has been purged of so many potential bugs through the use of it.

[bovi]

Thanks for telling me the usual way of handling animations. Right now I think it may serve its purpose to leave it as is, now that the list of vanilla animations is hopefully correct. The parameter name is a bit of a misnomer now though, so I'll change that to better reflect that it means vanilla animations are expected to be in the mod folder if you leave it to YES.

I'll add a list of thread responses to better track which ones still require attention - I'm getting a bit worried I'm forgetting things now.

[Gigantus]

Then I will download the entire mod I think. Easier to know I've got all I need then - unless your current files are too different from the released version.

Should be the updated 2.1 version - but honestly I am not sure if there are dev files in between as 2.2 is in full swing. The animation files are not unpacked in the installation

[bovi]

2.1? That post says 2.0 and is the one I downloaded, is it the wrong one?

I now listed the action points I see remaining in the end of the original post. Let me know if I overlooked anything.

Edit: sorry, I didn't look down to post #2 with 2.1. All fine!

[Gigantus]

Liked my file check at the end of installation?

Running the tool on your 1648 model DB, I see that you include the "1.12" not only on your first mount entry, but all of your horse mounts. I have special processing for the very first model db entry, but not for the subsequent ones, therefore it crashes on the second entry mount_heavy_horse.


In this position it expects a marker saying how many mesh files the entry has, rather than 1.12. What does that 1.12 do? I thought it was only some required fluff that you had to add to the first entry unless you had some blank dummy thing as the first entry like Tartaros had in his?

The special processing for the first entry expects three meaningless segments before the mesh count and then another two meaningless segments before those mesh references actually come, discarding those 5 segments around it.

If 1.12 is a meaningful value and you really need to have it there (we don't have it for any other mount entry than the first in EB2!), what other values could there be in its place? Would it be safe to look at the segment, and if it's 1.12 just discard it and look for the mesh count in the next segment instead?

No idea about the reason behind the 1.12 entry, a lot of custom modeldb's don't have it to that extent or at all - only that the vanilla modeldb has those seven instances of it. And a good number of older mods (like 1648) didn't use tools like gracul's checker on their modeldb to reformat the file.
I can only speculate that it was supposed to indicate a vanilla horse mount, camel and elephant do not have that entry. And that it was abandoned right after the start before they did the same (or similar) for camel and elephant.

Bare Geomod run
Looks fine, shows 'xyz is missing and therefore no check' but it shows a missing cross check involving descr_events (examples from a long script.error.log):

EDB 4358: Event counter "gunpowder_discovered" is checked but never set.
Historic event "earthquake_in_silicia" is never triggered in the script.

It's here in descr_events:

event    historic    gunpowder_discovered
date    210 220
movie    event/gunpowder_invented.bik

event    earthquake    earthquake_in_silicia
date    188
position    232, 76

The file is a left over from RTW and is typically not used (blank) in a mod as the events all trigger at the start of the first summer turn (which is fine for the vanilla campaign).

A question regarding these two sounds.error.log entries

Cannot find accepted accent prefixes for accent english
DSSV stratmap 471: ERROR - Referenced diplomacy sound name "english_diplomat_death_cry_1" is missing in diplomacy_speech.txt.

What is the target and the source of these two queries, eg what is being checked? I added the vanilla diplomat_speech.txt file to the data\text folder.

[bovi]

Then I think I will just check for 1.12 and discard that bit if it appears before the mesh count.

The only thing I read from descr_events.txt is event historic. I'll expand to also read event earthquake. Are there other event types I should know about when going about it? But it should have picked up on that gunpowder event. I'll see what is happening there in the bare geomod.

---------

The checker's reference to "DSSV stratmap" means export_descr_sounds_stratmap_voice.txt. It crosschecks against the file structure under sounds/Voice/Human/Localized/Camp_Map, to ensure you are not referencing nonexistent files, and that the files present there are in use. There is a similar crosscheck between export_descr_sounds_soldier_voice.txt and sounds/Voice/Human/Localized/Battle_Map, noted in reports as "DSSV soldier".

We had a problem with blank diplomacy messages turning up because we did not have all the required text in diplomacy_speech.txt. The tool therefore checks that all sounds related to the "diplomat" agent is also present in that file. Not all diplomat sounds are shown in a negotiation speech bubble, however, therefore there is a configuration file "unquotedDiplomacySounds.txt" which tells the checker which of the diplomat sounds it should not expect such text for. I forgot to add this bit to the setup for the checker.

