Might and Magic 6, 7, 8 merge based on mm8 engine. [upd. 05.11.2023]

[raekuul]

Jump spell is very weak compared to vanilla games. I cannot jump into many places which could have been reached in vanilla mm6.

Jump is weaker in 7/8 vs 6 to begin with, are you sure you're not just experiencing 8's jump inside 6's map design?

[tomchen1989]

@Templayer

1. OK looks like Q is still there but "Ethric" and "Agar" are lost and become Power Lich I guess (though I haven't checked them in the game)

2. OK thank you. For now, MMMerge's localization is a little messy. I'll see what I can do, perhaps make a script to convert all translatable strings to .po file (GNU gettext) which will be easy to manage

3. cp1250 (aka Windows-1250) is commonly used as one-byte encoding for Czech language. I guess I can safely assume Czech version MM uses cp1250. Therefore you have to save txt files with cp1250/windows1250 encoding, and use .fnt files from Czech version MM instead of English, French version's .fnt files.

Compare:
Windows-1250 Character set (Czech and central EU langs)
Windows-1251 Character set (Russian and Cyrillic script-based)
Windows-1252 Character set (French, German and Western EU langs)

Hex code CA (=binary code 11001010) represents Ę in cp1250, К in cp1251, and Ê cp1252. Switch your .fnt files (more like a code to character shape mapping file) to read those different encodings.

Each letter from the 26 basic latin letters always has the same code in these encodings. Therefore, lucky for English, it can use any one-byte encoding. However it usually uses Western EU langs' cp1252.

Vitually any text editor can save or open a file with a specific encoding. You could use EditPlus as your text editor, it preserves CRLF and LF, while VS Code, today's most popular text editor, does not allow CRLF and LF to co-exist in a same file, forcibly converting one to the other, making it not able to edit MM's txts, which use LF as in-string line break and CRLF as table line break.

1 and 2-byte encoding support would be enough. I don't see Unicode support is useful, unless multiple languages such as English, Russian, Greek, Chinese, Korean have to largely co-exist in the game.

GrayFace's TxtEdit works very nicely except it can't select another encoding, it has English and Russian interface, it seems that users have to use either Windows SystemLocale encoding for default English interface, or cp1251 (Russian) encoding when the interface is set to Russian version. Interface language is one thing, character encoding is another, it should allow to change character encoding like any other text editor. Actually, if TxtEdit has settings to allow encoding change, as well as EOL selection, comma/tab seperator switch, it'll become a very decent tool for editing not only MM's .txt, but ANY CSV file.

[tomchen1989]

@Templayer
I've checked https://www.dropbox.com/sh/y08ensa5qajm ... h35Ma?dl=0 Czech localization files. Text files seems to be encoded in cp1250 as I predicted, but English, French version's .fnt files are still being used, which cause the problem. I don't know if Czech language versions MM 6, 7 or 8 have ever been released. If so, get all .fnt files from Czech version MM to replace English, French version's .fnt files.

If no official version exists, that'll be a little complicated. You should:
- Either use my script (which I'll release soon) to generate all .fnt files, but you'll have to find similar fonts in normal font format. There are 12 .fnt in the game file, maybe 5 or 6 are actually used, you'll have to find similar font for each of them.
- Or edit each .fnt files that are currently used, to make the char code map correctly to cp1250's chars, you'll need to export BMP from .fnt files using .fnt editor, edit the BMP and import them back into .fnt.

[Templayer]

@Templayer
I've checked https://www.dropbox.com/sh/y08ensa5qajm ... h35Ma?dl=0 Czech localization files. Text files seems to be encoded in cp1250 as I predicted, but English, French version's .fnt files are still being used, which cause the problem. I don't know if Czech language versions MM 6, 7 or 8 have ever been released. If so, get all .fnt files from Czech version MM to replace English, French version's .fnt files.

