| Time |
Nick |
Message |
| 00:04 |
|
Warr1024 joined #minetest-dev |
| 00:54 |
|
hmmmm joined #minetest-dev |
| 01:36 |
|
Warr1024 left #minetest-dev |
| 03:05 |
|
kaeza joined #minetest-dev |
| 03:28 |
|
ssieb joined #minetest-dev |
| 03:40 |
|
kaeza joined #minetest-dev |
| 04:46 |
|
kaeza joined #minetest-dev |
| 06:23 |
|
kaeza joined #minetest-dev |
| 06:53 |
|
ImQ009 joined #minetest-dev |
| 07:20 |
|
Dormo joined #minetest-dev |
| 07:25 |
|
Dormo joined #minetest-dev |
| 07:25 |
Dormo |
Hey, can anyone help me? I fixed a bug in Minetest, and the fix doesn't appear to break anything else, but I think it might and I just can't think of an edge case. |
| 07:26 |
kaeza |
Dormo, can you be more specific? |
| 07:27 |
kaeza |
what bug? |
| 07:27 |
Dormo |
There was a bug filed that mentioned if youre falling from any hight while holding shift (sneak) and land on a slab (stairs), no damage occurs |
| 07:28 |
Dormo |
you also remain in the air above the slab until you let go of sneak |
| 07:29 |
kaeza |
Dormo, open a pull request for it so others can review it |
| 07:30 |
Dormo |
Alright. Thanks |
| 07:30 |
Dormo |
I was just worried it'd get accepted blindly and break other things. |
| 07:32 |
celeron55 |
that's not a concern here 8) |
| 07:33 |
celeron55 |
(it doesn't mean one should overwork the core devs by requesting crap to be merged though; it wastes everyone's time) |
| 07:39 |
|
emptty joined #minetest-dev |
| 07:42 |
Dormo |
Groovy. Submitted the pull request. Hopefully I didn't make an ass out of myself |
| 07:46 |
celeron55 |
seems very hacky |
| 07:46 |
Dormo |
That's why I was concerned. |
| 07:47 |
Dormo |
I didn't expect it to work, I thought I'd have to check the node type to make sure you were on stairs |
| 07:47 |
Dormo |
I just did that to see what would happen, and it didn't seem to break anything |
| 07:48 |
Dormo |
I thought "Okay, if this fixes the bug, I'll make a conditional that checks to make sure the node is stairs", but it just seemed to work fine in general |
| 07:48 |
celeron55 |
i don't think that can be merged as-is, but can you test what happens with that patch when you are sneaking on a pile of nodes and dig the topmost node while you keep sneaking? |
| 07:49 |
celeron55 |
there must be something that will break but i have no idea what :P |
| 07:49 |
Dormo |
Sure. And I know, right? |
| 07:51 |
Dormo |
It seems to work as expected. |
| 07:51 |
Dormo |
Let me try it on slabs... |
| 07:52 |
celeron55 |
at least for sure this patch won't work for things that aren't half-cube high |
| 07:52 |
Dormo |
That's true, but will that ever happen? |
| 07:53 |
Dormo |
What if |
| 07:53 |
celeron55 |
yes - and also it screws up something for sure, because it changes the scaling of a position variable that is used for other than vertical positioning too |
| 07:54 |
Dormo |
What if I made a conditional. If the node is stairs, do my patch. If the node isnt stairs, do the original code |
| 07:55 |
celeron55 |
it's still too hacky due to what i just mentioned |
| 07:56 |
Dormo |
I see. Alright. I'll try to get more familiar with what the code does, then |
| 07:57 |
celeron55 |
but if you want to check if something is stairs, you actually have to check if the node is of the "nodebox" type - that is the correct conditional from the engine's standpoint |
| 07:57 |
Dormo |
nodebox == stairs? |
| 07:57 |
celeron55 |
(nodebox means the node can be other than a full cube) |
| 07:58 |
Dormo |
Oh, I see |
| 07:58 |
Dormo |
Wait, are you sure? |
| 07:59 |
Dormo |
Because when I was on the slab it said m_old_node_below_type was "stairs:brick" |
| 07:59 |
Dormo |
Or is that the wrong way to check the node type? |
| 07:59 |
celeron55 |
oh it's a different thing |
| 08:00 |
celeron55 |
you need to look up the node's definition (a ContentFeatures struct from the nodemgr thing there) and see if it's drawtype is NDT_NODEBOX |
| 08:01 |
Dormo |
Is there any way to get the node's dimensions? |
| 08:02 |
celeron55 |
yes; if it's NDT_NODEBOX, all the axis-aligned boxes the node consists of are in the node definition field node_box |
| 08:02 |
celeron55 |
it's a list of axis-aligned boxes so it gets quite nasty from your standpoint |
| 08:05 |
celeron55 |
if you manage to whip up some nice generic way of handling that, it will be appreciated 8) |
| 08:05 |
|
Taoki[mobile] joined #minetest-dev |
| 08:07 |
Dormo |
Yeah, now that I know nodes can be things other than either a full cube or a half cube, I'll try to come up with a generic solution. Hopefully I can get something in the next few days. |
| 08:47 |
|
serengeor joined #minetest-dev |
| 08:47 |
|
Calinou joined #minetest-dev |
| 09:01 |
|
Deivan joined #minetest-dev |
| 09:03 |
|
PilzAdam joined #minetest-dev |
| 09:15 |
|
Zeg9 joined #minetest-dev |
| 09:23 |
|
nyuszika7h joined #minetest-dev |
| 09:40 |
|
smoke_fumus joined #minetest-dev |
| 09:48 |
|
MarkehMe joined #minetest-dev |
| 10:11 |
|
kaeza joined #minetest-dev |
| 10:39 |
|
Calinou joined #minetest-dev |
| 10:41 |
|
serengeor joined #minetest-dev |
| 10:47 |
|
kaeza joined #minetest-dev |
| 10:52 |
|
sapier joined #minetest-dev |
| 10:53 |
sapier |
http://ompldr.org/iaWQyeA what do you think about adding a ingame mod management tab like this? |
| 10:55 |
VanessaE |
nice |
| 10:55 |
sapier |
yes but for this to be of real use current mod/game administration needs a major overhaul |
| 10:56 |
sapier |
e.g. mods no longer should be spread over 30 folders |
| 10:58 |
sapier |
imho games should be just configuration files ... and yes I know my suggestions might add limitations that aren't there atm but they'd reduce complexity of mod/game handling by some factors |
| 10:59 |
VanessaE |
games have to be more than just config files, as a game consists of mods |
| 10:59 |
VanessaE |
where are they gonna get their content from? |
| 10:59 |
|
Taoki joined #minetest-dev |
| 11:00 |
sapier |
as far as I understand a game is a subset of all available mods installed on users machine |
| 11:00 |
sapier |
thus a game could be a config file selecting those mods required for it |
| 11:00 |
VanessaE |
no, not at all |
| 11:00 |
VanessaE |
a game is a collection of mods that may or may not be separately available anywhere |
| 11:01 |
VanessaE |
plus in most cases, some of the mods included in a given game will be different from what is officially available separately. |
| 11:01 |
sapier |
and what does "way of distribution" have to do with what a game is? Yes I admit "installing" a game is a special case |
| 11:01 |
VanessaE |
for example, my "game" has default in it, but modified to disable leaf decay. |
| 11:02 |
* darkrose |
is exceptionally pleased with the mod management in use on her server |
| 11:02 |
ironzorg |
inb4 hash |
| 11:02 |
PilzAdam |
dont you have a 0.3 server? |
| 11:02 |
darkrose |
exactly |
| 11:02 |
sapier |
if this is minimum requirement I'll stop ingame modmanager development... I don't se any way to create a usable modmanager this way |
| 11:02 |
VanessaE |
sapier: I'm not talking about methods of distribution. I'm talking about the content of the game and where that content comes from. |
| 11:03 |
VanessaE |
managing the download and install of individual mods the way you're thinking is probably fine |
| 11:03 |
sapier |
what you want is 200 degrees of freedom to implement this in a modmanager will result in a complex system not usable/understandable by anyone |
| 11:04 |
VanessaE |
but games must be handled as self-contained packages, not as some kind of generic "use this, this, and this mod" configs |
| 11:05 |
sapier |
meaning mod manager should be game specific, user needs to download and install a mod separate for each game |
| 11:05 |
VanessaE |
where did I say that? |