Additionally, it warns against possible copy-paste errors in the form of sound files having a name inconsistent with the culture (accent), such as steppe agents/units using a sound file starting with "greek". You need to

6. Configure the correct accent prefixes for battle sounds by culture in acceptedAccentPrefixes. The supplied file works for EBII's cultures and will not suit your mod.

I notice that I there only mention battle sounds, but it also goes for stratmap sounds. I'll have to amend.

[Gigantus]

Well, 'death_cry' is not going to show up in the diplomacy text file at all - that's the sound the guy makes when you kill him on the strat map

All these sound (vocal line) references are strat map only:

                  type Diplomat
                           vocal Death_Cry
                           vocal Entering_City
                           vocal Exiting_City
                           vocal Joins_Army
                           vocal Leaves_Army
                           vocal Movement_Completed
                           vocal Movement_Diplomacy
                           vocal Movement_Disembark
                           vocal Movement_Embark
                           vocal Movement_Ordered
                           vocal Movement_Ordered_Illegal_Move
                           vocal Selected
                           vocal Select_Ally
                           vocal Select_Neutral
                           vocal Select_Enemy

All three hundred odd error messages I have refer to these entries. The relevant sound entries for the diplomacy screen are after that, eg Offer_Countered_Open_Positive and haven't been triggered once.
Edit: I see in the unquotedDiplomacySound file I will need to generate multiple chunks of those exceptions with the prefix of each valid accent. That sucks, pardon my French, luckily I am pretty handy with the script replicator. Can you add all the entries for the vanilla accents? Or have a simple routine that looks up valid accents in descr_sounds_accents and then accounts for those exceptions?

There are no entries for the (vanilla) accents in acceptedAcccentPrefixes - maybe change the wording of the message? 'Caution: Prefixes for the [accent name] accent have not been defined.'
Will dummy entries work otherwise?:

[english]
none

If so then maybe add all vanilla accents (see descr_sounds_accents in BG) like that into the file. This works basically from the same source (valid accents from descr_sounds_accents) as the diplomacy text check.

As to events in descr_events: all of them generate (increase by 1 each triggering) event counters. A limited number have text in historic_events. I did a extensive, related guide for that. The entries in descr_events are basically the abbreviated form of it. You may want to incorporate the whole stuff into your script checker anyhow, see the detailed description which events can have messages.

[Gigantus]

All BG related

Right, after I added all those accent entries to unquotedDiplomacySound I am left with three massive sections of errors:

1. the missing accent prefix messages
2. Stuff from DSA:

DSA advice 421: Warning - Referenced sound name "prologue_prisoners_text_01_text1" has a prefix unlike its text (ui_scroll_prisoners_advice_text_01_text1). Intended?
DSA advice 441: Warning - Referenced sound name "ui_panel_settlement_auto_managed_advice_text_01_text1" has a prefix unlike its text (help_ui_panel_settlement_auto_managed_advice_text_01_text1). Intended?

3. And a lot of princess diplomacy:

Diplomacy text "east_european_princess_demeanour_declare_annoyed_1" is not used by any diplomat accent.

The 60 lines from the DSA errors have a 'help' prefix with the exception of two lines I think. Where do I add that?
The princess diplo issue is because of chauvinism - EBII doesn't use the girls. You basically have to duplicate your diplomat routine and replace diplomat with princess and then add another huge chunk to unquotedDiplomacySounds.

         accent East_European
                  type Princess
                           vocal Demeanour_Declare_Annoyed
                                    event delay 0
                                             folder data/sounds/Voice/Human/Localized/Camp_Map
                                                  East_European_Princess_Demeanour_Declare_Annoyed_1.mp3
                                                  East_European_Princess_Demeanour_Declare_Annoyed_2.mp3
                                                  East_European_Princess_Demeanour_Declare_Annoyed_3.mp3
                                    end

Here is the unquotedDiplomacySounds file for vanilla inclusive for princesses

Added vanilla prefix entries to acceptedAccentPrefixes:

[english]english


[scottish]
scottish


[french]
french


[german]
german


[mediterranean]
mediterranean


[east_european]
east_european


[arabic]
arabic


[mongolian]
mongolian

No more accent prefix messages , only princess diplomacy left and the advice mismatch.

