Minetest logo

IRC log for #minetest-dev, 2016-10-31

| Channels | #minetest-dev index | Today | | Google Search | Plaintext

All times shown according to UTC.

Time Nick Message
00:02 Tmanyo joined #minetest-dev
00:07 Fixer thats why minetest-dev in dissaray, someone wants to merge and fix, someone wants nuclear reactor grade software %)
00:08 Fixer it is so toxic right now
00:14 Hijiri nuclear reactor grade software probably would need more static verification
00:14 Hijiri anyway daily pls review post: #4685
00:14 ShadowBot https://github.com/minetest/minetest/issues/4685 -- Add control information to player interacts by raymoo
00:14 Hijiri don't let the uncertainty of drama prevent you from reviewing
00:16 Fixer haha, good luck with that
00:16 Fixer there was 3 pages of hate above
00:16 Hijiri I was here
00:16 Hijiri was afk but skimmed it
00:21 Fixer "oh shit we broke something because we didn't talk it over", "that's okay we have another commit" /// but why not? it is not like someone died.. I play another game called CDDA which has exactly this attitude and it is not only playable and gives you lots of fun, game makes huge progress between releases, while minetest adds (?) carts after 5 years of waiting %)
00:21 Fixer nobody is perfect anyway, including coders
00:22 Fixer and one thing i dislike about this debate
00:23 Fixer people bitching about some obscure refactoring, while gameplay or graphics or water or whatever bugs are left out for years for unknown reasons, no drama about that stuff :(
00:24 Fixer maybe coding requirements are too high for minetest and it is holding development? people afraid to even code smth
00:33 Fixer it is just a block game... don't ruin it, be good to each other, you will not loose job because of this, nobody will execute you, commit a fix and move... with current attitute minetest in 10 years will hardly move to anything...
00:33 Fixer good night
00:49 rubenwardy code review is very important though
00:50 rubenwardy oh, he went
01:01 paramat joined #minetest-dev
01:02 paramat fixer talking rubbish again
01:07 paramat #4686 seems good now
01:07 ShadowBot https://github.com/minetest/minetest/issues/4686 -- Remove optimization that causes underwater and cave rendering glitches. by lhofhansl
01:15 STHGOM joined #minetest-dev
01:30 turtleman joined #minetest-dev
01:57 paramat any more approval for 4686?
02:30 fling joined #minetest-dev
02:41 Zeno` joined #minetest-dev
03:06 ptv joined #minetest-dev
03:17 hiradur joined #minetest-dev
03:43 paramat game#1358
03:43 ShadowBot https://github.com/minetest/minetest_game/issues/1358 -- Add a telescope item that enables zoom privilege
03:48 paramat good, 4686 is in :]
03:53 hmmmmm is this the same zoom that kaeza (???) made a PR for years ago and not enough people liked it to get merged?
03:53 hmmmmm oh nvm it is in
04:19 Foghrye4_ joined #minetest-dev
04:21 paramat it was eventually merged in another form, it was made disabled by default and only allowed through privilege
04:32 hmmmmm yeah that sounds like a good compromise
04:40 sofar visually correct rendering has my preference, so +1 for 4686
04:40 sofar so many good visual changes in the past few days now, wow
04:45 paramat yes lots of good stuff all around
04:48 hmmmmm good stuff, but, there's one thing i really don't like, "Added my own approval"
04:51 paramat heh i noticed :]
04:51 hmmmmm it's a blatant attempt to subvert the whole point behind peer review
04:52 hmmmmm sure, you can have your own approval in the sense that you feel confident in the code you write, but that doesn't count for any points
04:53 hmmmmm but on the other hand... if you don't feel confident about your own code, what kind of message does that send to the rest of us?