| 11:05 |
sapier |
mixing game specific and global mods is a no-go imho either all mods are game specific or none |
| 11:05 |
VanessaE |
if your proposed solution removes even one option we have now, it is a non-starter. |
| 11:06 |
sapier |
ok then forget about it current mod/game selection mess can't be fixed |
| 11:06 |
VanessaE |
but if you can make it install mods on a per-game basis, the way we do it now, but of course with automated download-and-install like you're suggesting, go for it |
| 11:07 |
sapier |
I'm not willing to waste 2 weeks to create a mod manager supporting nonsense options sorry |
| 11:07 |
sapier |
the mod manager does only make sense if mod mess can be cleant up by adding it otherwise its waste of time |
| 11:08 |
VanessaE |
what mess? |
| 11:08 |
sapier |
vanessae we do not install mods in a per game basis now ... at least not completely |
| 11:08 |
VanessaE |
since when? |
| 11:08 |
sapier |
common mods |
| 11:08 |
VanessaE |
~/.minetest/mods/gameid/mods |
| 11:09 |
VanessaE |
for whatever gameid you want to consider |
| 11:09 |
VanessaE |
common is about to be re-absorbed. |
| 11:09 |
VanessaE |
PilzAdam: since he brings it up, can we just get that overwith now? |
| 11:09 |
sapier |
one option less still the duplication issue remains |
| 11:09 |
VanessaE |
who cares about duplication? disk space is cheap. |
| 11:10 |
VanessaE |
if it's that big of a concern, run a de-duplication fs :-) |
| 11:10 |
VanessaE |
or use symbolic links |
| 11:10 |
kaeza |
Windows? symbolic links? what? |
| 11:10 |
sapier |
so I need to add checksums recheck every time ... |
| 11:10 |
VanessaE |
kaeza: "shortcuts" in windows |
| 11:11 |
VanessaE |
checksums? you're overcomplicating this. |
| 11:11 |
kaeza |
VanessaE, shortcuts != symlinks |
| 11:11 |
VanessaE |
just download, install, done. |
| 11:11 |
VanessaE |
forget all the security-hardening shit for now |
| 11:11 |
sapier |
it's not security is just usability |
| 11:11 |
VanessaE |
kaeza: well they perform the same function as a symlink don't they? |
| 11:11 |
kaeza |
VanessaE, nope |
| 11:11 |
VanessaE |
sapier: like I just said - download, install, done. |
| 11:11 |
sapier |
user won't be very glad if his game won't work anymore once some other games mod is updated |
| 11:11 |
VanessaE |
don't make it more complicated than it truly needs to be |
| 11:12 |
VanessaE |
gah! |
| 11:12 |
VanessaE |
you're not listening |
| 11:12 |
kaeza |
A symlink is handled by the kernel/whatever in Linux, but a shortcut is a regular file on Windows |
| 11:12 |
VanessaE |
game mods are on a per-game basis already |
| 11:12 |
sapier |
no it is complicated vanessae you're just ignoring any problem that doesn't fit to your special way of using games |
| 11:12 |
VanessaE |
a mod in minetest/mods isn't going to interfere with build/mods is it? |
| 11:12 |
sapier |
if you use symlink it is interfering |
| 11:13 |
VanessaE |
what you propose is one mod to rule them all |
| 11:13 |
sapier |
no I'm just looking for better solutions than current mess |
| 11:13 |
VanessaE |
*facepalm* |
| 11:13 |
VanessaE |
again: |
| 11:13 |
VanessaE |
who cares about duplication? disk space is cheap. |
| 11:14 |
VanessaE |
= just install the damn mods wherever they're needed. |
| 11:14 |
VanessaE |
if one goes out of date, the user can fix it. |
| 11:14 |
sapier |
it's not about duplication but about duplicated work redownload and install a mod to multiple games |
| 11:14 |
sapier |
it's usability not disk space |
| 11:15 |
VanessaE |
well you have precisely two choices and no more: |
| 11:15 |
VanessaE |
either you maintain multiple copies (one per game), or you maintain one copy and link to it from everywhere else by some mechanism |
| 11:15 |
VanessaE |
and by "link to it" I also mean your overcomplicated config file idea. |
| 11:16 |
kaeza |
how about using "common" as symlink mechanism? |
| 11:16 |
sapier |
common is to be removed |
| 11:16 |
VanessaE |
well? |
| 11:16 |
sapier |
but I was thinking about replacing game folder by same mechanism used for common |
| 11:16 |
VanessaE |
pick one already |
| 11:16 |
sapier |
bot options are crap vanessae there must be another one |
| 11:17 |
VanessaE |
there isn't |
| 11:17 |
VanessaE |
you either have one copy, or you have more than one. |
| 11:17 |
VanessaE |
there is no third option. |
| 11:17 |
sapier |
and complexity isn't added by me but requirement to be compatible to all nonsense bugs and features added ages ago |
| 11:18 |
VanessaE |
what bugs? |
| 11:18 |
VanessaE |
you're overcomplicating it again |
| 11:18 |
sapier |
e.g. non working prefix check |
| 11:19 |
sapier |
one copy implies games to be config files instead of folders that's what you dont want to have |
| 11:19 |
kaeza |
sapier, you can have both config files *and* actual mods in a game |
| 11:19 |
sapier |
and don't say "symlinks" a folder containing symlinks is nothing different than a very special config file |
| 11:20 |
VanessaE |
you either have one copy, or you have more than one. PICK ONE. |
| 11:20 |
sapier |
if you do the implementation and cross checking kaeza ;-) |
| 11:20 |
kaeza |
sapier, the mod manager should not need to check individual mods in a game (for update, etc), but rather the game as a whole |
| 11:21 |
sapier |
imho this is to game centric |
| 11:21 |
VanessaE |
it's supposed to be! |
| 11:22 |
sapier |
if it's supposed to be of what use should any moddb be? |
| 11:22 |
kaeza |
sapier, look at all the games posted in the forums; most (if not all) of them are meant to be used stand-alone |
| 11:22 |
kaeza |
I mean, it's supposed you wouldn't need other mods if you download Dwarves for example |
| 11:22 |
sapier |
do you really expect me to implement a game only solution? you know I do NOT release any game |
| 11:23 |
VanessaE |
sapier: MOBF is practically a game by itself ;) |
| 11:23 |
sapier |
no it isn't |
| 11:23 |
VanessaE |
emphasis on *practically* |
| 11:23 |
sapier |
still it's never been meant as a full game |
| 11:23 |
VanessaE |
jeez man get a sense of humor |
| 11:23 |
sapier |
it's just a small piece |
| 11:24 |
kaeza |
sapier, treat "games" as simple "modpacks" |
| 11:24 |
VanessaE |
which is precisely all a game really i |
| 11:24 |
VanessaE |
is |
| 11:24 |
sapier |
I do I just dislike getting suggested to implement things that are of no use to me as only "possible solution" ;-P |
| 11:24 |
VanessaE |
it's a modpack with an extra file declaring that it isn't a modpack. |
| 11:24 |
kaeza |
you don't update a modpack by updating individual mods |
| 11:25 |
kaeza |
you update it as a whole |
| 11:25 |
VanessaE |
exactly. |
| 11:25 |
VanessaE |
let the game's author/maintainer decide if the mods therein need updates. |
| 11:25 |
sapier |
so we now have 3 things to handle |
| 11:25 |
kahrl |
hi, what are we arguing about today? :D |
| 11:25 |
sapier |
games/modpacks/mods |
| 11:25 |
VanessaE |
maybe the game author wants some mods in their game to be old, on purpose? |
| 11:25 |
sapier |
do you really think any user will understand what he's doing? |
| 11:26 |
VanessaE |
sapier: they can learn. |
| 11:26 |
sapier |
no they wont |
| 11:26 |
sapier |
kahrl almost same as last time ;-) |
| 11:26 |
VanessaE |
then they won't know what they're doing and they can either ask for help or go away |
| 11:26 |
kaeza |
sapier, it's better than browsing the entire forum and installing manually (plus possible renaming) |
| 11:27 |
kahrl |
hmm. why not add a button "games" to the bottom row of games when in the mod manager tab |
| 11:27 |
kahrl |
which will display a list of installed and available games |
| 11:27 |
sapier |
no it isn't ui in minetest is quite limited adding complexity like add games/add modpack/add mod linking everything together ... |
