Charon117

Ok, then here is the video nobody asked for :D.

Hm, I also had problems with the shield in vanilla. Dont know if that is related. The developers also hinted at an alternative way to disable the golem shield in conjunction with the stone it is throwing, but I never figured out how. For the latest EE,SS.C run I prepared an AWEO - 33% for the char, but it didnt make any difference. Would a video help you ?

About the GarColossus, I believe its primarily the shield which is able to soak up too much damage. Without dedicated bypassing attacks I do zero damage on the first difficulty. Also without having access to arcane reliqs his ranged attacks are devastating. I also believe the GarColossus to be too difficult in relation to the other bosses, making it a hard road block before you can access the rest of the game. The lethality level for bosses ranked from most to least dangerous: Mist Fog Stone golem Dragon Crystal Scorpion Octopus Guardian Bionic Flyer Demon For being the first boss to encounter he is the second most difficult to take on, as 90% of attacks are useless, and his melee damage is insane, giving he has a chance for doublehit. Double Hit paired with fast AS makes every 2 attacks have a chance for a full hp kill, which means premptively hammering your space is a prerequesite. So the tankines of the stone golem boss has nothing to do with his physical stats, but with his shield, which might have a secret trick to work around ?

I get it, it doesnt apply until somebodies uses the ingame "Reset to Default" button, or has a fresh installation. This is the logfile if you merge a complete EE2.1 build with an unedited EE2.2 content: logfile.txt

Flix, could you look into the following problem. EE2.2, SS, C: The Fog Boss Monster has an ultra hard hitting attack, which appears to be DOT, which one shots even the sturdiest builds. L91, 11k hp, Ice Mitigiation 15% - 20% ?, 3934 Ice Armour (3x legendary ice reliqs) and the Boss still seems to apply 4000 dmg every 0.5 seconds over the course of 2.5 seconds. Which equals to 200% of the chars total hp. Now I dont really want to complain about the difficulty on a Challenge run, but even though I have a ranged character it makes the boss fight even more lame than usual, not even talking about melee chars. So I would like to see a change which doesnt one shots every build in the game. A few example would be to spread out the damage from 2.5 seconds to the normal 5 or 7.5 seconds. Or maybe something else unexpected is going on. To the relevant people, is there a way to make the hp bar update instant ?

Eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeh .. DUH ?! You lost me there. Give me a rundown of the process: If I change teleport to F9 where is it getting saved ? Also are multiposts allowed now ?

@Flix Can be deleted.

Ok, so .. I assume a lot of ? in a row equal :). I assume thats autocorrect, and is made to save server space ? Im crying, and a laughing crazily a bit. 3 points With finishing the scripts folder I fail to see additional text files to write a modmerge system for. If you have any text files which still require some attention than please wishlist them now. relations.txt is still unaccounted for because of this issue: http://darkmatters.org/forums/index.php?/topic/23827-s2gcp16ee21ssc-esp-the-beginning-and-the-modmerging-system/&do=findComment&comment=7014812 Please take another look at the issue and suggest a solution. soundcluster.txt is still unaccounted for because of this issue: http://darkmatters.org/forums/index.php?/topic/23827-s2gcp16ee21ssc-esp-the-beginning-and-the-modmerging-system/&do=findComment&comment=7014812 Please take another look at this issue and suggest a solution. @Flix @dimitrius154 Phew. It .. works, and is nice. I suggest to implement this as for everybody else too. Or just wait until I release it as a standalone mod. Does it force change the options.txt permanently with every start ? Or is it a temporary change for one session ? Im kinda confused because if optionsDefault.txt would be working nobody would be able to change keys anyway ? Since autocorrect altered my original question, I still have no clue about roadmap.txt. Competent insight is welcome. Ill promise I will merge this post with the one above as soon as I get to it. Have a good start of the week :).

@Flix Did you ... edit my post ?

options_readme.txt: Does it actually do anything. And if yes, what ? optionsDefault.txt: I would make it modmergeable, unless you speak up now. roadmap.txt: :):):) startPos.txt: is empty. Almost. I doubt anybody misses the letters. @dimitrius154 @Flix

As far as the .diff states it, this files controls what files are even getting taken for the current gamesession. It first seems to load up general textures. Then shared data between server and client, which dont necessarily have to be in the "shared" folder. Then client only stuff. Can I deactivate the splash screen there ? And lastly server only stuff. This file looks so highly sensitive that I suggest it to be copyable only. Is everything working in the file as expected ? There doesnt seem to be an constructor for the behaviour above, which begs the question. If new behaviour is inserted should it be below or above the constructor ? Due to the assumed sensibility I suggest making it copyable only. Check that out. The first scaling = actualy has a tab before the =, but all others have one white space. Congratulations.