04:54 paramat the current practice by most is that an author's own contentment doesn't count as an 'approval' label
04:54 hmmmmm i noticed a lot of people like to strawman the push for code quality, comparing it to "nuclear reactor code" and other similar things
04:55 hmmmmm there's a huuuuge gamut of code quality
04:55 paramat oh fixer was doing the negativity thing, talking rubbish again
04:55 hmmmmm we're nowhere near nuclear reactor code quality
04:55 hmmmmm what i'm hoping for is a level of quality such that a new commit does not add new bugs
04:55 paramat yeah the required code quality is barely high enough as it is
04:56 hmmmmm because that's what a lot of these things tend to do
04:56 hmmmmm shuffle around the old bugs to make new ones
04:56 hmmmmm how much manpower and effort is put toward solving bugs that exist already?
04:57 hmmmmm how much does it upset the end users that there are so many bugs?
04:58 hmmmmm the thing is, the code review stage often misses many flaws
04:59 hmmmmm even if there are multiple reviewers and they're all extremely careful
04:59 hmmmmm the earlier a bug is seen, the lower the cost of remediation
05:00 paramat sofar is setting a higher quality requirement in mtgame, much appreciated
05:00 proller__ joined #minetest-dev
05:02 hmmmmm the thing about code reviews
05:03 hmmmmm i notice a lot of people just visually scan over it and note things that pop out to them, such as a buffer overrun or missing check
05:03 hmmmmm rarely do reviewers build up the mental state of the entire PR and make note of logical problems the changes introduce
05:05 hmmmmm i've personally noticed that with a group of decently competent coders, the great majority of bug types are logic errors, not using the wrong function or indexing an array out of bounds or overflowing a stack
05:05 hmmmmm it's mentally taxing to do this
05:05 hmmmmm i don't blame more reviewers for not
05:06 hmmmmm but at the same time it's perhaps an incentive to push harder for real unit tests that verify behavior on the smooth path as well as various edge cases
05:06 hmmmmm also it's a reason to demand easy to understand code
05:07 hmmmmm which is a big part of the reason why the shader setting cache PR is being shelved for the immediate time being
05:07 hmmmmm it's just far too difficult to verify the behavior as being correct with the many levels of indirection
05:08 hmmmmm so my point is - having a single other person "review" the code by scanning it over and looking for obvious errors is the bare minimum of a review.  you could probably get the same or better results from a static code analyzer
05:09 hmmmmm that's why two people other than yourself looking at it is considered the minimum for a change of moderate size and moderate complexity
05:09 hmmmmm and those code reviews (plural) themselves shouldn't be enough to feel confident in the code being considered, unit tests should also be there
05:10 hmmmmm the only reason why i haven't been harping on automated testing much more is because it's difficult to test
05:10 hmmmmm nothing is broken out into discrete units and refactoring it to do so would simply be a massive undertaking
05:24 gregorycu joined #minetest-dev
05:25 gregorycu I don't exactly think it's fair to characterise Fixer's points as rubbish or being negative
05:27 paramat many in the last few hours have been
05:27 gregorycu He's been around a while, and often the PRs he cares about fall by the wayside (for whatever reason)
05:29 gregorycu There are open source projects with more stringent requirements for merging, and also those with less stringent requirements. He thinks minetest is too stringent .