| 11:27 |
VanessaE |
kahrl: "Install..." "Delete..." |
| 11:27 |
sapier |
user will never know where he needs to add what |
| 11:28 |
VanessaE |
sapier: every mod, modpack, or game is already clearly labeled as such on the forums |
| 11:28 |
kahrl |
you don't need to differentiate between mod and modpack in the mod manager |
| 11:28 |
sapier |
if you have a look at tab list of my suggestion I already added a games tab |
| 11:28 |
VanessaE |
you expect us to hold the users' hands THAT far? |
| 11:28 |
VanessaE |
"connect all the little dots?" |
| 11:29 |
sapier |
there must be some reason why users tell minetest is to complicated ... they wouldn't do if it was as simple as you claim |
| 11:29 |
VanessaE |
they say it's too complicated because we don't have enough active content in the "default game" (e.g. minetest_game). |
| 11:29 |
kaeza |
oh dear |
| 11:29 |
VanessaE |
but it's still FAR less complicated than Minecraft! |
| 11:30 |
VanessaE |
(because, that is, we require users to install a few mods here and there to get the content they watn) |
| 11:30 |
VanessaE |
want8 |
| 11:30 |
VanessaE |
... |
| 11:30 |
sapier |
i just asked by bro he tells me he doesn't want t browse in formum look for compatible mods building a new game everytime |
| 11:30 |
sapier |
not knowing where to put them either |
| 11:31 |
kaeza |
sapier, the mod manager would eliminate the "don't know where to put what" problem |
| 11:31 |
sapier |
no its not less complicated than minecraft because minecraft is limited but easy |
| 11:31 |
sapier |
yes but only if user does understand HOW to use it |
| 11:31 |
kahrl |
until you want to install mods... then hello jar hell |
| 11:31 |
VanessaE |
minecraft easy!? |
| 11:32 |
VanessaE |
you have to patch the damn engine for every mod |
| 11:32 |
VanessaE |
and pray that everything you install is cross-compatible. |
| 11:32 |
VanessaE |
that's not anywhere near as much a problem with minetest |
| 11:32 |
kaeza |
not to mention you must use a matching client and server |
| 11:32 |
VanessaE |
yes, that too |
| 11:32 |
sapier |
those who want mods are skilled enough to do it but this is not the audience I'm talking about |
| 11:32 |
VanessaE |
sapier: then you're focused on the wrong audience. |
| 11:32 |
kaeza |
well... I'm confused |
| 11:33 |
sapier |
imho you're focused on wrong audience ;-) |
| 11:33 |
|
MarkehMe left #minetest-dev |
| 11:33 |
VanessaE |
those who find minetest too complicated are the ones who want more content than is the default and there are only two ways to solve that: 1. include more content. 2. make the user install mods. |
| 11:33 |
sapier |
you don't need a mod manager for developers |
| 11:33 |
kaeza |
and what is the intended audience again? |
| 11:33 |
kahrl |
I dislike the idea of dumbing down programs if it removes interesting choices |
| 11:33 |
|
Taoki joined #minetest-dev |
| 11:34 |
sapier |
the only removed option would be wasting memory by duplication and of course modders wouldn't have a "game" specific storage anymore |
| 11:34 |
VanessaE |
memory!? |
| 11:34 |
VanessaE |
DISK SPACE IS CHEAP. |
| 11:34 |
sapier |
don't know what use this could be either |
| 11:35 |
kaeza |
kahrl, I don't like it either, but it does not remove the "interesting choices" |
| 11:36 |
kaeza |
and let's face it. the average computer user of today is not the same as it was 20 years ago (or even 10 or 5) |
| 11:36 |
VanessaE |
I have somewhere north of 75 mods in my standard collection, of which 53 of them are part of my "game". it wouldn't matter if I had a thousand of them on hand, they aren't going to take any memory of they're not being used, they just need some disk space. Even my bloated-ass "game" is only 50 MB. |
| 11:36 |
VanessaE |
(the balance are not installed at all, mind you) |
| 11:36 |
sapier |
vanessae I understand your main focus is in game level thats good but it's not the only possible view |
| 11:36 |
kahrl |
the idea of having different games was not to have them all be "basically the same with a few differences" |
| 11:37 |
kahrl |
even if that's how it mostly looks today |
| 11:37 |
VanessaE |
sapier: my main focus is on everything - all three parts of the modding system |
| 11:37 |
kahrl |
so it must be possible to have different games with completely different mods even if those mods have the same name |
| 11:37 |
VanessaE |
mods, modpacks, and games - I'm concerned with all three. |
| 11:38 |
|
Taoki joined #minetest-dev |
| 11:38 |
kaeza |
I know, but there are games that can be easily extended, and those that come in "self-contained" version |
| 11:39 |
|
iqualfragile joined #minetest-dev |
| 11:39 |
kaeza |
there are some games out there in which adding a simple mod can break things disastrously |
| 11:39 |
sapier |
assuming we keep the game folders .... having a global mod manager isn't of any use |
| 11:39 |
kahrl |
kaeza: that's the fault of the mod then, I guess |
| 11:39 |
sapier |
the only possibility would be make the mod manager game specific |
| 11:39 |
kaeza |
kahrl, not always |
| 11:40 |
kahrl |
so the modder didn't test compatibility with that particular game? then it should not be marked compatible with that game |
| 11:40 |
sapier |
so first thing to ad would be a game manager |
| 11:40 |
celeron55 |
kahrl: personally i haven't yet figured out what kind of mod namespacing we want |
| 11:40 |
VanessaE |
sapier: so don't make it game specific. make it install games into ~/.minetest/games where they belong, and mods/modpacks into ~/.minetest/mods/<gameid>/mods where they belong. |
| 11:40 |
VanessaE |
anything more than this is overcomplictaed. |
| 11:41 |
VanessaE |
overcomplicated* |
| 11:41 |
kaeza |
anyway, we would need a mod versioning scheme |
| 11:41 |
sapier |
imho games do not belog to ~ |
| 11:41 |
VanessaE |
(or their appropriate folders for the run-in-place version of course) |
| 11:41 |
VanessaE |
sapier: you don't believe in multi-user OSs? |
| 11:41 |
sapier |
this is just crap if you have a development model like minetest |
| 11:42 |
sapier |
no i don't even believe in user global installing mods |
| 11:42 |
sapier |
i use RUN_IN_PLACE only |
| 11:42 |
kahrl |
sapier: if a distribution wants to distribute a game via their package manager they can still use /usr/share |
| 11:42 |
VanessaE |
what global? ~ is the user's homedir. |
| 11:42 |
sapier |
"user global" |
| 11:42 |
kaeza |
global for what? |
| 11:43 |
VanessaE |
sapier: there's no such thing. Either you've got per-user, or you've got system global (with appropriate permissions/acls) |
| 11:43 |
sapier |
if you have mods installed in ~/.minetest and you start 0.4.6 0.4.7 you'll have a real mess |
| 11:43 |
kahrl |
obviously the directory should be the minetest dir instead of ~/.minetest when RUN_IN_PLACE=1 |
| 11:43 |
kahrl |
just as it always was |
| 11:43 |
VanessaE |
kahrl: right. |
| 11:43 |
kaeza |
sapier, there you have a point, yes |
| 11:44 |
VanessaE |
sapier: non-sequitur. if you have 0.4.6, you aren't gonna have 0.4.7 (etc), unless you're using run-in-place installs, in which case ~/.minetest isn't even used to begin with |
| 11:44 |
sapier |
I know for distributions we need ~ and share too |
| 11:44 |
sapier |
just because you don't check your mods for compatibility against latest stable version you can't expect everyone else not to do so too vanessae |
| 11:45 |
VanessaE |
sapier: all of my mods are written to be as forward- and backward-compatible as is practical, thank youi. |
| 11:45 |
VanessaE |
you're going off on the wrong direction yet again |
| 11:45 |
sapier |
so any solution is required to support al three opions share,~,RUN_IN_PLACE |
| 11:45 |
VanessaE |