If no official version exists, that'll be a little complicated. You should:
- Either use my script (which I'll release soon) to generate all .fnt files, but you'll have to find similar fonts in normal font format. There are 12 .fnt in the game file, maybe 5 or 6 are actually used, you'll have to find similar font for each of them.
- Or edit each .fnt files that are currently used, to make the char code map correctly to cp1250's chars, you'll need to export BMP from .fnt files using .fnt editor, edit the BMP and import them back into .fnt.

This might be beyond my capability, at least for now, because I know virtually nothing about those.

A side effect of never having to deal with this kind of crap due to a very good software architecture at our company.

I do not think a really official version exists. My uncle has a 600 page Czech manual for MM6, which seems to be official, but the Czech is made through a Czech patch, I think.

This whole font thing is something I might need help with. And I have no way of testing that.
Hmm. I wonder if I can use the console to quickly finish the games and get to the problematic Lose text.

[tomchen1989]

Wait a minute, I re-downloaded the Czech localization files and found the .fnt files in it are indeed using cp1250 (Czech, Polish) encoding. I must have downloaded the wrong files or got my folders mixed up or something...

Then it's not a problem of the .fnt files. Your text files where "unreadable text" comes from must have been saved with wrong encoding, likely cp1252, or maybe ISO/IEC_8859-2. You should find those text files, open it with its current encoding, then save it with cp1250.

I actually used the Czech localization files and despite the interface (button images) are not localized, the translated text looks pretty normal to me. Don't know where your problem is.

(btw, here are the links to Czech patches for MM6 MM7 MM8, but you might have already had those.)

[Lucius]

Jump spell is very weak compared to vanilla games. I cannot jump into many places which could have been reached in vanilla mm6.

Jump is weaker in 7/8 vs 6 to begin with, are you sure you're not just experiencing 8's jump inside 6's map design?

This is exactly what I mean - I cannot jump to many places it was possible to do in vanilla mm6 because jump is weaker.
So I suggest to increase jump altitude.

[Templayer]

Jump spell is very weak compared to vanilla games. I cannot jump into many places which could have been reached in vanilla mm6.

Jump is weaker in 7/8 vs 6 to begin with, are you sure you're not just experiencing 8's jump inside 6's map design?

This is exactly what I mean - I cannot jump to many places it was possible to do in vanilla mm6 because jump is weaker.
So I suggest to increase jump altitude.

Just on Enroth I presume? Because the level design there WAS taking a stronger Jump spell into consideration...

My personal suggestion would be a setting of MM6VanillaJumpSpellHeight.
0 - disabled (i.e. newer formula for jump height used)
1 - enabled (older MM6 formula for jump height used)
2 - enabled only on the MM6 continent

[Ck-NoSFeRaTU]

Issue: Whenever I try to change interface (which is always from MM8, modified, no matter where I start the game), I'm getting CTD with the message "Application error: "ExcepctionElayout exception in module MM8patch.dll at 00058CA3. UI Layout line 60: Negative coordinate: (53, -21.. 106, 31)."

Weird. Did you do base merge + newest Merge patch + newest GF patch?
Anyway - the bug has been added to the Bug Tracker.

This error happens with UILayout=UI due to patch.UI.txt. Icons UI-BtnRd, UI-BtnRu, UI-BtnZd, UI-BtnZu seems missing so coordinates become negative for BtnZ.
If anyone need it I patched it a bit to use old mm7 icons for now. I replaced duplicated spell book button with rest button (so it is now in left bottom corner for 6 and 7 UI) and reused spell book icon for quick ref. But overall 6 & 7 UIs require much further work, currently it is complete mishmash.

BTW, "Community" version at https://gitlab.com/templayer/mmmerge/ is crashing due to stripped windows CR in txt tables.

[Templayer]

Issue: Whenever I try to change interface (which is always from MM8, modified, no matter where I start the game), I'm getting CTD with the message "Application error: "ExcepctionElayout exception in module MM8patch.dll at 00058CA3. UI Layout line 60: Negative coordinate: (53, -21.. 106, 31)."