05:35 sofar We all have different values, and they just conflict
05:39 paramat 'nobody fixes bugs' 'nobody cares about bugs' 'no drama about bugs' 'people afraid to even code something' fixer often does these ridiculous negative rants and exaggerates for effect, it's ok i'm used to it now, fixer is also a valued contributor :]
05:40 gregorycu I think they are true though
05:41 gregorycu Maybe a slight exaggeration
05:41 gregorycu And most of it is a function that minetest is an open source project
05:42 paramat we worrk a
05:43 paramat our asses off worrying about bugs, fixing them, discussing them. the actual issue is lack of dev time, which is what fixer needs to understand
05:43 paramat oops
05:44 gregorycu Lack of dev time is AN issue, agree
05:44 gregorycu I'm sure he understands it
05:47 paramat yeah, it's just not expressed during a rant
05:47 gregorycu Yeah
05:51 jin_xi joined #minetest-dev
05:57 Hunterz joined #minetest-dev
06:06 nrzkt joined #minetest-dev
06:43 CWz joined #minetest-dev
07:14 kaeza joined #minetest-dev
07:54 CWz_ joined #minetest-dev
08:15 paramat added docs to #4699 and will merge later
08:15 ShadowBot https://github.com/minetest/minetest/issues/4699 -- Lua voxelmanip: Add optional buffer param for 'get param2 data' by paramat
08:31 paramat tested too
08:53 Hunterz joined #minetest-dev
09:23 red-001 joined #minetest-dev
09:34 paramat joined #minetest-dev
10:16 est31 I agree with hmmm that reviews should be higher quality, but we don't have enough devs to do it
10:16 est31 the best way hmmm can make review quality go up is to review himself
10:17 est31 you know, because i rather introduce a bug than let minetest drown in 200 open PRs
10:17 est31 and demote potential contributors along the way
10:24 paramat yes, so an author's happiness with their PR counts as one of the 2 core dev agreements required
10:27 paramat (er as long as they are a core dev)
10:27 paramat that is also the minimum acceptable amount of review
10:43 Fixer joined #minetest-dev
10:50 paramat will merge #4681 #4699 in a moment
10:50 ShadowBot https://github.com/minetest/minetest/issues/4681 -- Replace call to obsolete get_look_pitch(). by sofar
10:50 ShadowBot https://github.com/minetest/minetest/issues/4699 -- Lua voxelmanip: Add optional buffer param for 'get param2 data' by paramat
10:52 nrzkt paramat, est31 #4155 is now ready and lighter (i removed the DB backend atm as it's not needed for the part 1 of this PR, this permit easy review/integration)
10:52 ShadowBot https://github.com/minetest/minetest/issues/4155 -- Implement player attribute backend by nerzhul
11:02 paramat now merging
11:02 est31 file backends are broken
11:02 est31 we should use databases, that's what they are made for
11:03 est31 many minetest admins regularly have to fix broken player files
11:03 est31 plain text files are fine AS LONG AS WE ARE NOT WRITING TO THEM
11:05 paramat merge complete
11:05 nrzkt est31, i agree with you for that
11:05 nrzkt but i should do the PR in two phasis atm
11:06 nrzkt because the db backend will be a little bit more huge PR because of our list of backends
11:06 nrzkt i already have the design to do it in pg & sqlite, others are not ready
11:07 nrzkt and sorry est31 but the purpose of the PR is not to rewrite the player file, just add a feature to players
11:07 nrzkt i remove your label as it's not adapted to its case
11:07 est31 no, it should remove the player file completely
11:07 nrzkt then merge this and let me finish the second part
11:07 est31 and add a migration system
11:07 est31 no
11:08 nrzkt i don't want to do a huge PR and hmmm will cry if we remove files, as you know
11:08 est31 finish it, then we can merge the combined pr
11:08 est31 we should remove them, sorry
11:08 est31 they are horribly broken
11:09 nrzkt paramat it's not controversial PR, the feature is not controversial, it only miss a db backend xD
11:09 est31 you know this when server admins get SPEEDUPs when removing player files
11:09 nrzkt est31, i agree with you
11:09 gregorycu Does minetest need to be rewritten?
11:09 nrzkt but wait a minute, each change at a moment
11:10 nrzkt i already have a full postgresql + sqlite3 player backend on my server + the migration process written in python, but it's not the purpose of this PR
11:10 nrzkt this PR only adds extended attributes on palyer,s not rewrite the player backend
11:10 nrzkt it's another PR
11:10 est31 nrzkt: its okay if you only add extended attributes for now
11:10 nrzkt just let the two features separated
11:10 est31 but do not store them in json files
11:10 est31 not at all
11:10 nrzkt est31, it's the only thing this PR does
11:10 est31 no json files
11:10 nrzkt no jsonfiles here
11:11 est31 store them in databases
11:11 nrzkt it's an attribute in the player file
11:11 est31 ah
11:11 est31 no to that as well
11:11 est31 we should not extend the broken system
11:11 nrzkt but let the two things separated, i need time to finish to write the second part :)
11:11 est31 the bigger a player file, the more time it will take to load it
11:11 est31 and to write it
11:11 nrzkt see this as a ... feature preview :)
11:12 gregorycu Do you want my opinion on the matter?