if you have multiple versions of minetest installed, YOU AREN'T USING ~/.minetest ! |
| 11:46 |
VanessaE |
no |
| 11:46 |
sapier |
it is |
| 11:46 |
VanessaE |
system-wide mods belong in the system's package manager ONYL. |
| 11:46 |
VanessaE |
ONLY* |
| 11:46 |
sapier |
not same time but it needs to support all options |
| 11:46 |
VanessaE |
userland should NEVER have to (or be allowed) to screw with /usr/share or similar. |
| 11:46 |
sapier |
ok ... then forget about the manager |
| 11:47 |
kaeza |
sapier, RUN_IN_PLACE should not touch ~/.minetest or /usr/share |
| 11:47 |
VanessaE |
sapier: doing so requires root privs, you really want to build THAT into a GAME!? |
| 11:47 |
sapier |
this is way to complicated I'll leve it for someone else I've got already enough open issues |
| 11:47 |
kahrl |
it's not complicated. the mod manager just doesn't touch /usr at all |
| 11:47 |
VanessaE |
kahrl: exACTly. |
| 11:48 |
sapier |
that's not am option kahrl it at least needs to know what s installed in usr |
| 11:48 |
VanessaE |
sapier: you're trying to do more than a standard OS security model allows for. /usr/share is strictly under the purview of the OS's packaging system and/or update manager. |
| 11:48 |
sapier |
if not doing so you'll have conflicts |
| 11:48 |
kaeza |
sapier, *touch* (as in *modify*) |
| 11:49 |
kahrl |
yeah, I meant writing |
| 11:49 |
VanessaE |
(of course I include Windows there, for whatever its equivalent of /usr/share is) |
| 11:50 |
kaeza |
minetest does not use it on Windows |
| 11:50 |
VanessaE |
didn't think so |
| 11:50 |
sapier |
as I said getting this done in a way supporting all usefull and non usefull features currently implemented is more time than I'm willing to spend for this |
| 11:50 |
VanessaE |
sapier: then just don't DO that part, jeez |
| 11:50 |
VanessaE |
stick to run-in-place and ~/.minetest |
| 11:50 |
VanessaE |
anything else is secondary |
| 11:50 |
VanessaE |
you seem to have this mental block to "do this part now, add this other part later on" |
| 11:51 |
kaeza |
sapier, KISS |
| 11:51 |
kaeza |
(no offense) |
| 11:51 |
VanessaE |
if it's deemed to be SO important that you JUST HAVE TO read /usr/share (and /usr/local/share btw), then add those features later on down the line. |
| 11:51 |
kahrl |
if RUN_IN_PLACE; basepath='minetest dir'; else; basepath='~/.minetest'; end; function update(modname,modpath) if modpath is below basepath; do the update; else; print "Can't update mod; end; end |
| 11:51 |
sapier |
it's not an option to not consider something if you redesign it at least desing needs to consider it ... implementing it is something completely different |
| 11:52 |
VanessaE |
sapier: designs can be RE-designed over time if needed. |
| 11:52 |
VanessaE |
if they couldn't, we'd still be stuck with 0.3.x and everything hard-wired into the engine. |
| 11:52 |
VanessaE |
(sorry, darkrose :) ) |
| 11:53 |
sapier |
RE-design? good god do you realize how much fighting is required to get evel small design cleanups into minetest? |
| 11:53 |
sapier |
:-) |
| 11:53 |
kaeza |
sapier, "oh, I'll think about all the possible problems that may arise from my implementation in the future" -- said no developer EVER |
| 11:53 |
VanessaE |
sapier: do it anyway :) |
| 11:54 |
sapier |
ok lets stop general discussions and back to original topic mod/game management |
| 11:54 |
VanessaE |
sapier: why bother? you already said it's too complicated :) |
| 11:55 |
sapier |
you're really close to driving me really giving up this task vanessae ;-) |
| 11:55 |
sapier |
back to topic |
| 11:56 |
sapier |
forget about my suggestion for mod manager for time beeing |
| 11:56 |
sapier |
if you want gam centric view of mintest first thing to add is a game manager |
| 11:56 |
celeron55 |
what's this all hassle |
| 11:56 |
celeron55 |
supporting read-only global unixy installs and user installs is trivial |
| 11:57 |
sapier |
yes everything is trivial if you're ignoring conflict cases |
| 11:57 |
celeron55 |
the first one doesn't touch and the second one does, and RUN_IN_PLACE means that the former just doesn't exist |
| 11:57 |
kahrl |
sapier: what conflict cases? |
| 11:58 |
sapier |
e.g. a game of same name in /usr and in ~ |
| 11:58 |
VanessaE |
simple: the one in ~ or run-in-place gets used |
| 11:58 |
VanessaE |
and the one in /usr is ignored. |
| 11:58 |
VanessaE |
how hard is that? |
| 11:58 |
celeron55 |
those are always handled with the ~ one overriding the /usr one |
| 11:58 |
celeron55 |
or if they aren't they can be; it's standard practice everywhere anyway |
| 11:58 |
VanessaE |
exactly. |
| 11:59 |
celeron55 |
we aren't designing anything new really; just following long-used ways in how to do stuff |
| 11:59 |
sapier |
you need to tell user of override happening in order to not cause frustration because of things simply not working ... users do mistakes |
| 11:59 |
VanessaE |
sapier: if the user needs to be informed of that, it belongs at the OS level, in a readme somewhere. |
| 12:00 |
VanessaE |
as c55 just said, this is standard practice anyway |
| 12:00 |
sapier |
there are at least 5 places where you could install a mod atm |
| 12:00 |
iqualfragile |
i have an ide |
| 12:01 |
* kahrl |
has a sata |
| 12:01 |
iqualfragile |
what about having 2 (exactly 2) places where to put mods |
| 12:01 |
iqualfragile |
the one is system wide in /usr/share |
| 12:01 |
celeron55 |
we have at max 2 proper places where to install additional mods |
| 12:01 |
iqualfragile |
and the other one is somewhere in ~/.minetest |
| 12:01 |
celeron55 |
everything else is wrong |
| 12:01 |
iqualfragile |
then you could define games via a configuration file wich states the mods to include |
| 12:02 |
VanessaE |
iqualfragile: /usr/share and /usr/local/share if you even consider those folders at all. |
| 12:02 |
sapier |
I'm not talking about "proper" places as I realized noone is interested if it's correct once it's working |
| 12:02 |
celeron55 |
oh god |
| 12:03 |
celeron55 |
at this rate i'm going to seriously make a minimal voxel mining game which nobody will ever modify, including me after it works |
| 12:03 |
sapier |
if I'm not blamed for breaking something working atm done wrong I'm fine with it too |
| 12:03 |
iqualfragile |
right now you can put mods into the games directory, into the word directory and into the mods directory |
| 12:04 |
celeron55 |
sapier: one of minetest's greatest features is that it will interact properly with distro package management systems, as well as traditional windows packaging, as well as just .zips |
| 12:04 |
sapier |
and each of this option is available twice |
| 12:04 |
sapier |
I don't want to change anything about that fact celeron |
| 12:05 |
sapier |
I suggest dropping support for world and gamedir stored mods |
| 12:05 |
sapier |
gamedir != game |
| 12:07 |
celeron55 |
wtf is "gamedir" |
| 12:07 |
sapier |
the "mods" folder parallel to "bin" dir |
| 12:08 |
celeron55 |
ehm |
| 12:08 |
sapier |
where almost every user places it's mods first because of the "mods_in_here.txt" file in that folder |
| 12:08 |
celeron55 |
that is $path_user/mods |
| 12:09 |
sapier |
containing mods not linked to any game |
| 12:09 |
kaeza |
sapier, that's what world configuration is for |
| 12:09 |
kaeza |
(enable/disable mods per world) |
| 12:10 |
sapier |
do we really NEED this feature? |
| 12:10 |
celeron55 |
what feature |
| 12:10 |
sapier |
world specific mods |
| 12:10 |
kahrl |
I'm having trouble following this conversation |
| 12:10 |
celeron55 |
why is that a problem |
| 12:10 |
kaeza |
just download and install to mods/, disable the mod by default |
| 12:10 |
celeron55 |
my conclusion about this is that sapier has gone insane |
| 12:11 |
sapier |