Weird. Did you do base merge + newest Merge patch + newest GF patch?
Anyway - the bug has been added to the Bug Tracker.

This error happens with UILayout=UI due to patch.UI.txt. Icons UI-BtnRd, UI-BtnRu, UI-BtnZd, UI-BtnZu seems missing so coordinates become negative for BtnZ.
If anyone need it I patched it a bit to use old mm7 icons for now. I replaced duplicated spell book button with rest button (so it is now in left bottom corner for 6 and 7 UI) and reused spell book icon for quick ref. But overall 6 & 7 UIs require much further work, currently it is complete mishmash.

BTW, "Community" version at https://gitlab.com/templayer/mmmerge/ is crashing due to stripped windows CR in txt tables.

Yes, Community. Since nobody can edit the first post but Rodril, and he doesn't integrate the community changes very much or very often.

The CR removal is due to a GitLab bug, actually. https://gitlab.com/gitlab-org/gitlab-foss/issues/44032

I need to reapply this and then repush stuff, I think:
$ git config --global core.autocrlf false

Code: Select all

Markus Koller
Markus Koller  @toupeira · 4 months ago
Developer

@phikai Reproduced this on Linux and Git 2.23 with default settings (which seems to be no auto-conversion).

    Created some test files at https://gitlab.com/toupeira/playground/tree/line-endings
    Edited on GitLab with ACE (single-file editor) and WebIDE (branches line-endings-ace and line-endings-webide)
        With both I end up with CRLF converted to LF
        WebIDE does detect CRLF/LF correctly and indicate it in the bottom status bar as @pslaughter pointed out
        Both editors also send the CRLFs correctly to the backend (visible as \r\n in the browser console), though through different endpoints (ACE uses a form submit, WebIDE uses the commits API)
    Pushed the repository to GitHub and edited there, CRLF is correctly preserved (pushed back to GitLab in line-endings-github)

core.autocrlf=input seems to always convert to LF according to https://stackoverflow.com/q/3206843, which would explain the current behaviour in GitLab:

    This means that Git will process all text files and make sure that CRLF is replaced with LF when writing that file to the object database. It will not, however, do the reverse. When you read files back out of the object database and write them into the working directory they will still have LFs to denote the end of line. This setting is generally used on Unix/Linux/OS X to prevent CRLFs from getting written into the repository. The idea being that if you pasted code from a web browser and accidentally got CRLFs into one of your files, Git would make sure they were replaced with LFs when you wrote to the object database.

I wonder if the "correct" solution would be to stop using core.autocrlf=input in Gitaly and preserve whatever clients send, but we probably have very good reasons for only storing LFs and this might break all sorts of things (such as diffs). On the other hand, if I understand correctly it's only applied to new commits and if we pull repositories from other sources they might contain CRLFs too.

I tried to use a .gitattributes file but it seems core.autocrlf takes priority over that (the gitattributes manpage somewhat confirms this as well). I'm not sure if Gitaly actually uses it, but I've observed the same behaviour locally so this seems like a no-go.

@smoe1 As for the final line ending: In all three editors including GitHub they're displayed as a trailing blank line (which is indeed what they are), and while GitLab just uses the input as-is (meaning you'll lose it if you're adding a new line without adding a blank one after it), GitHub always appends one even if there wasn't one before, which seems preferable to me. Git itself doesn't seem to have auto-conversion for this one, so we should probably spin this out into its own issue.

I'll get on it on the weekend. And I will add your fix to the community version as well. Because it might never appear in the base one, ever.

[justhoneypie]

Hello,

Not sure this is the right topic - I can't find armageddon spell on sale anywhere. Just got my necromant level 7 - Master trainer.

Thanks!

[Ck-NoSFeRaTU]

I need to reapply this and then repush stuff, I think:
$ git config --global core.autocrlf false

If your local repo copy is still intact this will fix it. Just don't forget to change modification date of affected files with something like touch command for git to see the changes for new commit.