11:12 gregorycu Does this change add a new datapoint, or does it make potential future migrations to db backend harder?
11:12 nrzkt no
11:13 nrzkt because it only add an attribute and i already have a working solution for this in relationnal DB which is wrote in another branch
11:13 gregorycu If it doesn't make it harder to refactor in the future...
11:13 nrzkt look at the result (i don't write the syntax of extended attribute, just the structure
11:13 nrzkt http://pastebin.com/3JALxy37
11:13 nrzkt oops i pasted two outputs, but the same :)
11:14 est31 okay, I've read the changes, still -1
11:14 nrzkt why ? except this is not stored in DB ?
11:14 est31 we can make that change once we store player data in the database
11:15 est31 but the bigger we make player files, the more likely they break
11:15 nrzkt you know, if i add the DB storage backend now i see at many kilometers that kwoelr will complain about the complexity of the PR and it's better to do kiss and working commits and he wants a file storage :p
11:15 est31 plus, we write player files each few seconds
11:15 nrzkt est31, yes, but we will write to DB each time too
11:15 est31 that doesnt have to be
11:16 est31 just write to the db each time you change something
11:16 nrzkt it's what we do with player files in fact
11:16 nrzkt the dirty flag on remote player is for that case
11:16 nrzkt we save if dirty flag is set or player disconnects
11:17 est31 yes
11:17 est31 but you store all attributes for a player
11:17 est31 not just those that changed
11:18 est31 if mods start using the system for all kinds of things this becomes a problem
11:18 est31 right now the system is only used for things that change often anyway
11:18 est31 like inventory or position
11:18 est31 but they are few things
11:18 est31 and very litle data for each
11:18 nrzkt yes, with relationnal DB we should be a little bit precise on what we save (inventory, player, extended attributes), but write 1kb data vs write to DB doesn't change the load on the server, but yes a DB is better for server owners to make a website interact with it
11:19 est31 but what happens when mods add many things and big data each
11:19 nrzkt in fact position changes everytime
11:19 est31 yes, you must assume that you rewrite the files for each emerged player
11:20 nrzkt est31, if mods store too many datas, storing in DB vs storing in files doesn't change anything, you write same or greater amount of data
11:20 est31 it depends on how you make your DB
11:21 est31 if you make your DB with one column for the player name (PRIMARY KEY) one for inventory one for position and one for extended attributes
11:21 nrzkt lol
11:21 nrzkt wrong design
11:21 est31 then you indeed have to rewrite everything each time
11:21 nrzkt inventory should have its own  table with (inventory_name, slot_id, item_id, item_count
11:21 est31 but if you make your DB with one column for the player name and one for the attribute name (they two combined make the PRIMARY KEY)
11:21 est31 then its okay
11:22 est31 nrzkt: that's overengineered IMO
11:22 nrzkt no est31 because this permits to dump easily the inventory o na website
11:22 nrzkt and it's very cheap to save
11:23 est31 you store more IDs and stuff than actual inventory data
11:23 nrzkt i have this on my server, let me a minute to give you stats
11:23 nrzkt not a problem for a relational db
11:23 est31 nrzkt: that should not be the measure we optimize against
11:23 est31 nrzkt: what do you propose next, storing each node separately in the db?
11:23 nrzkt if you use a relationnal DB to do NoSQL DB it's totally useless :)
11:24 nrzkt then use mongodb
11:24 est31 so your argument is "because we can use it we should"?
11:24 nrzkt i have 567849 stored in my db
11:25 est31 even though its worse?
11:25 nrzkt and 3077454
11:25 nrzkt items
11:25 est31 you just need to write some little deserialisation code on the webserver, that shouldn't be too hard should it?
