| --- Log opened Thu Oct 21 00:00:01 2021 |
| 02:22 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has quit [Connection closed] |
| 03:06 | | * ToxicFrog pokes this with a stick |
| 03:07 | <&ToxicFrog> | I have a seq of structs. |
| 03:07 | <&ToxicFrog> | I need to filter out the first N entries in the seq such that (from-me? %) is true; N will vary across calls. |
| 03:08 | <&ToxicFrog> | I need to do this without otherwise affecting the contents or order of the seq. |
| 03:09 | <&McMartin> | So, in Haskell, that would be a call to `take N` on a comprehension filter of the seq. |
| 03:10 | <&ToxicFrog> | drop N, surely? |
| 03:10 | <&McMartin> | I misunderstood what you meant by "filter out". drop N will not do it. |
| 03:10 | <&McMartin> | I read it as "extract". |
| 03:10 | <&ToxicFrog> | Oh no, I mean "remove" |
| 03:10 | <&ToxicFrog> | That does give me an idea though |
| 03:10 | <&McMartin> | Yeah |
| 03:11 | <&McMartin> | Doing it *efficiently* would probably end up producing a "drop-if" like function with a limiter on it, so that I could connect my filtered prefix to The Rest of The List. |
| 03:11 | <&ToxicFrog> | (let [to-remove (->> xs (filter from-me?) (take N) set)] |
| 03:11 | <&ToxicFrog> | (filter (complement to-remove) xs)) |
| 03:12 | <&McMartin> | That looks vaguely right |
| 03:12 | <&McMartin> | If xs is large, though, this might do a butt-ton of unnecessary allocation. |
| 03:16 | <&ToxicFrog> | xs and N are both usually going to be <5, although in rare cases xs will have 100-200 elements in it (and N will be 0) |
| 03:16 | <&ToxicFrog> | clojure does structural sharing, though, so the number of allocations needed should be minimal |
| 03:17 | <&McMartin> | What I'm worried about is that filter will not be able to deduce that it *can* share the structure as you've written it |
| 03:17 | <&ToxicFrog> | I mean, it'll allocate a new lazy-seq, but none of the actual seq contents will be copied |
| 03:17 | <&McMartin> | Because it's going to go "oh but what if the last element of the seq is in to-remove, better copy everything along the way" |
| 03:17 | <&McMartin> | Oh, these are lazy seqs |
| 03:18 | <&ToxicFrog> | These are lazy seqs of immutable data structures |
| 03:18 | <&ToxicFrog> | So it can take a lot of shortcuts because (a) only the parts I actually need will be evaluated and (b) it knows they cannot be modified in-place |
| 03:18 | <&McMartin> | OK, this "should" just be moving pointers around inside of it then. |
| 03:18 | <&McMartin> | It will potentially do a full traversal to compute to-remove, but it *has* to. |
| 03:19 | <&ToxicFrog> | (and filter, of course, returns a new seq with the filter applied, leaving the original seq as is, etc) |
| 03:19 | <&McMartin> | (Right. I was thinking of seqs as being something more like generic Iterables and thus potentially eagerly computed.) |
| 03:19 | <&ToxicFrog> | It won't have to, in fact! It only needs to do the first N that are from-me? |
| 03:20 | <&McMartin> | Right, but what if the Nth is the last one |
| 03:20 | <&ToxicFrog> | Then, and only then, will it realize the entire seq |
| 03:20 | <&ToxicFrog> | But like, if N=2 and the first two match, it only realizes the first two elements |
| 03:20 | <&McMartin> | Right. It's just "to-remove" *does* end up eagerly evaluated, is all. |
| 03:21 | <&ToxicFrog> | Yeah, by defining to-remove only requires evaluating enough seq elements to satisfy the call to take, not the entire seq. |
| 03:21 | <&ToxicFrog> | (not necessarily the entire seq, rather) |
| 03:23 | <&McMartin> | Right |
| 03:23 | <&McMartin> | It's been a long time since I've done any Clojure, I don't have which bits are eager and which bits lazy swapped in |
| 03:24 | < Mahal> | (sidebar: I am either really good at powershell or I've forgotten something. This script worked properly first time, so I don't trust it.) |
| 03:25 | <&ToxicFrog> | I don't remember them all offhand, but most of the "seq in, seq out" functions return lazy seqs |