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

Trying to make sense of this error message in EDCT: Trait "heresyimmunitynoncatholic" is hidden but has non-hidden description.

Trait HeresyImmunityNonCatholic
    Characters priest
    Hidden


    Level HeresyImmunityNC
        Description HeresyImmunityNC_desc
        EffectsDescription HeresyImmunityNC_effects_desc
        Threshold  1 


        Effect HeresyImmunity  1

{HeresyImmunityNC}HeresyImmunityNC
{HeresyImmunityNC_desc}DO NOT TRANSLATE - This man has an adultress ancillary
{HeresyImmunityNC_effects_desc}No Effects


I mean it absolutely doesn't matter what entry is in export_vnvs - as long as the curved bracket entries correspond. There is no 'hidden description'. I am guessing you are referring to the use of place holder\dummy text entry instead of a full entry:

Trait: Description Hidden_desc
EffectsDescription Hidden_effects_desc

export_vnvs: {Hidden_desc}hidden
{Hidden_effects_desc}hidden

Those follow exactly the same rule as the 'regular' descriptions, only that they are used for multiple (hidden) traits. That multiple use gets then picked up by some tools (withwar's)

[bovi]

I can make the unquoted diplomacy sounds contain patterns rather than specific full names. This should make it less of a hassle to configure it, possibly even making ours work for mostly every mod.

Adding dummy accent prefixes for the vanilla accents like that will not work. You will get rid of that one message then, but none of the vanilla sound files have names that start with "none", so all of them will still be flagged as suspicious. I'll see if I can add the vanilla accent prefixes, yes.

Thanks for the list of event types, I have incorporated them and fixed the reading of descr_events. It used to only add the listed events to USED events, not DEFINED events. Not uploaded a new version yet, but the bare geomod now only has these reports below for the script. I searched through the mod files and found nowhere those particular events are set, so I think that will work as it should now.

Spoiler Alert, click show to read: 

CS 41: Event counter "mongols_invasion" is checked but never set.
CS 57: Event counter "mongols_invasion" is checked but never set.
CS 173: Event counter "mongols_invasion" is checked but never set.
CS 302: Event counter "mongols_invasion" is checked but never set.
CS 443: Event counter "timurids_invasion" is checked but never set.
CS 459: Event counter "timurids_invasion" is checked but never set.
CS 602: Event counter "timurids_invasion" is checked but never set.
Historic event "first_printed_bible" is never triggered in the script.
Historic event "first_rudder" is never triggered in the script.
Historic event "the_mongols_emerge" is never triggered in the script.
Historic event "the_timurids_emerge" is never triggered in the script.

[bovi]

Didn't see your reply before the previous post.

The 60 lines from the DSA errors have a 'help' prefix with the exception of two lines I think. Where do I add that?

I'm not sure you can. I'll investigate. At worst, I can make a flag to allow turning off that test.

The princess diplo issue is because of chauvinism - EBII doesn't use the girls.

Here is the unquotedDiplomacySounds file for vanilla inclusive for princesses

Thanks for the vanilla list - I'll still see if I can make them patterns though.

You're exactly right about our chauvinistic overlooking the princesses. I'll fix that.

[bovi]

Just wanted to say thanks a lot for the work you guys are putting into uncovering the weaknesses in the checker. I'm afraid I may not be able to work through them very quickly, but I'll make sure to get around to all of it sooner or later.

[Gigantus]

Note: I updated the unquotedDiplomacySounds file because I had small error in my template, in case you get a 'file missing' message. I also added to post #50 long after you had posted.
I'll check if I find some connection for those remaining errors, the emergent stuff sounds familiar.

Edit:

CS 41: Event counter "mongols_invasion" is checked but never set.


Here the add_events sequence is what sets the event counter:

        add_events
            event    counter    mongols_invasion
            date    4 8
        end_add_events

The moment the event fires the event counter will be incremented (can be multiple triggering), meaning the first time the counter will get set from zero to one, next to two, etc. See next section as well.

----

Historic event "the_mongols_emerge" is never triggered in the script.
Historic event "the_timurids_emerge" is never triggered in the script.