11:25 nrzkt if you want a DB USE the DB, not only store data because it's a DB
11:26 nrzkt a website should read the properties directly, not doing deserialization, don't forget server owners are mainly using Lua, they don't want to deserialize but have a properly designed object
11:27 nrzkt if you want a db backend then i propose to do a DB PR first, merge it and then i will rebase and fix the current PR to use it
11:27 nrzkt not merge the two features in one PR
11:27 est31 well we can provide deserialisation libraries in python and in ruby
11:27 est31 then server owners can use those for their websites
11:28 est31 or if you want in php too
11:28 nrzkt est31, and who maintain those libraries ? us ?
11:28 est31 nrzkt: that proposal sounds good
11:28 nrzkt sorry but giving a proper relationnal DB because we are using it vs: adding overhead server & client side is not the good solution
11:29 est31 we can do it in two prs
11:29 est31 but this PR is blocked
11:30 nrzkt you blocked, if two others wants it this doesn't block anything. But okay for adding DB before merging that if you want
11:30 Calinou <gregorycu> Does minetest need to be rewritten?
11:30 est31 why should it
11:30 Calinou maybe, every time I talk about the world storage issue, people tell me "write your own binary format"
11:30 Calinou "and don't use relational SQL databases for this"
11:31 gregorycu I have a feeling that both non-immediate graphics, along with client side (predictive) lua, and maybe even SQL databases probably require a ground-zero rewrite of the architecture
11:32 est31 well maybe someone thinks this is an abuse of the relational db model
11:32 est31 but we use dbs not for their relational features
11:32 est31 we use them for their storage features
11:33 est31 that you can crash and your db still works fine
11:33 gregorycu For their ACID?
11:33 est31 gregorycu: yes
11:34 est31 also, if you use many little files there is this second issue that filesystems weren't really designed for that case
11:35 est31 like the fact that a fs always allocates at least one hdd block
11:35 est31 and that some filesystems are case insensitive
11:36 gregorycu And some filesystems have characters that are illegal
11:36 gregorycu (Which you may want to use for filenames)
11:36 Zeno` what's ground-zero mean?
11:37 est31 https://github.com/nerzhul/minetest/blob/master/src/player.cpp#L409
11:37 est31 just look at this
11:38 gregorycu est31: Yeah, it's retarded. I'm not sure nrzkt disagrees with us however.
11:38 est31 he agrees I think
11:38 gregorycu Zeno`: Bottom up rewrite
11:39 est31 I do disagree with him about making everything totally relational though
11:39 Zeno` hmm ok. I'd never heard the term ground zero used that way before
11:39 gregorycu I'm special
11:42 nrzkt yees est31 this code is an aberation, i didn't touch it because i didn't know the side effects :p
11:43 nrzkt est31, if i have a website and i want an inventory based on its name or find all items of a player based on item name, i really should load all the inventory data to do an update or a delete ? this is overkill
11:44 gregorycu nrzkt: I suppose this is a question of usecase
11:44 est31 yes
11:44 est31 and, for what usecase do you need to do this kind of thing
11:44 paramat website linkage is low priority
11:45 nrzkt gregorycu, yes, but having serialization doesn't permit this usecase, and to do stats or massive removal of an item into db (if node name changed or was removed by the admin)
11:45 est31 nrzkt: that's a valid point
11:46 gregorycu Feels like hyper-normalisation to me
11:46 est31 although removals can be handled by removing it when loading the inventory
11:46 est31 and stats by nightly run scripts
11:47 est31 but I admit both those approaches are worse than some sql statement .)
11:47 est31 still, the removal problem needs to be solved by the map as well
11:48 est31 and now excuse me, I've got to do some stuff for uni today
11:48 * est31 walks away
11:54 nrzkt yes we can fix it when loading inventory, but core didn't know better than admin what to change
11:54 nrzkt interesting discussion anyway
12:18 Wuzzy joined #minetest-dev
12:18 Wuzzy hi all
12:19 Wuzzy hi paramat
12:19 Wuzzy I want to talk about the workflow on GitHub
12:20 nrzkt yes ?