I really don't have any idea how you do conlude celeron |
| 12:11 |
kaeza |
(or even conslidate somewhat the mod manager with the world config) |
| 12:11 |
sapier |
one simple question is a mod a part of a world or of a game or both? |
| 12:12 |
PilzAdam |
https://github.com/minetest/minetest/pull/716 <- tested and it works fine; any objections? |
| 12:12 |
sapier |
if it's both I don't have any idea how to create an easy to use way to manage mods |
| 12:12 |
celeron55 |
sapier: write a clear proposal that explains your vision for mod handling and handles all the use cases |
| 12:13 |
celeron55 |
sapier: nobody can follow and comment everything you say in here this way |
| 12:13 |
sapier |
I obviously don't even know of all use cases that are implemented currently |
| 12:13 |
celeron55 |
the dev wiki has a page for such |
| 12:13 |
sapier |
i stopped using wiki after not beeing able to decide cats from dogs for 10 minutes |
| 12:14 |
celeron55 |
just make an account |
| 12:14 |
celeron55 |
anyway, pastebin is equally fine |
| 12:16 |
celeron55 |
or maybe gist.github.com is best, i think it allows comments |
| 12:16 |
sapier |
I had an account and still couldn't change anything ;-) |
| 12:17 |
sapier |
I'll create a new suggestion and paste screenshots |
| 12:17 |
sapier |
obviously my first approach was too mod centric ... lets try to create a game centric one |
| 12:19 |
celeron55 |
well, it could be decided that should MT focus on mods per world or mods per game |
| 12:19 |
celeron55 |
there are good arguments towards both |
| 12:20 |
sapier |
yes but both at same time is crazy |
| 12:21 |
sapier |
still I may have an idea how to get this done at least a little bit userfriendly |
| 12:22 |
sapier |
my last suggestion was absed on plain mods ... another option would be hooking mod install to a selected game |
| 12:22 |
celeron55 |
the original design is having mods per game |
| 12:22 |
sapier |
a third option would be hooking it to a world, fetching mods from two sources |
| 12:23 |
celeron55 |
but the problem is, people tend to want to have different mods per different world |
| 12:23 |
celeron55 |
it can't be ignored |
| 12:23 |
sapier |
so mod manager should be even world specific instead of game specific |
| 12:23 |
celeron55 |
but also people want to automatically have all or part of their mods in newly created worlds |
| 12:24 |
PilzAdam |
the best for me would be world specific mods, i.e. place mods in $path_user/mods/ and enable them for worlds in the GUI |
| 12:24 |
sapier |
displaying mods currently active for a world isn't a problem |
| 12:24 |
PilzAdam |
(mods should be disabled by default for worlds) |
| 12:24 |
celeron55 |
this all is quite heavily reflected in our current implementation |
| 12:25 |
iqualfragile |
yeah, that would be nice (enabling mods) |
| 12:25 |
sapier |
we can't have everything at once we need at least to decide for a default behaviour |
| 12:26 |
celeron55 |
it seems to me that everything that we have currently would work just fine if mods would default as disabled for new worlds |
| 12:26 |
sapier |
enabling mods conflicts with with to make mods game specific |
| 12:26 |
kaeza |
sapier, the "standard" behavior for the user to install mods is to put them in $path_user/moods/minetest |
| 12:26 |
kaeza |
(as specified in the README) |
| 12:26 |
sapier |
but those mods will be active in all games am I correct? |
| 12:26 |
celeron55 |
sapier: decisions on whether mod namespace is game-specific or global is under discussion |
| 12:27 |
PilzAdam |
well, since we move back to minetest_game only I guess game specific namespace would work too |
| 12:27 |
sapier |
so designing a mod manager at present time is nonsense this decision has to be made prior creating a manageer |
| 12:27 |
PilzAdam |
(it was just annyoung for survival and build) |
| 12:27 |
kaeza |
sapier, yes, but disable by default (as PilzAdam and I said), and let the user enable them per-world |
| 12:27 |
PilzAdam |
*annoying |
| 12:28 |
celeron55 |
sapier: more like so that deciding that is part of the process of designing a mod manager |
| 12:28 |
sapier |
kaeza this would reuire games beeing at least hyprid config file + folders |
| 12:29 |
sapier |
as far as I understood if I'm creating the mod manager I'm not allowed to drop any feature. so how can I decide about the namespace without droping any feature? |
| 12:30 |
PilzAdam |
for the modmanage it would be best if the mod namespace would not be gamespecific |
| 12:30 |
PilzAdam |
+r |
| 12:30 |
sapier |
I don't have any problem with different oppinions but plz stop requesting conflicting things |
| 12:31 |
celeron55 |
sapier: this is not a light issue so it must be chosen carefully, so don't do it by yourself :P |
| 12:31 |
sapier |
pilzadam that's my current suggestion but vanessae doesn't agree to this at all |
| 12:31 |
* PilzAdam |
would be happy with both ways, although I slightly prefer the global namespace |
| 12:33 |
sapier |
as far as I see there are three options 1) "global mod namespace" 2) game specific namespace 3) world specific namespace |
| 12:33 |
sapier |
first one is most easy to implement and doesn't require lots of checks |
| 12:33 |
sapier |
second is something in between |
| 12:33 |
celeron55 |
my starting point here is that somebody will make a space travel / flight simulator / RTS game on the engine - and the mod system must not become a mess at that |
| 12:33 |
sapier |
and in third one we need to ask user to where to install a mod ... again |
| 12:34 |
celeron55 |
(none of those will use any mods from minetest_game) |
| 12:34 |
celeron55 |
(and thus will not be compatible in any way) |
| 12:36 |
PilzAdam |
the problem with the global namespace is that if somone makes a mod for minetest_game called "a", and another one makes a mod called "a" for a different game you cant install both |
| 12:36 |
sapier |
I'll create a suggestion the game specific way knowing that dropping world specific mod support is hard to accept for some ppl |
| 12:36 |
celeron55 |
of course there is the question of whether such is a reasonable goal, but i think that it would be extremely dumb to limit ourselves out of it |
| 12:36 |
PilzAdam |
but I think the chance for that to happen is very low |
| 12:36 |
celeron55 |
because we have no way of knowing what will happen in the future |
| 12:37 |
sapier |
ok we don't need to drop it but it'd just not be handled in in game mod manager correctly |
| 12:37 |
celeron55 |
sapier: it won't be handled; don't worry about that - it's for pre-made puzzle worlds and such |
| 12:37 |
celeron55 |
only thing it needs to do is do nothing if it detects worldmods |
| 12:39 |
celeron55 |
PilzAdam: it can't be left in the hands of chance - if the namespace is combined, the mention about that must be added to modding tutorials and such so that people won't even make a mod of the same name for two different games |
| 12:39 |
|
BlockMen joined #minetest-dev |
| 12:39 |
sapier |
ingame mod manager won't even detect world mods |
| 12:40 |
sapier |
the game specific way will only handle mods in games |
| 12:40 |
celeron55 |
sapier: it must so that it won't screw up such pre-made worlds; it's a matter of checking existence a single directory so it can be easily done |
| 12:40 |
celeron55 |
+of |
| 12:40 |
BlockMen |
hi, i have noticed, that "max_hear_distance = w/e" only works when starting to play the sound. but when it is playing and the player moves away the sound still plays. is that wanted? |
| 12:41 |
sapier |
celeron55 it can't starting at a game I can't check for all worlds and if there might be a mod in there |
| 12:41 |
celeron55 |
sapier: eh what |
| 12:41 |
celeron55 |
sapier: now you are again just deliberately not understanding |
