00:12:35 -!- iovoid has changed nick to iovoid_.
00:12:42 -!- iovoid_ has changed nick to iovoid.
00:19:40 -!- sawyerschumacher has joined.
00:29:57 -!- oerjan has joined.
00:31:09 -!- sawyerschumacher has quit (Quit: WeeChat 1.9.1).
00:34:22 <lambdabot> ENVA 212350Z 10011KT CAVOK M10/M14 Q1011 RMK WIND 670FT 11007KT
00:35:11 <Roger9> Hmm... well, I understand the 'WIND' bit. And maybe the '670FT' bit.
00:35:40 <oerjan> the 212350Z bit is date and time
00:36:32 <oerjan> M10/ is temperature, /M14 is dew point temperature (afaiu if they're close it's humid.)
00:36:45 <oerjan> the rest is pretty vague to me as well.
00:38:05 <boily> quintopia: I am worked. I hate bureaucracy :D
00:39:40 <Roger9> It's a political system in which squirrels rule the nation, and burrow (aka bureau) their nuts.
00:42:14 -!- fungot has quit (Ping timeout: 258 seconds).
00:42:29 -!- fungot has joined.
00:43:54 <oerjan> `# Stolen from Roger9//`learn Bureaucracy (from French bureau, "burrow") is a political system in which squirrels rule the nation, and burrow their nuts.
00:43:57 <HackEgo> Learned 'bureaucracy': Bureaucracy (from French bureau, "burrow") is a political system in which squirrels rule the nation, and burrow their nuts.
00:50:58 <lambdabot> CYUL 220000Z 15005KT 15SM SCT160 BKN240 05/M02 A2978 RMK AC4CI2 SLP087
00:51:03 <boily> strangely warm today.
00:52:27 <boily> Roger9: where is your nearest airport?
00:55:52 <Roger9> boily: I have no idea.
00:58:02 <boily> no airports at all where you live?
01:09:15 <Roger9> I'm... not that into airports to be honest.
01:11:38 -!- jaboja has quit (Ping timeout: 268 seconds).
01:15:17 <lambdabot> EGLL 220050Z AUTO 23015KT 9999 BKN019 14/10 Q1004
01:15:32 <fizzie> You said something about strangely warm?
01:15:48 <fizzie> They're saying it's going to be 16 tomorrow.
01:16:05 <fizzie> (Then it'll go back to normal.)
01:20:13 <oerjan> here it's about -10 now but will go up to nearly +10 in a couple days
01:23:23 <lambdabot> KOAK 220053Z 31005KT 10SM FEW100 BKN180 17/13 A3008 RMK AO2 SLP186 T01670133
01:32:48 <boily> helloochaf. you and your tropical weather :P
01:32:56 <boily> quintopia: what's a demoli?
01:33:12 <lambdabot> KSAN 220051Z 33004KT 10SM FEW200 FEW250 24/04 A3001 RMK AO2 SLP162 T02390044 $
01:33:21 <lambdabot> LLBG 220120Z 27013KT 9999 BKN045 17/07 Q1015 TEMPO -RA BKN030
01:34:07 <boily> fungot: do you do you saint tropez?
01:34:07 <fungot> boily: i guess, but you probably get banned for behaving like an asshole, and not for anything else
01:37:13 <fizzie> fungot: Stop being coherent, it's weird.
01:37:13 <fungot> fizzie: bindings from other programs? i have fnord) gold. :)
01:47:16 -!- xkapastel has joined.
02:13:12 <quintopia> demoli is the best psychonauts speedrunner
02:14:34 -!- jaboja has joined.
02:14:38 * ais523 wonders what sort of non-asshole behaviour would typically be considered banworthy
02:14:56 <shachaf> Are Psychonauts speed runs interesting?
02:15:09 <shachaf> Do they get all the figments?
02:15:23 <shachaf> I guess they do sometimes.
02:19:41 -!- jaboja has quit (Ping timeout: 240 seconds).
02:25:42 <oerjan> /ban doesthiswork1 What a stupid question
02:32:46 <ais523> doesthiswork1: places in general
02:33:08 <ais523> right, this is what 100% speedruns are normally for
02:33:13 <ais523> making sure you collect everything collectable
02:33:19 <ais523> any% lets you collect any proportion of it
02:33:21 <shachaf> Apparently for Psychonauts it's 101% speed runs.
02:33:38 <ais523> (in some games that's 0% but in many games, collecting some proportion allows you to go faster)
02:34:04 <shachaf> "This trick requires levitation. While in midair, press and hold 'Float' (default: Shift) to begin floating. Now, while still holding Float, rapidly alternate pressing 'Jump' (default: Spacebar) and the psi power key that you assigned levitate to (typically this is the 3rd slot). If you alternate fast enough, Raz will begin to rise slowly. This can be repeated indefinetly to reach any height, and skip ma
02:34:04 <ais523> there are even a few games where any% = 100%, typically because the reward for collecting everything is so good it makes the rest of the game considerably faster
02:34:10 <shachaf> ny parts of the game. If you alter your cadence, you can also use it to cover horizontal distance quickly."