In a long perspective IMHO It'll be better to allow both non-CR and CR tables/scripts on read and use non-CR on write. I personally hate any presense of CR as non-windows user. Especially in repositories. But it will require changes all over the place from scripts to grayface's patches.

[Templayer]

Hello,

Not sure this is the right topic - I can't find armageddon spell on sale anywhere. Just got my necromant level 7 - Master trainer.

Thanks!

If you mean in the Merge - some spells are now harder to get. It was weird that anybody could buy a spell that literally literally (hehe) means the End of the World on some shelf in a shop. Identified, i.e. the shop owner knowing what he is selling.

[justhoneypie]

Hello,

Not sure this is the right topic - I can't find armageddon spell on sale anywhere. Just got my necromant level 7 - Master trainer.

Thanks!

If you mean in the Merge - some spells are now harder to get. It was weird that anybody could buy a spell that literally literally (hehe) means the End of the World on some shelf in a shop. Identified, i.e. the shop owner knowing what he is selling.

Thanks for the heads up

I was wondering if you could...reveal to me where would I be able to find more information about how to obtain those spells and move between mm6-7-8,please?

So far - I love it, especially the higher resolution!

[Templayer]

Hello,

Not sure this is the right topic - I can't find armageddon spell on sale anywhere. Just got my necromant level 7 - Master trainer.

Thanks!

If you mean in the Merge - some spells are now harder to get. It was weird that anybody could buy a spell that literally literally (hehe) means the End of the World on some shelf in a shop. Identified, i.e. the shop owner knowing what he is selling.

Thanks for the heads up

I was wondering if you could...reveal to me where would I be able to find more information about how to obtain those spells and move between mm6-7-8,please?

So far - I love it, especially the higher resolution!

From random drops for example. I also remember there being a pre-placed drop in one of the castles (Darkmoor? For Armageddon? In vanilla?) in MM6, if that still applies.

To move between the continents, check the Tracker in my signature. There is even a spoiler section for that specifically, so that you could jump VERY early (so that for example you could run all the starting locations first, etc.). I might have to rename the Tracker to a Manual at this point.

[justhoneypie]

If you mean in the Merge - some spells are now harder to get. It was weird that anybody could buy a spell that literally literally (hehe) means the End of the World on some shelf in a shop. Identified, i.e. the shop owner knowing what he is selling.

Thanks for the heads up

I was wondering if you could...reveal to me where would I be able to find more information about how to obtain those spells and move between mm6-7-8,please?

So far - I love it, especially the higher resolution!

From random drops for example. I also remember there being a pre-placed drop in one of the castles (Darkmoor? For Armageddon? In vanilla?) in MM6, if that still applies.

To move between the continents, check the Tracker in my signature. There is even a spoiler section for that specifically, so that you could jump VERY early (so that for example you could run all the starting locations first, etc.). I might have to rename the Tracker to a Manual at this point.

I think I misunderstand...Did I get the hint that there's stone circle in murmurwoods, but there's no portal in that druid circle?

[Templayer]

Thanks for the heads up

I was wondering if you could...reveal to me where would I be able to find more information about how to obtain those spells and move between mm6-7-8,please?

So far - I love it, especially the higher resolution!

From random drops for example. I also remember there being a pre-placed drop in one of the castles (Darkmoor? For Armageddon? In vanilla?) in MM6, if that still applies.

To move between the continents, check the Tracker in my signature. There is even a spoiler section for that specifically, so that you could jump VERY early (so that for example you could run all the starting locations first, etc.). I might have to rename the Tracker to a Manual at this point.

I think I misunderstand...Did I get the hint that there's stone circle in murmurwoods, but there's no portal in that druid circle?

If I remember correctly - touch it and then cast Town portal. Or something like that.

[raekuul]

When you see the message "you feel a high magic presence here", you're good to go. Murmurwoods is frustrating to TP out from before GM Water though due to an overabundance of Basilisks.

[Templayer]

BTW, "Community" version at https://gitlab.com/templayer/mmmerge/ is crashing due to stripped windows CR in txt tables.