| 12:41 |
celeron55 |
stop it |
| 12:41 |
celeron55 |
it's dumb |
| 12:41 |
sapier |
you said mod manager should handle world mods |
| 12:42 |
sapier |
the game specific mod manager wouldn't even have an idea what a "world" is |
| 12:42 |
celeron55 |
it must *not interfere* with them, by doing exactly nothing and letting the world simply run with the default mods of a game combined with the worldmods |
| 12:42 |
sapier |
i think we're talking of different things |
| 12:42 |
celeron55 |
okay, you can't do that |
| 12:43 |
|
ImQ009 joined #minetest-dev |
| 12:43 |
celeron55 |
it has been in discussions that eg. deciding between a survival world and a creative world would be done by the means of enabling a survival or creative focused mod on a world |
| 12:43 |
celeron55 |
or other stuff like that |
| 12:43 |
celeron55 |
so having it per-game is really out of the question |
| 12:44 |
celeron55 |
(and yes, we talked about different things) |
| 12:44 |
sapier |
the more we're talking the more I think the modmanager containing all features required by different ppl just can't ever be developed |
| 12:45 |
sapier |
maybe I should create a modmanager guessing who's currently playing and enabling exactly this feature set /ironic/ |
| 12:45 |
kaeza |
sapier, then set a new standard |
| 12:45 |
PilzAdam |
sapier, the only thing the modmanager has to do is download/install/remove mods into/from $path_user/mods/ |
| 12:45 |
celeron55 |
and my SSH connection sucks horse dick |
| 12:45 |
celeron55 |
this makes me three times more annoyed |
| 12:45 |
PilzAdam |
(in a game namepsace or not, depends how we decide) |
| 12:46 |
sapier |
pilzadam that's YOUR feature set |
| 12:46 |
PilzAdam |
what do other people expect from it? |
| 12:46 |
kaeza |
sapier, it's the general user "feature set" |
| 12:46 |
sapier |
for vanessaE I'll create a gui showing games only |
| 12:46 |
iqualfragile |
i.e. worldmods > gamemods and the modmanager ignores worldmods? |
| 12:47 |
sapier |
and for celeron one supporting decision of download place based on some complex calculations where waht mod should be added |
| 12:47 |
celeron55 |
fuck this connection |
| 12:47 |
celeron55 |
now let me propose some crap if this connection might happen to let me |
| 12:47 |
sapier |
iqualfragile that suggestion has just been denied by celeron |
| 12:47 |
PilzAdam |
sapier, mods by the mod manager should always go into $path_user/mods/ |
| 12:48 |
iqualfragile |
yeah, im sorry, i have not scrolled down :D |
| 12:48 |
PilzAdam |
and its a modmanager, not a gamemanager |
| 12:48 |
sapier |
pilzadam doing so games need some way to enable/disable mods |
| 12:48 |
PilzAdam |
no, the user decides wich mods to use in a per world setting |
| 12:48 |
kaeza |
sapier, the mod enable/disable mechanism will come in handy here |
| 12:48 |
celeron55 |
actually no i won't, it won't work |
| 12:49 |
kaeza |
(per-world) |
| 12:49 |
sapier |
might be afouth option |
| 12:49 |
iqualfragile |
maybee we should get our definitions straight: |
| 12:50 |
sapier |
"global" mods handled by mod manager enabled at world creation |
| 12:50 |
iqualfragile |
game: some completly independent set of nodes and scripts |
| 12:50 |
celeron55 |
what if we drop having game selection in the UI and define a scheme of how games can start up themselves in minetest by their own shortcuts or bundles? |
| 12:50 |
iqualfragile |
mod: a modifictaion to a game |
| 12:50 |
celeron55 |
it will give more space for mod selection |
| 12:50 |
|
hmmmm joined #minetest-dev |
| 12:50 |
iqualfragile |
i think the question is: should a mod be usable in more then one game? so: are mods game-specific or not? |
| 12:51 |
iqualfragile |
thats a good idea, actualy |
| 12:51 |
celeron55 |
iqualfragile: the answer is, they can be, but in the future many may not be |
| 12:51 |
|
Zeg9 joined #minetest-dev |
| 12:51 |
sapier |
a mod can be even world specific iqualfragile |
| 12:52 |
iqualfragile |
sapier: yeah, but a world allwas has a gamemode |
| 12:53 |
sapier |
there are too many options to handle all at once in an easy modmanager |
| 12:53 |
iqualfragile |
celeron55: because right now most games are quite similar and a bit interchangable (like minetest_game) but in future people might create crazy stuff on the minetest engine |
| 12:53 |
celeron55 |
iqualfragile: that is the idea, and the source of all confusion and problems |
| 12:53 |
celeron55 |
8) |
| 12:53 |
celeron55 |
of course we could just drop the "gameability" and be done with that |
| 12:54 |
celeron55 |
but then most of our framework is useless |
| 12:54 |
iqualfragile |
well, there are two solutions in my opinion: either we ignore the posibilities and just say: minetest is a gameengine for minetest_game like games |
| 12:54 |
celeron55 |
well, not most, but some parts |
| 12:54 |
BlockMen |
what about user-mods are global in general, but for custom games deisabled by default? |
| 12:54 |
iqualfragile |
and all mods are put into one folder and every gamemode enables a subset of those per config file |
| 12:54 |
BlockMen |
*disabled |
| 12:54 |
iqualfragile |
and that selection can be extended by world-specific configuration |
| 12:55 |
iqualfragile |
or we say: minetest is a general gameengine |
| 12:55 |
iqualfragile |
wich would result in putting mods in gamespecific directories |
| 12:55 |
celeron55 |
this all comes to what i have said before: minetest as an engine must work more like a library or a script interpreter; not as the game-engine hybrid it is |
| 12:56 |
celeron55 |
if the goal is to reduce confusion, that is |
| 12:56 |
celeron55 |
which sapier seems to aim towards |
| 12:56 |
kaeza |
game-specific mod install would require downloading and installing one copy per game mode |
| 12:56 |
PilzAdam |
my ideas: https://gist.github.com/PilzAdam/5554242 |
| 12:56 |
sapier |
yes this is my aim |
| 12:56 |
iqualfragile |
ok, so you favour the sec. possibility? |
| 12:56 |
sapier |
-aim + target |
| 12:56 |
kaeza |
PilzAdam, +1 |
| 12:57 |
celeron55 |
and there are living examples of such engines like löve2d |
| 12:57 |
iqualfragile |
kaeza: yeah and that would be absolutely sane if we assume that gamemodes are compleetly independent |
| 12:57 |
iqualfragile |
but they are not atm |
| 12:57 |
kaeza |
iqualfragile, some games are meant to be stand-alone (f.ex: Dwarves) |
| 12:57 |
sapier |
pilzadam if you create a new world you'll always have to enable lets say 30 mods |
| 12:57 |
Zeg9 |
PilzAdam, I think the same |
| 12:57 |
PilzAdam |
sapier, yes |
| 12:57 |
BlockMen |
PilzAdam, +1 (but the user chosen mods should be enabled by default for the default games) |
| 12:57 |
sapier |
this is not what I try to achieve |
| 12:58 |
PilzAdam |
well, other way round I have to disable 30 mods |
| 12:58 |
PilzAdam |
its just dont mess up existing worlds |
| 12:58 |
sapier |
it's to easy to click wrong mod and can't really disable it afterwards |
| 12:58 |
PilzAdam |
if you intsall a mod you first have to tell a world to use the new mod |
| 12:58 |
kaeza |
and it's easier to leave it by default |
| 12:58 |
sapier |
there should at least be some "templates/games" to use for world creation |
| 12:58 |
iqualfragile |
maybee we should force mods to have a remove.lua file wich aliases the registered nodes to some default nodes |
| 12:58 |
kaeza |
(and not be able to disable it afterwards) |
| 12:59 |
kaeza |
iqualfragile, umm |
| 12:59 |
PilzAdam |
sapier, well, if you accidentally use a mod in a world at startup, just remove the world and create a new one |
| 12:59 |
PilzAdam |
currently if you install a mod it would go directly into every world |
| 13:00 |
sapier |
can be done of course but clicking 30 mods is still anoying |
| 13:00 |