02:34:37 <shachaf> When is Psychonauts 2 coming out?
02:34:41 <quintopia> shachaf: it doesnt work on xbox or with vsync on
02:34:57 <shachaf> Do you really believe that?
02:38:18 <quintopia> they have been meeting all their milestones so far. i think wed know in advance if it werent going to happen
02:47:02 -!- boily has quit (Quit: SONG CHICKEN).
02:48:50 -!- ais523 has quit (Quit: sorry for my connection).
02:49:46 <shachaf> How much would you pay for a contract that paid $100 if they release it on or before Dec 2019?
03:10:00 -!- watered has joined.
04:19:10 <doesthiswork1> how much are you offering for the option to purchase that contract for $50 on Jan 2020?
04:20:42 <shachaf> That sounds needlessly complicated.
04:44:36 -!- heroux has quit (Ping timeout: 246 seconds).
04:50:20 <oerjan> shachaf: sounds a bit derivative
04:59:52 -!- heroux has joined.
05:44:00 -!- doesthiswork1 has quit (Quit: Leaving.).
06:14:34 -!- friendlyGoat has joined.
06:32:43 -!- doesthiswork has joined.
06:45:09 -!- friendlyGoat has left.
06:45:36 -!- friendlyGoat has joined.
06:45:41 -!- friendlyGoat has left.
06:47:41 -!- oerjan has quit (Quit: Nite).
06:59:05 -!- doesthiswork has quit (Read error: Connection reset by peer).
07:11:54 -!- FreeFull has quit.
07:15:11 -!- sleffy has quit (Ping timeout: 248 seconds).
07:38:41 -!- watered has quit (Ping timeout: 255 seconds).
08:15:17 -!- brandonson has quit (Ping timeout: 268 seconds).
08:15:48 -!- brandonson has joined.
08:46:19 <Vorpal> what the hell is going on with all the /me?
08:47:22 <Vorpal> also why did this show up in here * Sigyn sets modes [#freenode +qz $~a]
08:47:31 <Vorpal> maybe my IRC client is bugging out and mixing up channels???
08:48:13 <Vorpal> well znc or hexchat might be messing up
08:49:02 <Vorpal> hm something is very broken. Nothing happens on /join or /part of various channels
08:50:44 <Vorpal> restarted the IRC client, lets see if it helps
09:34:26 -!- zseri has joined.
09:35:21 -!- AnotherTest has joined.
09:36:27 -!- LKoen has joined.
09:36:42 -!- xkapastel has quit (Quit: Connection closed for inactivity).
10:00:59 -!- sftp has joined.
10:02:17 <HackEgo> [wiki] [[User:Zzo38/Untitled 1]] https://esolangs.org/w/index.php?diff=53395&oldid=53388 * Zzo38 * (-16) Now the name is given. Will move
10:02:44 <HackEgo> [wiki] [[Special:Log/move]] move * Zzo38 * moved [[User:Zzo38/Untitled 1]] to [[Crement]]: Named by [[User:Ais523]]
10:02:45 <HackEgo> [wiki] [[Special:Log/move]] move * Zzo38 * moved [[User talk:Zzo38/Untitled 1]] to [[Talk:Crement]]: Named by [[User:Ais523]]
10:21:11 -!- watered has joined.
10:23:24 -!- LKoen has quit (Remote host closed the connection).
10:48:10 -!- LKoen has joined.
11:27:24 -!- LKoen has quit (Remote host closed the connection).
12:40:51 -!- watered has quit (Ping timeout: 250 seconds).
13:01:24 -!- watered has joined.
13:13:08 <b_jonas> In M:tG, is the name of the "By Gnome Means" card https://media.wizards.com/2017/ust/en_xqPx77FNyw.png supposed to be something like Gnome Ann https://www.xkcd.com/1704/ ?
14:00:32 -!- doesthiswork has joined.
14:35:43 -!- sftp has quit (Remote host closed the connection).
14:36:04 -!- sftp has joined.
15:15:59 -!- watered has quit (Ping timeout: 250 seconds).
15:30:24 <HackEgo> [wiki] [[XTW]] https://esolangs.org/w/index.php?diff=53400&oldid=53391 * Zseri * (+251) +local variables
15:39:20 -!- LKoen has joined.
15:39:21 -!- LKoen has quit (Remote host closed the connection).
15:39:27 -!- LKoen has joined.
15:50:31 -!- LKoen has quit (Remote host closed the connection).
16:08:10 <zseri> I added support for local variables to XTW, because it was too complex to simulate them using nested objects.
16:10:20 <zseri> (which are simulated using nested objects and local vars are then simulated using objects which save the state of every local var, which is complex, error-prone and confusing)
16:27:06 <Roger9> zseri: A concept I had in mind was 'first-class scopes': a scope would have a parent, map variable names to variable values, and the code would currently be running in the 'current scope'.
16:32:09 <Roger9> This first-class implementation would make supporting closures much easier. A function would have a parent scope (that refers to the current scope when the function was created), a list of parameters, and a block of code. The code would be run in a new scope, whose parent would be the function's creator scope, and would match the parameters to their arguments (along with maybe a return keyword to match to the caller's continuation).