The CR removal is due to a GitLab bug, actually. https://gitlab.com/gitlab-org/gitlab-foss/issues/44032

I need to reapply this and then repush stuff, I think:
$ git config --global core.autocrlf false

Code: Select all

Markus Koller
Markus Koller  @toupeira · 4 months ago
Developer

@phikai Reproduced this on Linux and Git 2.23 with default settings (which seems to be no auto-conversion).

    Created some test files at https://gitlab.com/toupeira/playground/tree/line-endings
    Edited on GitLab with ACE (single-file editor) and WebIDE (branches line-endings-ace and line-endings-webide)
        With both I end up with CRLF converted to LF
        WebIDE does detect CRLF/LF correctly and indicate it in the bottom status bar as @pslaughter pointed out
        Both editors also send the CRLFs correctly to the backend (visible as \r\n in the browser console), though through different endpoints (ACE uses a form submit, WebIDE uses the commits API)
    Pushed the repository to GitHub and edited there, CRLF is correctly preserved (pushed back to GitLab in line-endings-github)

core.autocrlf=input seems to always convert to LF according to https://stackoverflow.com/q/3206843, which would explain the current behaviour in GitLab:

    This means that Git will process all text files and make sure that CRLF is replaced with LF when writing that file to the object database. It will not, however, do the reverse. When you read files back out of the object database and write them into the working directory they will still have LFs to denote the end of line. This setting is generally used on Unix/Linux/OS X to prevent CRLFs from getting written into the repository. The idea being that if you pasted code from a web browser and accidentally got CRLFs into one of your files, Git would make sure they were replaced with LFs when you wrote to the object database.

I wonder if the "correct" solution would be to stop using core.autocrlf=input in Gitaly and preserve whatever clients send, but we probably have very good reasons for only storing LFs and this might break all sorts of things (such as diffs). On the other hand, if I understand correctly it's only applied to new commits and if we pull repositories from other sources they might contain CRLFs too.

I tried to use a .gitattributes file but it seems core.autocrlf takes priority over that (the gitattributes manpage somewhat confirms this as well). I'm not sure if Gitaly actually uses it, but I've observed the same behaviour locally so this seems like a no-go.

@smoe1 As for the final line ending: In all three editors including GitHub they're displayed as a trailing blank line (which is indeed what they are), and while GitLab just uses the input as-is (meaning you'll lose it if you're adding a new line without adding a blank one after it), GitHub always appends one even if there wasn't one before, which seems preferable to me. Git itself doesn't seem to have auto-conversion for this one, so we should probably spin this out into its own issue.

I'll get on it on the weekend.

I've spent half a day trying to fix it, but no matter what I set or do, if the .zip generating is used (i.e. for non-git users, anonymous people just wanting to download the crap), it fuc.ks up the line breaks. Even if I rename the .txt to .table before a commit with files having CRLF, it still screws them up in the repository.

So for now I've fixed it by making a .bat file that fixes those files and needs to be run after extraction. Check the updated installation instructions on GitLab.

I've tested it and the game works, and after browsing through the fixed files I can see no problems...

... if I've screwed up something again, or overlooked something, please report.

[cthscr]

I've spent half a day trying to fix it, but no matter what I set or do, if the .zip generating is used (i.e. for non-git users, anonymous people just wanting to download the crap), it fuc.ks up the line breaks. Even if I rename the .txt to .table before a commit with files having CRLF, it still screws them up in the repository.

Well, I don't know, but zip archive generated from merge request works fine to me.

[Ck-NoSFeRaTU]

I've spent half a day trying to fix it, but no matter what I set or do, if the .zip generating is used (i.e. for non-git users, anonymous people just wanting to download the crap), it fuc.ks up the line breaks. Even if I rename the .txt to .table before a commit with files having CRLF, it still screws them up in the repository.

Dunno what's your issue but I just checked and all is working as intended on gitlab part.
Created a merge request if you want it. You can check prior to merging this zip generated by gitlab