There is no chance that somebody would want to have additional modmerging functionality to add entries ? autoexec.txt: What kind of modmerge functionality is wished for here ? behaviour.txt: "name =" is the identifier ? There is also a cute local newBehaviour = {} in the middle Somebody explain genMipMapInfo.txt to me, and what kind of modmerge functionality it would need. heightmap.txt: "mgr.defineScaling( NUMBER," is the identifier ? landscape.txt: Similarily to weather.txt this seems to simply spawn a landscape mask for coordinates. Should we take the "mgr.defineLandscape(SEC POS," as unique identifier ? @dimitrius154 @Flix

Welcome to itemtype.txt, ill just quickly merge 332k lines. Or not. Maybe take a seat until its ready. Coffee ? 264 pre compiler optimised seconds later ... Still not bad for a file which takes up 30 percent space of all EE text files. Optimised time is 97 seconds. For 10 mb it could be better, could be worse. material.txt: "mgr.createMaterial(NUMBER;" is the unique identifier ? spells.txt: "mgr.defineSpell( "NAME"," and "mgr.addTokenBonus( {"NAME"," are the unique identifiers ? staticinfo.txt is empty ? typification.txt: "mgr.createTypification(NUMBER," is the unique identifier ? @Flix @dimitrius154

books.txt: I assume BOOK_TYPE + NUMBER is the unique identifier. Btw I always tried to read that "BOOK_ORCISH_INTRODUCTION" thing, but never could. Does it even exist ? creatureinfo.txt: I assume "type =" to be the unique identifier defines.txt is getting used ? iteminfo.txt: "type =" is the unique identifier ? itemtype.txt: "mgr.typeCreate(NUMBER," is the unique identifier ? @Flix @dimitrius154 Thx for the help.

Which means we are missing a useable identifier. Copy-able only it is.

Im sorry, I still have to come back to the total_density issue. Does anybody know if how total_density works ? If total density only checks the total number of creatures in sector pos + layermap_id we can make a triple identifier build. But if it counts creatures inserted from previous layermap_ids its a different story. Edit: Can I assume "Pfad" in waypoints to be the unique identifier ? It seems to be except for "Pfad_Wache2_Gabelung", which doesnt contain data. Edit2: I assume "dbid =" to be the unique identifier for weaponpool.txt. No dublicates exist. Edit3: weather.txt seems to be written in a similar fashion than spawn.txt. A simply fashion of "spawning" weather. In what kind of way would it be desired to get merged, if at all ? Edit4: worldobjecthints.txt is empty. Edit5: "mgr.addWorldObject( NUMBER," seems to be the identifier for worldobjects.txt, but there are a few dublicates: @Refresh @Flix @dimitrius154.

Yes, thats a very frustrating limitation. Every time somebody wants to upload a fullscreen image they have to compress it with an image processing software. I always thought this was to save space on the servers. But thats not the only drawbacks the forum has. Just one of them.

No, wait for me. I need to finish the modmerging system first.

The drop.txt has a single entry that has 2 numbers, {"SUBFAM_LIFE_DRAGON",1544,4123}, all other entries have 1 number only. Is that intended ?

If people havent figured it out yet, the last number in a position is most likely the azimuth. Following example strenghtens this claim:

Oh. oh. I thought we had .. a little more than that. I get the picture now. Because the software should be 110% of the best quality. If some files differ in their behaviour reverse searching wouldnt cover everything. So the base files will be searched for backwards now, to guarantee modding content to be actual taken by the game, even if the file has dublicated entries. Merging files will still be read from the top, so the last entry will take precedence as well. Both dublicated entries in base and merge files shouldnt happen, but if they do, we got em covered. The ideal state would be clean files for everybody, but we need a medium to get us there, so the modmerge system will be able to work with clean and dirty files alike. It especially covers situations like: An yet unknown file has dublicates, which no special rule has been written for. Merging dirty files with dirty files, like upgrading from EE 2.1 to EE2.2. And other situations.

I am currently having an issue. It is the "What if any file has dublicates ?" problem. While I can write extensive cleanup programs for individual files it doesnt solve the question whether some random base files might have dublicates. This obviously should be avoided by base file authors, but again, doesnt answer the question. Sooooooooooooooooooooooooooo ... 1) I could write a generic cleanup program which checks the base file for dublicates and cleans them up in one fellow swoop. The big downside of this is that such cleanup programs are very expensive in resources, and would take 80% of the the total time to search for dublicates which might not even exist. I mean after you´ve gone once over all your files they definitely dont exist anymore. 2) Solution number 2 would be to outsource such a cleanup program which users would have to use before using the modmergse software. Im very averse to this solution, as relying on the user is generally not a good strategy. Software is there to adapt to humans, not to make humans adapt to software. 3) Number 3 would be to cleanup the community database. But that again would not really answer the question when somebody digs up a dirty file from somewhere else. 4) Dimitrius mentioned that if dublicates exist the last entry takes precedence. If that would be the case for all txt files than the software could simply modifiy the last entry it finds, instead of the first one. This would be a very elegant solution as it would take up zero additional resources, and still reliably produce the proper modmerge effect. So I would ask Dimitrius to look into the Sacred 2 source code and make some reliable checks whether or not somebody without a reasonable doubt could say that the last dublicated entry takes predence over all other ones for at least 99% of files.

There are no such dublicates, dublicates only refer to sector pos + layermap_id data.

Im surprised by this answer, as I thought it would be more important to preserve the current data format. Can you tell me in what way you want to merge the dublicated entries ?