16:32:34 <Roger9> Closures would simply be functions which access variables in their creator scope.
16:35:46 -!- LKoen has joined.
16:48:43 <zseri> I think I could integrate this concept in XTW by mapping ::l::__ to the parent scope and so on like ::l::__::..
16:50:10 <zseri> I need to keep track of special variables in the interpreter (in the symbol table, which maps variable name to numbers and uses them at runtime)
16:57:39 -!- sftp has quit (Max SendQ exceeded).
16:58:04 -!- sleffy has joined.
16:58:08 -!- sftp has joined.
17:34:20 -!- xkapastel has joined.
17:38:58 <zseri> but I think that one level (::l::__::var) is enough
17:55:58 -!- Phantom_Hoover has joined.
17:55:58 -!- Phantom_Hoover has quit (Changing host).
17:55:58 -!- Phantom_Hoover has joined.
17:57:45 -!- sebbu2 has joined.
17:58:31 -!- sebbu has quit (Ping timeout: 240 seconds).
17:59:04 -!- FreeFull has joined.
18:01:17 -!- sebbu2 has changed nick to sebbu.
18:08:57 -!- LKoen has quit (Remote host closed the connection).
18:13:40 -!- LKoen has joined.
18:17:24 <zseri> ok, I resolve parent namespaces recursively
18:18:35 <zseri> NOTE : I don't implement first-class scopes, but emulate that with dynamic resolved namespaces.
18:24:02 -!- LKoen has quit (Remote host closed the connection).
18:25:46 -!- watered has joined.
18:31:46 <zseri> done. https://github.com/zserik/zxtw/blob/master/examples/lclparent.xtw (output: before: 1\nafter: 2\n)
18:36:35 <zseri> it's easier to emulate first-class scopes with dynamic namespace with resolving them partly at startup-time (before the interpreter builds the bytecode) and the rest done at runtime (could be done at startup-time too, I would just need to cache how many scopes to skip)
18:38:25 <zseri> instead of emulating first-class scopes with compose at scope-push and decompose at scope-pop (with dereferencing)
18:39:30 <quintopia> namespaces, as the man said, are one honking great idea
18:52:47 <zseri> One disadvantage of the interpreter is that it suppports functions only at global scope (not inside a namespace) but a function name prefixed with a namespace. It supports namespace at global scope and inside functions
19:13:32 -!- Cale_ has joined.
19:14:21 -!- MrBusiness has quit (Quit: https://www.youtube.com/watch?v=xIIqYqtR1lY -- Suicide is Painless - Johnny Mandel).
19:18:20 -!- LKoen has joined.
19:18:23 -!- LKoen has quit (Remote host closed the connection).
20:19:51 -!- watered has quit (Ping timeout: 240 seconds).
20:20:19 -!- watered has joined.
20:42:55 -!- zseri has quit (Quit: Leaving).
20:43:22 -!- ais523 has joined.
20:49:24 -!- ais523 has quit (Remote host closed the connection).
20:50:37 -!- ais523 has joined.
20:51:48 <Roger9> 'Ello, ais523. Having trouble with your connection or somethin'?
20:52:13 <ais523> I basically always have trouble with my connection
20:52:22 <ais523> that's why my default quit message apologises for it
20:52:29 <ais523> (in the rare cases where it's actually working, I override it manually)
20:52:51 <Roger9> * ais523 has quit (Remote host closed the connection)
20:56:38 -!- ais523 has quit (Remote host closed the connection).
20:57:51 -!- ais523 has joined.
21:04:27 <ais523> Roger9: the weather here tonight is terrible (some amount of rain, and a huge amount of wind); that might have something to do with the connection's unreliability
21:07:07 <ais523> it's pretty bad even normally though
21:14:01 -!- sftp has quit (Remote host closed the connection).
21:14:23 -!- sftp has joined.
21:21:24 -!- ais523 has quit (Quit: quit).
21:55:41 -!- Cale_ has quit (Remote host closed the connection).
22:03:19 -!- ais523 has joined.
22:03:37 -!- wob_jonas has joined.
22:05:30 <wob_jonas> so it looks like neither git nor mercurial currently supports partial checkouts, as in cloning parts of a large repository selected by pathnames without having to store the rest of the data on my disk or transfering it through hard disk. Svn supports this, although there are some limitations.
22:05:38 <wob_jonas> This is a very important feature to me.
22:05:43 <ais523> so it turns out Half-Squirrel Half-Pony actually got errata
22:06:03 <wob_jonas> If you happen to know that git or mercurial has gained ability to do this since I heard these info, then PLEASE tell me.
22:06:08 <ais523> wob_jonas: I don't think the git/mercurial storage model is intended for repositories where that feature would be useful
22:06:31 <wob_jonas> I know a little about how a git repo works, and I'm quite sure that it should be possible to support this, it's just that the commands don't currently implement it.