PilzAdam |
this will mess up worlds a lot more than disabling every mod by default |
| 13:00 |
kaeza |
^ including existing ones |
| 13:00 |
celeron55 |
sapier: just add an "enable all mods" button |
| 13:00 |
PilzAdam |
sapier, there is "Select all" |
| 13:00 |
celeron55 |
what's the problem with that |
| 13:00 |
PilzAdam |
also: modpacks |
| 13:00 |
celeron55 |
anyway, i'm ready to go along PilzAdam's suggestion |
| 13:01 |
sapier |
select all doesn't help if you want a subset |
| 13:01 |
celeron55 |
it's the KISS way and i'm tired of trying to come up with better ones |
| 13:01 |
PilzAdam |
users can make their own "Survival" and "build" modpacks to easily enable/disable mods in their worlds |
| 13:01 |
PilzAdam |
(or the mod manager can create them for them) |
| 13:01 |
celeron55 |
sapier: that's more of a UI problem anyway; but what about a button "use mod set from an existing world" that gives a list of worlds from where to copy the mod configuration |
| 13:02 |
kaeza |
or add a "Enable by default" checkbox in the mod manager |
| 13:02 |
PilzAdam |
kaeza, i.e. "Select all" |
| 13:02 |
kaeza |
PilzAdam, per-mod |
| 13:02 |
PilzAdam |
nah |
| 13:02 |
PilzAdam |
that would screw up the concept |
| 13:02 |
kaeza |
so when you create a new world, it will enable only those selected by default |
| 13:02 |
sapier |
is vanessae still here? |
| 13:03 |
sapier |
she's been the one denying the mod centric approach currently discussed |
| 13:03 |
kaeza |
(of course, the "Select all" button would be good too) |
| 13:03 |
celeron55 |
sapier: we simply won't care; she doesn't have better alternatives anyway |
| 13:04 |
sapier |
http://ompldr.org/iaWQyeA this is the suggestion fitting for a mod centric mod manager |
| 13:04 |
sapier |
second part needs to be added in world creation |
| 13:06 |
sapier |
is this what you've been thinking about pilzadam/celeron55? |
| 13:06 |
PilzAdam |
the mod manger needs to be discussed later on |
| 13:06 |
PilzAdam |
but it will be relatively simple with my system |
| 13:06 |
sapier |
ok lets get back to beginning |
| 13:07 |
sapier |
guys I'm gonna leave now to prevent saying something I'm gonna regret |
| 13:07 |
|
sapier left #minetest-dev |
| 13:07 |
kaeza |
? |
| 13:09 |
Zeg9 |
huh? |
| 13:09 |
celeron55 |
lol i don't even care, whatever that is |
| 13:09 |
iqualfragile |
i guess he is slightly annoyed |
| 13:09 |
iqualfragile |
so whats the decision? all mods into one folder and enabling them per world? |
| 13:10 |
PilzAdam |
yes |
| 13:10 |
celeron55 |
it seems to me that we don't have any other viable options |
| 13:10 |
PilzAdam |
(see my gist) |
| 13:10 |
kaeza |
looks simple enough |
| 13:11 |
kaeza |
I don't know why sapier is always finding unrelated problems or looking way too much ahead |
| 13:11 |
iqualfragile |
the other option would be to keep it like its atm: |
| 13:11 |
iqualfragile |
gamemodes provide sets of mods |
| 13:12 |
kaeza |
"problems"* |
| 13:12 |
celeron55 |
it's simple, but we may end up having to put in place pretty strict naming conventions after a long time, breaking it |
| 13:12 |
celeron55 |
... |
| 13:12 |
celeron55 |
fucking SSH |
| 13:12 |
iqualfragile |
people extend gamemodes by putting mods into a gamemode-specific folder |
| 13:12 |
iqualfragile |
and that can be extended by putting mods into worlds |
| 13:12 |
celeron55 |
it's like sapier thought that discussion was in any way enjoyable to others |
| 13:13 |
celeron55 |
discussing about this sucks because everyone has different usage and sapier tries to go forward way too fast |
| 13:14 |
celeron55 |
iqualfragile: the problem is, sticking with the current way complicates all mod-related tools |
| 13:14 |
iqualfragile |
i, too think that pilzadams aproach would be the best idea for now |
| 13:14 |
celeron55 |
and especially the UI |
| 13:15 |
iqualfragile |
as most gamemodes are quite simmilar |
| 13:16 |
iqualfragile |
we should reserve the modname "local" so serverowners can create custom extensions and are guaranteed to never run into collisions |
| 13:20 |
celeron55 |
i don't think such is an issue at all |
| 13:20 |
celeron55 |
it could be a recommended name for sure though |
| 13:22 |
kaeza |
"Games have nothing to do with mods; the user (and dependencies) has to decide if a mod fits to a game or not." |
| 13:22 |
kaeza |
I have a possible solution to this |
| 13:22 |
kaeza |
maybe add a "standalone =" setting to game.conf |
| 13:22 |
iqualfragile |
? |
| 13:23 |
kaeza |
if it's true, only mods within the game dir will be loaded |
| 13:23 |
iqualfragile |
what should that signalize? dont put mods in here? |
| 13:23 |
iqualfragile |
i dont think thats a good idea |
| 13:25 |
kaeza |
ok, just a brainfart :P |
| 13:26 |
celeron55 |
while 0.4's design is quite intensely solid compared to what was needed at the time it was designed, these days i'm hoping it would've been designed to stand even further to the future... |
| 13:26 |
celeron55 |
seriously, all that we are building on was completely overkill 1.5 years ago, but now it's not enough to anyone |
| 13:26 |
|
BlockMen left #minetest-dev |
| 13:27 |
celeron55 |
or maybe i'm fooled by the naysayers once again |
| 13:27 |
celeron55 |
it's hard to tell |
| 13:27 |
kaeza |
minetest has shifted from being a "game" to being an "engine" |
| 13:27 |
celeron55 |
naysayers should be illegal, they do nothing good to anything |
| 13:28 |
celeron55 |
kaeza: it was well-known back then; it's starting to be more like a monster than an engine now 8) |
| 13:28 |
RealBadAngel |
hi folks, whats up? |
| 13:28 |
kaeza |
so perhaps in the near future you could develop RPGs, FPSs, and things like that on top of it (as you said) |
| 13:29 |
celeron55 |
RealBadAngel: the million kilometer long backlog is up there |
| 13:29 |
RealBadAngel |
hehe, i will scroll and read |
| 13:29 |
RealBadAngel |
just wanted to say hello |
| 13:29 |
kaeza |
hey RBA |
| 13:29 |
kaeza |
:) |
| 13:29 |
celeron55 |
i don't recommend reading it, it's painful |
| 13:30 |
iqualfragile |
hi RealBadAngel |
| 13:30 |
celeron55 |
kaeza: the plan of 0.4 is to create a platform on which every corner of the genre set up by MC can be... tested |
| 13:30 |
iqualfragile |
and i think thats enought |
| 13:30 |
kaeza |
if 3rd person view ever gets implemented, you could do a mario-like platformer in it :P |
| 13:30 |
iqualfragile |
KISS |
| 13:31 |
iqualfragile |
there is no need for minetest to be a monstrous game engine where you can create every thinkable game on |
| 13:31 |
iqualfragile |
there are lots of other engines for other games |
| 13:31 |
celeron55 |
it's just that MT combines that with server-side modding which complicates things |
| 13:31 |
celeron55 |
so it divides and confuses people, and makes things more tedious to implement |
| 13:32 |
celeron55 |
basically MT doesn't really know what it is, but it certainly has many likeable properties |
| 13:33 |
RealBadAngel |
why painful? again renaming engine or something similar? |
| 13:33 |
iqualfragile |
jup |
| 13:33 |
iqualfragile |
i allways say that in minetest the game is created by the player (or the serveradmin) |
| 13:33 |
iqualfragile |
and not the developers |
| 13:34 |
RealBadAngel |
developers? never heard of them |
| 13:34 |
RealBadAngel |
we just code here |
| 13:35 |
iqualfragile |
thats why i did not put game designers there :D |
| 13:35 |
iqualfragile |
afk |
| 13:39 |
RealBadAngel |
http://i.imgur.com/KrqC9RM.png |
| 13:40 |
RealBadAngel |
i just learned how to properly UV map spheres in blender :) |
| 13:40 |
kaeza |
RealBadAngel, nice |
| 13:42 |
RealBadAngel |
blender is another app very hard to master, as gimp before |
| 13:42 |
RealBadAngel |
but once you get how the things work, you can do miracles |