The format has some hard coded stuff:

            add_events
                ;;; SCOUTING INVASION
                event      emergent_faction    mongols
                date    0
                ;                 region     Tbilisi_Province
                region     Yerevan_Province
                region    Bulgar_Province
                region    Sarkel_Province
                region     Baghdad_Province
                movie    event/mongols_invade.bik
                ;ADD MONGOL INVASION 1
                event    counter    mongols_invasion
                date    2
            end_add_events

----

Historic event "first_printed_bible" is never triggered in the script.
Historic event "first_rudder" is never triggered in the script.

That's truly nowhere.

[bovi]

Not all problems have been addressed, but I've done some work on this and uploaded a new version.

Changes made:

The checker now survives if the meaningless segment before the mesh count is 1.12 as well as 1, whether in the first entry or not.

The checker now survives units having no sprite definition, or no weapon definition (wardogs).

The checker now survives duplicate spaces, at least more places in the model DB than previously. These are not correct, but better to not die.

Reading events in descr_events and campaign script now includes all kinds of events, such as earthquakes and storms. Previously only "historic" events.

Diplomacy sounds now include princesses.

Diplomat / princess sounds not meant for appearing in speech bubble are now defined by pattern in unquotedDiplomacySounds.txt, rather than having to make specific entries for every single file.

[bovi]

But seriously, blank edct and eda means it has crashed somewhere else. I'll take a look at your logs, hoping there's a screenshot among them.

No, there wasn't a screenshot. Tartaros, to investigate I need either your mod in its entirety, or a screenshot of the crash to know where to look next. Please use the version I just uploaded to make the screenshot, as that will point me to the correct code location.

[Gigantus]

Can you please add this section to the acceptedAccentPrefixes file: It's the vanilla settings:

[english]english


[scottish]
scottish


[french]
french


[german]
german


[mediterranean]
mediterranean


[east_european]
east_european


[arabic]
arabic


[mongolian]
mongolian

If memory serves these two event counters (BG script) disable the related hidden resource function, hard coded, see pic :

CS 758: Event counter "disable_no_brigands" is set but never checked.
CS 760: Event counter "disable_no_pirates" is set but never checked.

Other then advice text mismatch and the two 'true' historic_events BG is clear of errors otherwise. (No false positives)

[Gigantus]

New\missed stuff from RotK - modelDB log

Skeleton "strat_merchant" is never used by model DB entries.
Skeleton "strat_priest" is never used by model DB entries.
Skeleton "princess" is never used by model DB entries.
Skeleton "witch" is never used by model DB entries.

I was wondering why the other strat skeletons (spy, assassin etc) were not flagged when I remembered that the above strat character models are not used by EBII, please add to your exception list.


The sound log is miles long as is the above modelDB log - will need some time to go through that in detail.

[bovi]

Done - although not made into a downloadable version yet. You can add them to acceptedUnusedSkeletons.txt yourself in the meantime. Full contents:

strat_named_with_army
strat_spy
strat_assassin
strat_diplomat
strat_navy
strat_merchant
strat_priest
princess
witch

[bovi]

I have made a new version. There are a number of changes - most notably fixing some of the stuff reported here and adding strat map models to the check subjects. I need to update the setup instructions as the location of configuration has changed.

Configuration moved to own file rather than the runcheck.bat file.
The tool no longer dies if an event invocation does not include a specific event name.
New report: Duplicate declared counters.
The tool can now be configured to suppress reports about hidden traits having specific descriptions.
Added EB2 specific unquoted diplomacy sound filename patterns.
New report: faction symbols and other file references crosscheck against file structure.
New report: crosscheck of descr_model_strat file references, strat models' skeleton usage, and descr_character usage of strat models.
Silver surfer reports now correctly take mercenary attribute into account for unit ownership, previously only entries matching EB2's way of handling mercenaries would pass.
Usage of battle models now also take into consideration battle_model references in descr_strat and campaign script.
Model DB error log and java class renamed as it no longer concerns only model DB related reports.
Configuration of accepted files not being referenced is now filename patterns rather than specific filenames.
Vanilla accent prefixes for sound filenames added to configuration.

By the way, as strat models are now read as well as battle models, accepted unused skeletons is currently empty, contrary to what I said in the previous post.