22:06:33 <ais523> as it happens I'm currently working on a new VCS (scapegoat-inspired but less innovative and thus more likely to be useful/used in practice)
22:06:46 <ais523> wob_jonas: no, there'd be no way to create a new commit
22:06:53 <ais523> as the hash of that commit would involve the hash of all the files in other directories
22:07:03 <ais523> which you don't have because you didn't check it out
22:07:24 <wob_jonas> ais523: but the hash of a tree is computed from just the hash of tree and file objects it contains
22:08:07 <wob_jonas> ais523: if I have a tree where some children aren't stored locally, their *hashes* would still be stored locally, so I can create a new version of that tree by copying those hashes (and names), and recompute the new hash for it
22:08:15 <wob_jonas> ais523: this is for git, I don't know about mercurial
22:08:32 -!- watered has quit (Ping timeout: 255 seconds).
22:08:48 <Phantom_Hoover> i didn't care about vcs at all back when it was a thing but i've learned git since
22:09:14 <ais523> Phantom_Hoover: basically, the /original/ big idea was that each line would have a record of where it came from, meaning that merges could be implemented safely with less context in the files themselves
22:09:16 <wob_jonas> ais523: I still think it is likely possible to extend svn to a distributed vcs, and this would have some advantages: (1) it could be compatible with existing svn servers and repositories, that is, I could pull from a remote repo without that repo knowing anything about the distributed extension,
22:09:18 <ais523> (and likewise, unsafe merges could easily be detected)
22:09:35 <ais523> wob_jonas: how would you handle commit numbers?
22:09:43 <wob_jonas> (2) svn already handles the storage of objects and delta compression for you, which you can definitely reuse, and you can probably resuse its diff and merge tools as well
22:10:43 <ais523> I thought svn was famously bad at merging? or was that a different VCS?
22:11:13 <wob_jonas> ais523: that was older versions of svn
22:11:13 <wob_jonas> ais523: versions before 1.5 or 1.6 or something
22:12:16 <ais523> anyway, the problem I've come up across, which is leading me to think a new VCS is really needed
22:12:28 <ais523> is a scenario where two related forks of a project exist that want to share some code, but not all of it
22:12:40 <wob_jonas> ais523: basically I'd have branches in the local repo that track the remote repo, and you can't commit into them directly, just like in git. the extension would pull into them, and store the correspondence with remote objects (with path and version number in remote repo), technically stored in svn properties
22:12:44 <ais523> (this is obviously the case in the whole NetHack variants situation, but I can't imagine that's the only place where it happens)
22:13:04 <ais523> wob_jonas: so you're basically trying to use svn branches like git branches, which are like git remotes?
22:13:13 <ais523> that doesn't seem completely nonviable
22:14:02 <wob_jonas> ais523: svn already lets you have a read-only mirror or a dump-restore of a repository with only parts of the repo included, and can omit commits that don't modify the parts kept, thus skewing commit numbers. it has tools for this. what it can't do is pushing back, which I admit is a MUCH harder problem.
22:14:08 -!- jaboja has joined.
22:15:10 <ais523> hmm… in the VCS I'm working on, you could check out a single directory but only if no commits had been made to that directory and to a file outside that directory at the same time
22:15:23 <wob_jonas> ais523: yes. the idea behind that is that this way you can use such a branch for any read-only operation that svn already supports, such as childing or merging from it. the extension would have to map parents and mergeinfo in a tricky way.
22:15:26 <ais523> how would that interact with your usecase for checking out a single directory?
22:15:31 <ais523> (what is the reason, incidentally?)
22:15:42 <lambdabot> EGLL 222150Z AUTO 20020KT 9999 SCT026 BKN031 15/10 Q0990 TEMPO RA
22:15:56 <lambdabot> EGBB 222150Z 20020G34KT 160V220 9999 SCT023 SCT043 14/10 Q0983
22:16:14 <wob_jonas> ais523: I don't understand your question
22:16:27 <fizzie> I think we're having some "G"s as well, even if it's not listed.
22:16:29 <ais523> wob_jonas: well, you want to check out a single directory from a repo
22:16:46 <ais523> I'm asking why, because that influences what implementation techniques a VCS could use for that
22:18:53 <ais523> like, if a VCS supported checking out a single directory but had limitations, would the limitations affect you?
22:20:01 <wob_jonas> ais523: yes, they'd affect me. more like I want to clone and fetch a single directory, or more generally any set of objects, from a remote repo, and I want to be able to unfetch (delete local copy) any object, fetch new objects, and modify what will be fetched later (set paths to automatically fetch new children or not).
22:20:09 <wob_jonas> But I know that supporting this in full generality is difficult:
22:20:45 <ais523> wob_jonas: I'm still interested in what's in this repository?
22:20:58 <wob_jonas> it's especially difficult if I want to fetch an object that is the parent of an object that was already fetched, because to do that with the svn representation, I have to recreate the local version of all the descendant objects as new vanilla svn objects.
22:21:03 <wob_jonas> that's still possible, but difficult.
22:21:41 <ais523> what do you mean by parent/child here?