12:21 Wuzzy i dont mean to offend or insult me, but why is it you just dont commit the changes you request from a PR yourselves? in PR 1357, you seem to have already done the work, but you are still asking me to basically repeat what you have already did. Why? Personally, I find this practice pretty annoying
12:21 Wuzzy This goes mostly to paramat
12:22 Wuzzy oh, in case you missed it: GitHub has now a feature called “Allow edits from maintainers.”. you can directly edit on a branch of a PR when you're maintainer
12:22 Zeno` #1357
12:22 ShadowBot https://github.com/minetest/minetest/issues/1357 -- Allow custom liquids to have drops by sfan5
12:23 Wuzzy no, this is minetest_game.
12:23 Wuzzy sorry
12:23 Zeno` oh :)
12:23 Wuzzy oops I mean Minetest Game
12:26 Wuzzy hmmm paramat is not there, it seems. :(
12:27 Wuzzy i hate it when in IRC people seem to be always “there” 24/7 while, in fact, they're not actually there ... lol
12:27 paramat you wrote you could not do the texture yourself, sofar suggested the devs do it instead, so to simplify things and speed up the process i did the texture myself and said you can use it in the PR if you want
12:27 paramat it's an all-round good thing i did
12:28 paramat it's also normal practice for devs to ask for changes to a PR, to be done by the author, instead of doing the work themselves
12:29 paramat you have got annoyed by this before, it's weird
12:33 red-001 opinions on #4642 ?
12:33 ShadowBot https://github.com/minetest/minetest/issues/4642 -- Allow the join/leave message to be overridden by mods. by red-001
12:33 Wuzzy paramat, it is not about that you're asking for changes in PRs. i mean, if you found a serious bug you have good reason to expect from the submitter to fix it
12:34 Wuzzy its just that youre asking me to do something what you apparently already did. It's weird. Maybe I just misunderstood something. idk
12:38 paramat well, i just made a texture you could use, this way you save us having to alter your commit or add another commit, after all you are the contributor and it's your PR, you're expected to do the work
12:39 Wuzzy did you actually touch the code or how did you make the screenshot?
12:39 paramat thanks for using it
12:39 paramat and for updating the PR, appreciated
12:40 paramat i just made the texture and put it into my local copy of mtgame to test it, i didn't go near your PR
12:40 Wuzzy i will talk to you later.
12:40 Wuzzy OOOHhh...
12:40 Wuzzy i see
12:40 paramat ah :]
12:40 Wuzzy in that case, forget my complaint
12:40 paramat no problem then
12:41 Wuzzy I thought you have actually edited the code, and then asked me to do the same changes you did, and submit them ... XD
12:41 paramat i see
12:42 paramat whenever i ask for changes i rarely touch the commit itself
12:47 Wuzzy ok
12:48 Wuzzy paramat: Out of curiosity: What do you think of the “Allow edits by maintainers” feature in general? Is it good or will you more tend to avoid it? (btw: this question goes out to other Minetest and Minetest Game devs)
12:49 paramat that may have been the misunderstanding last time too :] anyway cool ..
12:50 paramat erm i haven't dared use that, i'm hesitant about these new github features
12:50 Wuzzy i can kinda understand why
12:50 Wuzzy but the PR submitter can always opt to disable it
12:50 paramat i tend to work in terminal with git as much as possible and avoid fancy github stuff
12:51 Wuzzy lol ok :D
12:51 Wuzzy hmmm idk maybe this feature also implies that it grants maintainers commit access to the PR branch. regardless of GitHub or commandline. i havent tested it yet
12:52 paramat i actually haven't thought about 'allow edits by' at all, don't even understand it
12:52 paramat i doubt there is access to your branch, just edits to the commit
12:52 Wuzzy as far I understand it, it means that maintainers of the repository to which the PR is posted to (minetest_game in this case) can directly commit to the branch of the PR
12:53 Wuzzy wait, what? *editing* commits? if this is true it this would be very wrong
12:53 paramat i hope not
12:53 est31 yes that is my understanding too
12:53 est31 (I'm semi present)
12:53 Wuzzy ah, i see. you want to avoid features you dont fully understand. makes sense :D
12:54 Wuzzy OK, as far I understand it, as long I have PRs for Minetest or Minetest Game, i should always expecting to see more change requests. And I guess this won't change in the near future.