| 14:32 |
|
kaeza joined #minetest-dev |
| 14:55 |
PilzAdam |
RealBadAngel, if you dont wanna read logs: https://gist.github.com/PilzAdam/5554242 this pretty much sums the result up |
| 15:06 |
|
Jordach joined #minetest-dev |
| 15:11 |
|
kaeza1 joined #minetest-dev |
| 15:36 |
|
rubenwardy joined #minetest-dev |
| 15:43 |
|
Calinou joined #minetest-dev |
| 16:00 |
|
dexter0 joined #minetest-dev |
| 16:07 |
|
rubenwardy left #minetest-dev |
| 16:07 |
|
rubenwardy_ joined #minetest-dev |
| 16:09 |
|
emptty joined #minetest-dev |
| 16:14 |
rubenwardy |
game filter is so ugly |
| 16:14 |
PilzAdam |
it should be only visible in singleplayer |
| 16:15 |
rubenwardy |
and the texture does not wrap around |
| 16:15 |
rubenwardy |
I think that it should be more like a context menu |
| 16:15 |
rubenwardy |
with a dropdown with over game types |
| 16:21 |
PilzAdam |
also I dont get why would change the background image |
| 16:23 |
rubenwardy |
I quite like that as it is kind of skinning, but it might get annoying |
| 16:30 |
VanessaE |
*looks at PilzAdam's proposal* ... isn't that what I said 5 hours ago? :) |
| 16:30 |
VanessaE |
*returns to scrollback* |
| 16:33 |
VanessaE |
ok, done. |
| 16:34 |
VanessaE |
sorry for the delay, had a technician over working on something. |
| 16:34 |
Calinou |
technician ö_ö |
| 16:34 |
Calinou |
that's pejorative |
| 16:34 |
Jordach |
did he use a mining drill? |
| 16:34 |
VanessaE |
my only contention with a mod manager is that it shouldn't try to overcomplicate shit. |
| 16:34 |
Calinou |
^ |
| 16:35 |
VanessaE |
it should manage exactly one thing: loose mods/modpacks in whatever directories they go in. Games should not be dealt with at all. |
| 16:35 |
VanessaE |
exactly as PilzAdam suggested. |
| 16:36 |
VanessaE |
<offtopic> (and if you MUST know, the technician was installing a powered stair lift for my husband) </offtopic> |
| 16:36 |
ironzorg |
somebody should stop merging requests and fix all those warnings |
| 16:36 |
Jordach |
rubenwardy, nick spam must stop |
| 16:36 |
rubenwardy |
I am making a hexchat script |
| 16:37 |
rubenwardy |
to automatically ghost, nick, identify |
| 16:37 |
rubenwardy |
I will exit all minetest channels |
| 16:37 |
|
rubenwardy left #minetest-dev |
| 16:52 |
|
celeron55 joined #minetest-dev |
| 16:56 |
|
rubenwardy joined #minetest-dev |
| 17:00 |
|
doserj joined #minetest-dev |
| 17:01 |
|
doserj left #minetest-dev |
| 17:15 |
|
doserj joined #minetest-dev |
| 17:15 |
|
doserj left #minetest-dev |
| 17:17 |
|
ssieb joined #minetest-dev |
| 17:22 |
|
sapier joined #minetest-dev |
| 17:22 |
sapier |
https://github.com/minetest/minetest/pull/677 any (additional) comments? |
| 17:24 |
sapier |
plz tell me whats still wrong with it if it can be fixed I'll fix it if it can't be fixed I'll drop this pull request |
| 17:24 |
PilzAdam |
sapier, about your sky leak fix: how does the server use the sky? |
| 17:25 |
sapier |
I've already closed it as I can recreate the memory leak it once triggered |
| 17:25 |
PilzAdam |
it seems to work fine if I just drop the sky at game.cpp:3337 |
| 17:25 |
sapier |
if this leak is currently only hidden I'll fix it once it reapears |
| 17:26 |
sapier |
if I remember correctly there was some issue with quick select menu using sky when rendering for some unknown reason |
| 17:26 |
sapier |
but I'm not sure about it |
| 17:27 |
|
celeron55 joined #minetest-dev |
| 17:28 |
sapier |
btw pilzadam you don't need any cleanup mod with mobf as mobf has its own cleanup capabilities |
| 17:28 |
PilzAdam |
see my comment in the forum |
| 17:30 |
sapier |
he did ask about removing vombies he didn't say everything ... for this usecase built in cleanup is best ... and to tell with vanessae's words storage is cheap just because there's sitting around lots of dead code this doesn't mean there's any performance impact with disabled mobf |
| 17:31 |
rubenwardy |
Would it not be possible to, on block load, check if the node is defined, and if it is not then delete it? |
| 17:31 |
sapier |
any real results on mod discussion after I left? |
| 17:31 |
rubenwardy |
I know it would be slower than an abm |
| 17:32 |
PilzAdam |
sapier, so, this could cause problems: https://github.com/PilzAdam/minetest/commit/782d06be460e81c9fe4760c78003ac6522676e82 ? |
| 17:32 |
sapier |
this has sideeffects too rubenwardy |
| 17:34 |
sapier |
if it doesn't result in double free in your case pilzadam it most likely will be fine. Maybe the real bug wasn't sky beeing deleted to early but some other things beeing done to late |
| 17:35 |
sapier |
rubenwardy e.g. if you accidently disable mobf you'll never get some mobs in spawned worlds again if any entity is deleted immediatly |
| 17:35 |
sapier |
and no I can't fix it because if I do everybody starts screaming "its so slow" again ;-) |
| 17:35 |
rubenwardy |
maybe you would have to activate it by going /clean |
| 17:36 |
sapier |
currently if you have set mobf to cleanup all disabled mobs will be deleted on activation |
| 17:36 |
sapier |
cleanup is not default setting to avoid mentioned problem |
| 17:37 |
sapier |
if you have any idea how to make this even more clear plz tell me |
| 17:38 |
sapier |
currently ppl seam to be so used to bad mod design they can't even imagine someone giving them possibility to cleanup correctly |
| 17:39 |
PilzAdam |
sapier, any idea how to trigger the double free bug (if its possible with this code at all) |
| 17:41 |
sapier |
the one I intended to fix happened on any shutdown after opening inventory (everything within valgrind) |
| 17:41 |
sapier |
but I couldn't reproduce it with current master |
| 17:42 |
PilzAdam |
well, the one I wanted to fix is: https://gist.github.com/PilzAdam/5556066 |
| 17:42 |
PilzAdam |
and this appears always when shutting down |
| 17:43 |
sapier |
thats not related to sky |
| 17:44 |
PilzAdam |
hm? |
| 17:45 |
sapier |
did you check if guitext_status is cleaned up? |
| 17:45 |
PilzAdam |
what? where? |
| 17:46 |
PilzAdam |
you mean this line is not related to sky? https://github.com/minetest/minetest/blob/master/src/game.cpp#L1262 |
| 17:46 |
sapier |
opps of by one line |
| 17:46 |
sapier |
codeline |
| 17:46 |
sapier |
oops? this code is completely different to what I was looking at |
| 17:47 |
sapier |
ok yes thats the issue I tried to fix by dropping it |
| 17:47 |
sapier |
but I didn't see it last time ... maybe I just overlooked |
| 17:48 |
sapier |
if you want to fix this you have to move pointer out of try block to have it available lateron |
| 17:48 |
sapier |
you can't delete it within the try block without causing problems |
| 17:49 |
PilzAdam |
that was my question: what problems and how to trigger them? |
| 17:50 |
sapier |
just closing a game |
| 17:50 |
sapier |
not directly returning to game |
| 17:50 |
sapier |
to os |
| 17:51 |
sapier |
problem was access of already freed memory |
| 17:52 |
PilzAdam |
that works fine |
| 17:53 |
sapier |
ok so the problem is gone |
| 18:27 |
|
ImQ009 joined #minetest-dev |
| 18:37 |
|
mrtux joined #minetest-dev |
| 18:45 |
|
Calinou joined #minetest-dev |
| 18:56 |
|
kaeza joined #minetest-dev |
| 19:24 |
|
ImQ009 joined #minetest-dev |
| 22:16 |
|
kaeza joined #minetest-dev |
| 22:44 |
|
ecube joined #minetest-dev |
| 22:53 |
PilzAdam |
sapier, the NodeDefManger in run_tests() leaks a lot of memory, do you have any idea why? |
| 22:54 |
sapier |
no I don't I've already fixed some leaks in tests but as they didn't care about cleaning up anything I'm not surprised there is more |
| 22:55 |
|
KrisEike joined #minetest-dev |
| 22:56 |
|
KrisEike left #minetest-dev |
| 23:07 |
|
BrandonReeseS3 joined #minetest-dev |
| 23:42 |
|
iqualfragile1 joined #minetest-dev |