22:22:56 <wob_jonas> ais523: in git, every commit has usually one but generally any number of parents, which are earlier versions of the commit. in svn, commits don't have that, but each directory/file version has one (or sometimes zero) parent which is the immediate previous version of that object, and can be at any path as long as it's in an earlier commit,
22:23:31 <wob_jonas> and zero or more additional objects marked as extra ancestors merged into it. (there's also the restriction that directory and non-directory objects can't be mixed in parental relationship)
22:23:58 <wob_jonas> so svn tracks moves or copies of individual files or directories.
22:24:03 <ais523> oh, "earlier versions of the commit" isn't a way to describe it, really
22:24:06 <ais523> but I know what parent are in git
22:24:22 <ais523> just wasn't sure that was the terminology you were using
22:25:16 <wob_jonas> an svn working tree (checkout) can already be partial, it has local copy of usually just the latest versions of objects, but it can mix-match any local path in the checkout with any path-version in the remote repo, and stores pristine copies of each of those files
22:26:12 <wob_jonas> you can switch any local path to point to any path-version of the remote repo, set them to automatically check out new files in a checked out directory or not, and update paths to later versions of the same path easily
22:26:56 <wob_jonas> (there's some limitations, some obvious because only directories can have files or directories in them, some apparently just limitation of the programs about what you can modify, and some limitation about subrepositories which are like git modules)
22:27:53 <wob_jonas> ais523: so anyway, about your main question about why partial checkouts are useful for me
22:28:49 <wob_jonas> ais523: as long as you're always connected to a single writable central svn server (there can be any number of read-only mirrors, but no read-write ones), you can conveniently use svn to store big files or big directory trees.
22:29:26 <wob_jonas> these aren't source code, but dependencies (possibly binary) that you often want to check out together with the rest of the project to build the project, or data, or whatever
22:29:32 <wob_jonas> you can also store multiple projects together
22:29:43 <wob_jonas> svn partial checkouts make this easy
22:29:53 -!- AnotherTest has quit (Ping timeout: 248 seconds).
22:29:57 <ais523> storing multiple projects together is in general a bad idea because there's no obvious advantage over doing it separately
22:30:16 <ais523> although, what I'm working on /would/ be able to handle that if nobody ever committed to both projects in the same commit
22:30:21 <wob_jonas> you can replace an entire binary dependency with a new version, such that their difference doesn't compress well, and only the server will have to store older versions
22:30:38 <ais523> for binary dependencies, would it be OK to store the hashes of all those dependencies? just not download the dependencies themselves?
22:30:57 <wob_jonas> we want to store multiple projects together because one project can want to pull part of the code of another project, without making full dependency, and then want to merge bugfixes from them
22:31:03 <ais523> I assume it would be, because the number of files involved is still manageable, it's just the files themselves that are too large
22:31:32 <wob_jonas> ais523: often you want to download the big binary dependency, but not all previous versions of that dependency.
22:31:41 <wob_jonas> but sometimes you only want to download some of the dependencies, because they're big.
22:31:44 <ais523> most VCSes can do that already, including git
22:31:52 <ais523> that is, download only the most recent version
22:32:10 <wob_jonas> ais523: I think git can do that only on a full repo. you could try to store deps in git submodules
22:32:10 <fizzie> wob_jonas: I've not read the context here, but we have this thing called https://bitbucket.org/Google/narrowhg that sounded like plausibly relevant.
22:32:39 <ais523> wob_jonas: yes, git needs the most recent version of each file
22:33:11 <wob_jonas> ais523: but also you can't get full history of some files, but only latest version of other files, within the same git repo
22:33:17 <wob_jonas> it should be possible, but isn't with current tools
22:33:30 <ais523> right, that's just a UI problem I think?
22:33:40 <ais523> in that the UI doesn't give any easy way to do that
22:33:49 <wob_jonas> ais523: yes, which is why I'm asking if git has solved this since
22:33:58 <wob_jonas> ais523: they've already solved multiple worktrees (checkouts) with only one object storage
22:34:13 <wob_jonas> ais523: another thing though is that even without partial checkouts, I would want a distributed version of svn
22:34:29 <ais523> I think the way to go in terms of implementing this would be a lazy repository in which objects are retrieved only when their content is requested, and only directly to the filesystem
22:34:31 <ais523> you'd need to work online
22:34:39 <ais523> but it means that no more would be downloaded than actually necessary
22:34:57 <wob_jonas> ais523: two things you could do with a distributed version are (a) maintaining a forked project of a project where I never push back because they don't like me, but I can pull newer versions of that original project and merge into mine
22:35:27 <wob_jonas> and (b) committing revisions locally while I don't have access to the central writable server because no network connection or something
22:35:42 <wob_jonas> or even committing local branches that I don't want to share until I clean them up
22:35:57 <wob_jonas> people do all these things in git,
22:36:47 <ais523> why would you want svn in particular for that? because the people-who-don't-like-you happen to be using svn and you can't persuade them to change?
22:36:49 <fizzie> wob_jonas: For your case (a), I was using a thing where I had git and SVN share a checkout.