12:55 paramat i mean i hope the feature doesn't allow devs to edit a contributor's branch
12:55 paramat maybe it does ..
12:55 nore Well, best way to test how it works is to enable it for some test PR and see what happens
12:56 Wuzzy hmm, i bet GitHub has some help page on this new “fancy” feature.. lemme see
12:56 paramat anyway, yeah, if you submit a PR unfortunately it can be a long frustrating series of requests from devs for changes
12:56 Wuzzy I don't want to pollute your repos with “Test PRs”. lol
12:57 paramat 1 test PR is ok
12:57 Wuzzy hmm, what about this: I post PR, wait for a week for requests to pour in, then start editing everything at once. xD
12:57 nore paramat: yeah, that's why allowing maintainers to edit the PR to fox a few style issues could be useful
12:58 paramat i see
12:58 nore since that kind of change is asked quite often, and it is a bit annoying to the one who opened the pr
13:01 Wuzzy Welcome in my world. XD
13:01 Wuzzy i try to abide by coding style guide from now on, however
13:01 Wuzzy "do it right the first time"
13:01 Wuzzy i just didnt know that even Minetest Game has coding style
13:02 nore I just read the help page, and how I understand it is that it allows push access by maintainers to the branch of the pull request
13:02 Wuzzy thx
13:02 nore So that can include editing commits
13:03 Wuzzy oh right. by git rebase, and then git push... O SHI-
13:04 Wuzzy but doing rebasing like this is evil anyway :)
13:08 paramat if i alter a commit i do it from git in terminal, that way your branch isn't accessed but you still remain the author
13:09 paramat so there's another way for devs to alter
13:11 Wuzzy I also use only Git in terminal for my commits
13:11 Wuzzy By the way: https://github.com/minetest/minetest_game/pull/1346
13:15 paramat +1
13:19 AcidNinjaFWHR joined #minetest-dev
13:20 Taoki joined #minetest-dev
13:23 paramat left #minetest-dev
13:32 Samson1 joined #minetest-dev
13:34 STHGOM joined #minetest-dev
13:53 Taoki joined #minetest-dev
14:57 garywhite joined #minetest-dev
14:59 fapper joined #minetest-dev
15:08 hmmmm joined #minetest-dev
15:17 Foghrye4_ joined #minetest-dev
15:28 garywhite_ joined #minetest-dev
15:35 AcidNinjaFWHR joined #minetest-dev
15:36 Taoki joined #minetest-dev
16:07 twoelk joined #minetest-dev
16:35 jin_xi joined #minetest-dev
16:57 Taoki joined #minetest-dev
16:58 Krock joined #minetest-dev
17:10 Taoki joined #minetest-dev
17:52 johnnyjoy joined #minetest-dev
17:55 johnnyjoy joined #minetest-dev
18:11 Foghrye4__ joined #minetest-dev
18:13 nrzkt joined #minetest-dev
18:40 * sofar sees fire, adds oil
18:40 Krock The school's burning. Get more wood.
18:45 jin_xi *weed
18:46 ssieb joined #minetest-dev
18:52 DI3HARD139 joined #minetest-dev
19:18 thePalindrome What are the plans for more secure password hashing?
19:18 thePalindrome I'd like to suggest bcrypt
19:18 thePalindrome Failing that, anything that's not base64
19:18 Calinou thePalindrome: uh, we use SRP with hashing and all
19:18 Calinou we're not using Base64 for a *long* time now!