| 03:26 | <&ToxicFrog> | And even the ones that don't are generally zero-copy (or at least, the copy only the pointers, not the whole data structure) because immutability. |
| 03:26 | <&McMartin> | Yeah. I have much longer experience with Scheme, Python, and ML, where such things generally copy the prefix and then share the suffix. |
| 03:27 | <&McMartin> | (Though I guess Python it would act like Clojure nowadays because you'd use generators) |
| 03:29 | <&McMartin> | Python ultimately *consumes* its generators though so you don't get the repeatability there you get from Clojure |
| 03:32 | | Degi_ [Degi@Nightstar-s8e23t.pool.telefonica.de] has joined #code |
| 03:34 | | Degi [Degi@Nightstar-vtah9h.pool.telefonica.de] has quit [Ping timeout: 121 seconds] |
| 03:34 | | Degi_ is now known as Degi |
| 03:43 | <&ToxicFrog> | Mahal: that's how I feel about anything that works the first time no matter how comfortable I am in the language |
| 03:43 | <&ToxicFrog> | There's always a bug, so if it works on the first go the bug must be something really subtle and nasty! |
| 03:43 | < Mahal> | https://cloud.netlifyusercontent.com/assets/344dbf88-fdf9-42bb-adb4-46f01eedd629/ddfa7889-c131-4a6b-9773-883b2fd2fb02/03-programming-humor-opt.jpg |
| 03:50 | <&McMartin> | I'm now at the point of my career where I have a sense for a change of a certain size where it's actually OK if the tests pass the first time. |
| 03:51 | <&McMartin> | (e.g., I'm doing a simple graphical renderer right now. The hard part was getting a single triangle to appear on the screen. Once that part works, the distance from there to "a bunch of rectangles bounce around the screen" is short enough that it's plausibly might-work-first-try.) |
| 03:55 | <&McMartin> | ... though I guess that one also passes the "single screen of 80s-style BASIC code" test |
| 04:50 | | himi [sjjf@Nightstar-v37cpe.internode.on.net] has quit [Ping timeout: 121 seconds] |
| 04:51 | | Kindamoody is now known as Kindamoody|afk |
| 04:51 | | himi [sjjf@Nightstar-v37cpe.internode.on.net] has joined #code |
| 04:51 | | mode/#code [+o himi] by ChanServ |
| 05:53 | | abudhabi_ [abudhabi@Nightstar-1cji25.adsl.tpnet.pl] has joined #code |
| 05:54 | | abudhabi__ [abudhabi@Nightstar-ftopsj.adsl.tpnet.pl] has quit [Ping timeout: 121 seconds] |
| 07:18 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 07:18 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 07:20 | | catalyst [catalyst@Nightstar-np9mhl.dab.02.net] has joined #code |
| 07:22 | | catalyst_ [catalyst@Nightstar-ejd4sd.cable.virginm.net] has quit [Ping timeout: 121 seconds] |
| 07:22 | | catalyst_ [catalyst@Nightstar-ejd4sd.cable.virginm.net] has joined #code |
| 07:24 | | catalyst [catalyst@Nightstar-np9mhl.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 08:16 | | himi [sjjf@Nightstar-v37cpe.internode.on.net] has quit [Ping timeout: 121 seconds] |
| 08:24 | | himi [sjjf@Nightstar-v37cpe.internode.on.net] has joined #code |
| 08:24 | | mode/#code [+o himi] by ChanServ |
| 10:05 | | catalyst [catalyst@Nightstar-ejd4sd.cable.virginm.net] has joined #code |
| 10:05 | | catalyst_ [catalyst@Nightstar-ejd4sd.cable.virginm.net] has quit [Connection reset by peer] |
| 10:08 | <@sshine> | does anyone here experience having to share confidential stuff between your co-workers and having to compromise protocol by emailing API keys or dumping them in some Slack message? |
| 10:09 | <@sshine> | we have compliance rules that everyone is breaking because we just don't have an internal end-to-end-encrypted expiring dropbox of sorts. |
| 10:09 | <@sshine> | e.g. when a developer hands over keys that operations are supposed to put in somewhere. |
| 10:28 | < [R]> | GPG is your friend |
| 11:10 | <~Vornicus> | uch. need a real hashtable in lua. :( |
| 12:03 | | * TheWatcher runs a zendto for such things |
| 13:58 | <@sshine> | [R], GPG is on my list of candidates. sending a GPG email doesn't expire my attachment. |