22:36:49 <wob_jonas> and they aren't incompatible with pulling from a central server, and in (b) you can push some or all of your commits to the central server later, while each repo has an immutable history
22:37:11 <ais523> but git-svn works very well for both those cases
22:38:46 <wob_jonas> ais523: if I understand correctly, in git-svn the local branches and commits are git branches and commits. I like svn and the way it handles files and renames and stuff, and don't want to lose that. but yes, that's a good point, git-svn may help with some of the model, especially if you already like git
22:39:37 <ais523> what git-svn basically does is to create a git remote that mirrors a given branch of an svn repository, I think
22:39:47 <ais523> although it can also commit to the branch
22:39:53 <ais523> as long as you have a linear history
22:40:18 <ais523> (the git-svn equivalent of git pull is git svn rebase, in order to ensure that the history stays linear; it pulls the remote changes then rebases your local changes on top)
22:40:45 <wob_jonas> ais523: yes, and it's the commit back part with proper tracking of ancestry and merges that's hard. we already have tools to read-only mirror part of a remote svn repository to a local svn repo.
22:42:57 <wob_jonas> because the moment you want to push back, you have to figure out how to set all the parent and mergeinfo for the new objects in the remote repo.
22:42:58 <fizzie> Regarding the earlier discussion fork, narrowhg lets you have a shallow copy of all files and a deep history (optionally up to N revisions) of some specific subset of files, though (IIUC) you can't have different depths for the not-entirely-shallow files.
22:43:34 <wob_jonas> fizzie: let me look at that. can you also have no versions at all of some paths?
22:44:54 <wob_jonas> Just to briefly mention other version control systems. (1) cvs can do partial checkouts, but I think svn can already do anything that cvs can do, and cvs hasn't gained new features for like a decade, so it's safe to discard it except for compatibility,
22:45:02 <fizzie> wob_jonas: I forgot. Maybe it was actually more limited than that, and you actually could only get either files (up to depth N) or no files at all, but not shallow copies of everything.
22:45:48 <ais523> wob_jonas: what about rcs? :-P
22:45:57 <wob_jonas> (2) darcs is like developped by some alien race and I didn't understand anything they say about how their vcs works, and (3) fossil now lets you create forks of a read-only repository and merge to it from the remote, sort of like private branches, but I believe it doesn't do partial checkouts,
22:46:03 <Roger9> This has probably been done before, but I'm messing around with a hypothetical language that compiles into BF.
22:46:05 <wob_jonas> (4) I didn't look at bazaar at all.
22:46:51 <ais523> (in RCS you're supposed to work directly on the machine holding the repository, the "check out" operation is more like a locking operation)
22:46:57 <wob_jonas> ais523: I think rcs is something that used to count as a "version control system" back when it was created, but doesn't count as one in the modern sense. or something.
22:47:08 <ais523> wob_jonas: it's a non-concurrent VCS
22:47:31 <ais523> it does control versions, it just doesn't control access in any but the crudest way
22:47:49 <ais523> (you have one repo per file, checking out the repo locks the file, then checking the repo back in will unlock the file and record a commit)
22:47:59 <ais523> come to think of it, this is probably where the terms checkout/checkin come from
22:48:09 <ais523> even though they make no sense in the context of something like cvs or git
22:49:36 <wob_jonas> ais523: yes, and "commit" is cleverly chosen so that the "cvs ci" abbreviated command would sound like either "commit" or "checkin", depending on whether you're a new cvs user or someone used to rcs
22:49:56 <wob_jonas> because cvs was originally designed as a better rcs, but later got important in itself,
22:50:25 <wob_jonas> just like how svn was originally a better cvs, but these days it develops on its own without regard to cvs,
22:50:46 <wob_jonas> and both rcs and cvs are probably not developed further with new features, and cvs hasn't even had a new release for several years
22:51:36 <wob_jonas> mind you, no new release for several years is fine for a mature tool if there aren't important bugs, I had no problem with no how the 7z compressor utility with windows GUI hasn't had a new release for like six years
22:52:45 <wob_jonas> ais523: so wait, you said you're developing a new version control system that isn't scapegoat. do you wish to tell a bit more about this?
22:53:15 <wob_jonas> or did you only say you're planning this new system?
22:53:38 <ais523> currently at the stage of a) planning, b) learning enough git internals to be able to write an importer
22:53:59 <ais523> I'm planning to use git as the backend in order to get a lot of functionality for free and take advantage of things like public git hosting sitse
22:54:12 <ais523> also because git was intended to be a VCS backend rather than a VCS itself
22:55:01 <wob_jonas> "Half-Squirrel Half-Pony actually got errata" wait what? I haven't read the setFAQ yet
22:55:30 <ais523> I'm really disappointed in it
22:55:52 <ais523> mostly because with many cards, how they're meant to work, and how the FAQ says they work, is different from what you'd assume based on the wording of the card
22:56:08 <ais523> there's a big departure from the "cards mean what they say" that you normally get on black-bordered cards
22:56:29 <ais523> presumably the templating team ran away from the set
22:57:16 <wob_jonas> ais523: my take on that is that MaRo is really bad at rules, he should never have been allowed to be even the un-rules manager, and we should just ignore everything he says about un-rulings and make up our own rules that make sense
22:57:53 <wob_jonas> ais523: even for unhinged, many of his rulings make no sense. for unstable, there's much more cards that would work in black-bordered, and he's publishing a lot of rulings about them.