19:18 Calinou with salting too
19:18 Calinou our password storage is quite secure, secure enough for a game
19:18 Calinou (this is not a banking application)
19:18 thePalindrome Ah, the wiki is out of date
19:18 Calinou I'd still like a "remember password" option though :p
19:19 Calinou (feel free to complain about its security, but it's up to the user to use it)
19:19 thePalindrome eh, I can at least get behind that
19:19 thePalindrome at least the insecurity actually is convienent
19:19 thePalindrome Could somebody update the wiki?
19:19 * thePalindrome doesn't have an account :um:
19:19 Calinou sure, but I need confirmation from ShadowNinja first
19:19 Calinou (who implemented SRP, IIRC)
19:20 thePalindrome Which hashing algo?
19:21 Calinou I don't remember
19:21 * thePalindrome dives into the code
19:22 thePalindrome sha-1 it appears
19:24 Calinou hmm, I wonder why that was used
19:24 Calinou no dependency on other libraries?
19:24 thePalindrome Dunno
19:24 thePalindrome bcrypt? :eager:
19:24 hmmmm est31 was the one who implemented srp
19:25 hmmmm pretty sure it's using SHA-256 in compliance with the SRP 6a standard
19:26 thePalindrome The code seems to be using sha-1 though
19:26 thePalindrome Unless it's in another branch
19:26 hmmmm no.... hold on
19:28 hmmmm https://github.com/minetest/minetest/blob/master/src/util/srp.cpp#L53
19:28 Tmanyo joined #minetest-dev
19:29 thePalindrome Huh
19:29 thePalindrome auth.cpp uses sha-1
19:29 thePalindrome https://github.com/minetest/minetest/blob/master/src/util/auth.cpp#L24
19:30 thePalindrome Note: I believe you, I'm just trying to figure the code out here :P
19:30 hmmmm auth.cpp implements the legacy password and provides an abstraction layer between the various authentication methods
19:30 hmmmm legacy password system
19:30 thePalindrome Oh duh :P
19:31 hmmmm because the way it was implemented, srp was made into its own library that est31 maintains on his own
19:31 thePalindrome I see
19:31 hmmmm and the legacy auth system remains in auth.cpp because it's so small it doesn't deserve its own source file
19:31 thePalindrome the wiki *really* needs updating then
19:31 hmmmm nobody uses the wiki
19:31 * thePalindrome does
19:39 turtleman joined #minetest-dev
19:40 Taoki joined #minetest-dev
19:40 twoelk obviosly those that know all the secrets of minetest have invested a lot of hard work and much time in understanding their little corner of the code. This has cost way too much sweat, blood and tears to be shared with any wanabee script kid that stumbles through some wiki. Documenting one's knowledge is one of the hardest things to do anyways besides being abysmally boring and draining and unrewar
19:41 thePalindrome lol
19:41 thePalindrome At least I have an accepted patch :P
20:52 Manzano_ joined #minetest-dev
21:03 Hunterz1 joined #minetest-dev
21:06 Taoki joined #minetest-dev
21:12 Foghrye4__ joined #minetest-dev
21:19 red-001 joined #minetest-dev
21:24 FirePowi joined #minetest-dev
21:32 thePalindrome joined #minetest-dev
21:35 troller joined #minetest-dev
21:36 sofar in my experience, wiki's function best if developers aren't asked to maintain them
21:37 sofar if developers have to maintain them, it's a sure sign that the community doesn't have enough support to let developers code and people tracking what is happening in the code and update the wiki for them
21:39 sofar fortunately we have #minetest-project :)
21:42 Player_2 joined #minetest-dev
21:49 hmmmmm joined #minetest-dev
22:30 twoelk joined #minetest-dev
22:39 Gael-de-Sailly joined #minetest-dev
23:24 twoelk|2 joined #minetest-dev
23:28 Samson1 joined #minetest-dev
23:33 AnotherBrick joined #minetest-dev
23:45 werwerwer joined #minetest-dev
23:48 DI3HARD139 joined #minetest-dev

| Channels | #minetest-dev index | Today | | Google Search | Plaintext