| 13:59 | <@sshine> | [R], also, asking people in my company to use GPG is already futile. it's supposed to be about as easy as copy-pasting something into Slack. |
| 14:01 | <@sshine> | TheWatcher, sounds like what I want. :) it even does OAuth, which is sort of a requirement (so that you can't anonymously post something confidential, the recipient has to be tied within the company's 2FA.) |
| 14:28 | | catalyst [catalyst@Nightstar-ejd4sd.cable.virginm.net] has quit [Connection closed] |
| 14:30 | | catalyst [catalyst@Nightstar-ejd4sd.cable.virginm.net] has joined #code |
| 16:50 | | Emmy [Emmy@Nightstar-l49opt.fixed.kpn.net] has joined #code |
| 16:57 | | abudhabi_ [abudhabi@Nightstar-1cji25.adsl.tpnet.pl] has quit [Ping timeout: 121 seconds] |
| 17:07 | | abudhabi_ [abudhabi@Nightstar-1cji25.adsl.tpnet.pl] has joined #code |
| 17:32 | | catalyst [catalyst@Nightstar-ejd4sd.cable.virginm.net] has quit [Connection closed] |
| 17:32 | | catalyst_ [catalyst@Nightstar-ejd4sd.cable.virginm.net] has joined #code |
| 17:36 | | catalyst_ [catalyst@Nightstar-ejd4sd.cable.virginm.net] has quit [Ping timeout: 121 seconds] |
| 17:37 | | catalyst [catalyst@Nightstar-bq1l3m.dab.02.net] has joined #code |
| 17:38 | | catalyst_ [catalyst@Nightstar-ejd4sd.cable.virginm.net] has joined #code |
| 17:38 | <&ToxicFrog> | Vornicus: what do you need it for that the builtin table type doesn't provide? |
| 17:38 | <~Vornicus> | value-based comparison |
| 17:40 | <&ToxicFrog> | Aah |
| 17:41 | | catalyst [catalyst@Nightstar-bq1l3m.dab.02.net] has quit [Connection closed] |
| 17:44 | <~Vornicus> | or to be more precise: I need to be able to put tables in as keys and use something other than object identity to tell whether things are identical or not. I am perfectly cool with providing my own hash function but I need to be able to go "this object is a key, and it looks like this other object" and that is not a thing the built in table type can do |
| 18:05 | | Kindamoody|afk is now known as Kindamoody |
| 18:26 | <&ToxicFrog> | Yeah, that's something you can implement in lua but at minimum you're going to be faffing around with __index/__newindex to accomplish it |
| 18:29 | <~Vornicus> | yeah that's fine with me too, but I was hoping someone else had written it already |
| 18:29 | <~Vornicus> | and because there technically "is" a hashtable built in to lua's table implementation it's a little tricky to find |
| 20:18 | | Netsplit Traal.Nightstar.Net <-> Krikkit.Nightstar.Net quits: catalyst_, Alek, @McMartin, @ToxicFrog, macdjord, @Tamber, Degi, [R], Emmy, @PinkFreud, (+4 more, use /NETSPLIT to show all of them) |
| 20:18 | | Netsplit over, joins: [R], &jerith, Degi, Alek, macdjord, @himi, @gnolam |
| 20:19 | | Netsplit over, joins: catalyst_, &ToxicFrog, @Tamber, Emmy, &McMartin |
| 20:22 | | Syloq [Syloq@NetworkAdministrator.Nightstar.Net] has joined #code |
| 20:44 | | PinkFreud [WhyNot@NetworkAdministrator.Nightstar.Net] has joined #code |
| 20:44 | | ServerMode/#code [+o PinkFreud] by *.Nightstar.Net |
| 20:44 | | mode/#code [+o Syloq] by ChanServ |
| 20:44 | | mode/#code [+o ErikMesoy] by ChanServ |
| 21:31 | | JustBob [justbob@Nightstar.Customer.Dissatisfaction.Administrator] has quit [Ping timeout: 121 seconds] |
| 21:32 | | JustBob [justbob@ServerAdministrator.Nightstar.Net] has joined #code |
| 21:32 | | mode/#code [+o JustBob] by ChanServ |
| 21:37 | | catalyst_ [catalyst@Nightstar-ejd4sd.cable.virginm.net] has quit [The TLS connection was non-properly terminated.] |
| 21:43 | | JustLurk [justbob@ServerAdministrator.Nightstar.Net] has joined #code |
| 21:43 | | JustBob [justbob@Nightstar.Customer.Dissatisfaction.Administrator] has quit [NickServ (RECOVER command used by JustLurk)] |
| 21:43 | | JustLurk is now known as JustBob |
| 21:43 | | mode/#code [+o JustBob] by ChanServ |
| 21:44 | | Kindamoody is now known as Kindamoody[zZz] |
| 22:02 | | Emmy [Emmy@Nightstar-l49opt.fixed.kpn.net] has quit [Ping timeout: 121 seconds] |
| --- Log closed Fri Oct 22 00:00:02 2021 |