22:57:57 <ais523> it says something that the half-squirrel half-pony errata was done by eli shiffrin (i.e. the /real/ rules manager)
22:58:06 <wob_jonas> I'll read them for flavor, but I'll not try to take them as meaningful rulings.
22:58:24 <ais523> he says he was really surprised that the errata actually works in black border, because it inspects a property of the card that no current card inspects
22:59:00 <wob_jonas> ais523: where's this errata? is it in MaRo's FAQ?
22:59:13 <ais523> wob_jonas: yes, but the metadiscussion about it happened on the Reddit thread about the FAQ
22:59:25 <ais523> or possibly about the erratum itself
22:59:48 <wob_jonas> oh yeah, the discussion for future errata used to happen in the old-style web forum that wizards maintained, until they closed it
22:59:53 <wob_jonas> there's even one comp rules modification that I promoted
23:01:38 -!- boily has joined.
23:03:14 <wob_jonas> (reading article) WHAT? that's not even what the card says.
23:05:16 <ais523> wob_jonas: which card in particular?
23:05:31 <wob_jonas> Ah! So those red creatures on the packaging are beebles!
23:05:56 <wob_jonas> I didn't recognize them. Beebles used to be of a more rounded shape and less red.
23:06:40 <\oren\> I ordered a supervillian laptop
23:07:32 <wob_jonas> what does a supervillian laptop mean?
23:07:53 <\oren\> https://www.sagernotebook.com/Notebook-NP9175.html
23:08:32 <\oren\> means it will probably double as a haridryer
23:08:35 <ais523> wob_jonas: hmm, which part of Better Than One surprised you? the fact that you have a shared life total?
23:09:01 <ais523> oh, the fact that you get attacked as a team, probably
23:09:04 <wob_jonas> ais523: that, and that it uses the rules for two-headed giant in general, not the default rules
23:09:10 <ais523> but yes, the card makes it look more like traditional multiplayer than 2HG
23:09:58 <wob_jonas> \oren\: mind you, I ordered a new copy of my heavy clicky keyboard (after ruining mine by spilling tea on it), and it's branded "Demon King". not quite a "supervillian", but still an antagonist.
23:11:14 <wob_jonas> ais523: the augment and especially the contraption rules are also a bit strange, but at least there there's clearly a new keyword or supertype where the rules have to be decided, not just a plain card where you don't expect strange rules
23:11:42 <ais523> the functional reprints of black-bordered cards are also surprising
23:11:51 <ais523> like the card costing 4GG with trample and no other abilities
23:12:13 <ais523> why would a card like that need a silver border?
23:12:33 <ais523> (IIRC, it doesn't have any silly stuff on its typeline or the like, either)
23:16:16 <\oren\> but wob_jonas that keyboard doesn't glow
23:16:29 <wob_jonas> \oren\: I don't want a glowing one
23:16:39 <wob_jonas> I don't understand the point of a glowing keyboard
23:17:15 <fizzie> wob_jonas: You can indicate the status of different things (like passing builds) by differently glowing keys.
23:17:34 <fizzie> \oren\: Sager == Clevo, right?
23:17:50 <wob_jonas> fizzie: I have a monitor for indicating things
23:18:13 <ais523> oh, the most obvious reason for glowing keys for me would be to easily place my hands on the keyboard when I'm working in near-total darkness
23:18:16 <wob_jonas> and I'm looking at the monitor a lot, but not much at the keyboard
23:18:18 <wob_jonas> it seems the whole set is revealed now
23:18:35 <ais523> there are indents on the F and J (and numpad 5) on most keyboards but they're often very subtle and hard to find
23:19:13 <ais523> so normally I either switch to a window with a white background, getting more illumination from the screen, or else move to a text boxt that accepts input and experimentally type a few letters to determine where my hands are
23:19:24 <wob_jonas> ais523: I don't want to work in near-total darkness, because it hurts my eyes to look at a screen when the environment is too dark. I know there are people who this doesn't bother.
23:19:29 <ais523> an illuminated keyboard would be a more direct fix to the problem
23:19:45 <ais523> wob_jonas: eyestrains caused by looking at something that's dimmer / harder to make out than the surrounding environment
23:20:01 <ais523> so darkness around the computer should in theory be the optimum for avoiding eyestrain
23:20:16 <ais523> but eyes are also hurt by focusing continuously at the same range for a long period of time
23:20:21 <wob_jonas> ais523: no, that's not true. the problem is that if the environment is dark, your pupils extend too much
23:20:33 <ais523> if the only thing you can see is your nearby screen, you have problems with that
23:20:38 <ais523> wob_jonas: well, I try to make the screen very dark too
23:20:45 <ais523> minimum brightness and black backgrounds almost everywhere
23:23:18 <wob_jonas> wait. Eager Beaver? how does green get an untap ability that works on any type of permanent? that doesn't make sense color-wise.
23:23:39 <wob_jonas> blue or rarely white can get that ability, green shouldn't
23:23:57 <ais523> green always used to be able to untap, I think?
23:24:15 <ais523> but then they moved away from that into just untapping lands
23:24:15 <wob_jonas> (which is already hard to balance, mind you)
23:24:21 <ais523> white can untap creatures, can it untap anything else?
23:24:38 <ais523> wob_jonas: can't remember any offhand
23:24:50 <ais523> cards which mess around with mass untapping tend to be UG nowadays
23:25:20 <ais523> not sure, I don't remember seeing it in the list of functional reprints but I didn't memorise the list in detail
23:25:24 <wob_jonas> UG would be fine. but beaver is pure green.
23:26:27 <\oren\> dimensions: 16.46" (w) x 11.63" (d) x 1.61" (h)
23:26:45 <ais523> does it even make sense to say which dimension is which?
23:26:58 <ais523> as in, isn't which dimension is which determined by their relative sizes rather than vice versa?
23:27:14 <\oren\> yeah well it's certainly inferrable
23:27:14 <wob_jonas> ais523: it can make sense for some types of objects
23:27:34 <wob_jonas> ones that have fixed elements on a certain side
23:27:49 <fizzie> I think even a hypothetical laptop could well have d > w.
23:28:07 <wob_jonas> ais523: like a washing machine would have its door on either the front, or the top and open backwards, so it helps to know which is the depth and which is the width
23:28:31 <ais523> fizzie: you're assuming that the hinge is on the boundary between the wide and high sides
23:28:45 <ais523> I'm not sure that's any more valid an assumption than that the wide side is longer than the deep side
23:29:01 <fizzie> ais523: I don't think I'm assuming that.
23:29:03 <ais523> I guess the distinction between wide and deep here is the orientation of the keyboard, though?
23:29:17 <fizzie> I'm assuming width is the axis the spacebar is congruent to.
23:29:40 <\oren\> like, a laptop could have a portrait display
23:30:00 <\oren\> i've never heard of such thing, but it could...
23:30:00 <wob_jonas> ais523: the hinge doesn't have to be close to the edge. older laptops have a screen narrower than the base part, and a large battery part behind the hinge in the base part.
23:30:14 <ais523> and at least one much-advertised laptop has no hinge at all
23:30:27 <ais523> the screen component is entirely separate from the keyboard component
23:30:48 <ais523> maybe that helps prevent it breaking, the hinge tends to be the most fragile part of a laptop…
23:31:36 <ais523> anyway, I'm going to see if anywhere round here sells water at 11:30pm
23:31:46 <ais523> normally I'd be home this late, but the water supply is under maintenance
23:31:47 <wob_jonas> so now we know what the mythic vanilla is. tricky.
23:32:05 <\oren\> like, what if you had a cube-shaped computer where the top of the cube hinges up to become a square screen
23:32:07 <ais523> so I'm at work, and am not sure how easy it'll be to get something to drink
23:32:15 -!- ais523 has quit (Quit: quit).
23:32:28 <\oren\> would that even count as a 'laptop'
23:32:48 -!- Phantom_Hoover has quit (Ping timeout: 240 seconds).
23:34:18 <\oren\> hmm, or wait. suppose instead you just have a cube with a display on one side and a keyboard that slides out of the space under the screen
23:34:35 <\oren\> I guess that isn't a laptop
23:35:12 <wob_jonas> \oren\: no sliding, but very old "portable" computers in the 80s were mostly big cubes, with most of the space being taken up by a CRT, and there's a keyboard plus floppy drive that extends from below the CRT part
23:35:37 <wob_jonas> you know, before even laptops got LCDs
23:36:23 <fizzie> I think rarely really literally cubical, more rectangular cuboids.
23:36:35 <wob_jonas> https://commons.wikimedia.org/wiki/File:IBM_Portable_PC_5155_-_Retrosystems_2010.jpg
23:36:43 <wob_jonas> ^ this one actually has an ordinary keyboard connected by cable
23:36:57 <wob_jonas> fizzie: yes, and some of them have the front side slanted
23:37:47 <wob_jonas> https://commons.wikimedia.org/wiki/File:Ibm5100_(2297950254).jpg this I think has a hinged keyboard in fact
23:38:43 <wob_jonas> (and we've done a full circle, because apparently many people nowadays are agent content to work with such tiny screens, it's just that the rest of the computer got lighter)
23:41:09 <wob_jonas> whoa. why are Prowling Pangolin and the new Bumbling Pangolin subtyped Beast? isn't wizards using that only for mythical animals, not cards based on real animals?
23:42:54 <quintopia> prowling pangolin--the mobile version of Ubuntu 12.04?
23:43:26 <quintopia> it's a bit less precise, but it gets around more
23:50:01 <wob_jonas> wow, they made a second card whose name starts with a quotation mark
23:56:18 -!- oerjan has joined.
23:58:06 <wob_jonas> Knight of the Kitchen Sink is an actually nice silver-bordered card
23:58:22 <wob_jonas> one you couldn't do in black border, but also has clean rules