🦋 Welcome to the MAIN() IRC channel of the Raku Programming Language (raku.org). Log available at irclogs.raku.org/raku/live.html . If you're a beginner, you can also check out the #raku-beginner channel!
Set by lizmat on 6 September 2022.
00:00 reportable6 left, loken left 00:02 reportable6 joined
Nemokosch what's the rationale behind only looking up the modules provided by the META6.json file when one appears in a folder? This seems like an unfeature that ruins any testing if the distro (and consequently the test) has dependencies 00:32
the advice Rakudo gives (add the module to the provides section) is really harmful, even 00:34
ispyhumanfly NPM packages are similar, right? Perhaps the rationale is related to the fact that META6.json is for the "packaging" of your files, and, their needs to be a proper manifest of its deps. 00:46
jaguart I agree - the META6 supplanting -Ilib is really unexpected
ispyhumanfly I dunno 🙂
Nemokosch Let me get this straight... this is a problem whenever you want to include your own, declared dependencies in your tests. 00:47
The irony is that dependencies are not even distributions (which I would say would generally make more sense) but modules, plain and simple. 00:48
Which means that they could really be allowed straight from the same META6 file.
ispyhumanfly I see what you're saying... you're talking about the modules that are part of your own distribution, not just the deps. 00:51
It would be pretty cool if it just autodetected those like a good boy. 🙂
Nemokosch Maybe jaguart is saying that - I'm really just saying that you might REALLY want to use modules you depend on anyway, in your tests 00:52
Sure, probably one could just build some wrapper that encapsulates the depended-upon functionality only for tests - but honestly, to do this for the sake of a bad decision sounds a little awkward 00:53
01:02 linkable6 left, evalable6 left 01:03 evalable6 joined, xinming left 01:04 xinming_ left, linkable6 joined 01:05 xinming joined 01:06 xinming_ joined 01:14 lichtkind__ joined
[Coke] Can you restate what your issue is? It sounds like you have dependencies for testing, but I'm unclear what's wrong with the mechanism to declare them in the META6.json file 01:15
Nemokosch How do you declare them in the META6.json file in a way that Rakudo acknowledges for module lookup? 01:21
01:21 razetime joined 01:23 lichtkind__ left 01:30 Manifest0 left 01:54 Bocaneri joined 01:55 Bocaneri is now known as Guest9402 01:56 Sauvin left 02:05 Guest9402 is now known as Sauvin 02:07 jpn joined 02:12 jpn left
well... this is the issue. That I don't even know a way to do that. If you want to do testing on a distribution, you are suddenly limited to the modules of the "provides" section in META6.json. 02:24
02:42 rf left
ugexe Don’t declare modules used only in your tests in meta6 02:46
Put that stuff in t/lib, problem solved 02:47
use lib t/lib in the tests
Nemokosch could be that I fail to communicate properly but one thing is sure: something is seriously off with the whole communication 02:48
let's take the concrete example then. Digest and Digest::HMAC again
Digest::HMAC depends on Digest and it uses Digest in tests as well - except that it cannot 02:49
it doesn't seem appropriate to just take the source of Digest and put it somewhere (anywhere) inside Digest::HMAC 02:50
ugexe If it depends on it I don’t see why that wouldn’t work. I’d it doesn’t then something else is wrong with it 02:52
im on my phone and can’t really dig in
Meta6 supplanting lib is a security feature 02:54
its bizarre to me one could think ltherwise 02:55
there are commits somewhere (or prs) that explain all this somewhere
Nemokosch the error message about META6 surely doesn't happen figuring out what goes wrong 02:57
something else... is there a sufficiently large number of people caught by zef install --depends not actually installing the dependencies if the module itself is installed, for this to be changed? 02:58
if the module installation was rock solid, perhaps I wouldn't bring this up - actually, perhaps nobody would bring this up 03:00
but it's not. it's quite common, without any deliberate pushing of the system, that you end up in a situation where for some reason a module is installed while some dependencies are not
by the way: caught the problem here: Digest (the distribution) itself doesn't provide Digest (the module/namespace) anymore 03:09
03:42 evalable6 left, linkable6 left 03:43 linkable6 joined 03:45 evalable6 joined 03:55 razetime1 joined 03:57 razetime left, razetime1 is now known as razetime
jaguart weird inconsistency with CompUnits - if I have a CompUnit with a versioned module and two classes, I get three entries in GLOBALish - so far so good 04:01
the ordering is random - if the Module is listed first, it will have its version information available. If one of the classes is listed in GLOBALish first, the Module is missing it's version and auth 04:02
04:29 ispyhumanfly left
jaguart actually it's a bit more subtle then that - Module loses its ver/auth if preceeded by a Package inferred from a class (e.g. package Bar from class Bar::Bob) 04:31
04:54 discord-raku-bot left, discord-raku-bot joined 05:05 PolarBearXL joined 05:07 Xliff left 05:25 ispyhumanfly joined 05:30 Sgeo_ joined, Sgeo left 05:43 ispyhumanfly left 05:54 discord-raku-bot left 05:55 discord-raku-bot joined 06:00 reportable6 left 06:02 reportable6 joined 06:11 ispyhumanfly joined 06:25 jpn joined 06:45 dustinm` left 06:50 dustinm` joined 06:54 discord-raku-bot left 06:55 discord-raku-bot joined 07:55 discord-raku-bot left, discord-raku-bot joined 08:12 jpn left 08:29 jjido joined 08:34 Sgeo_ left 08:36 jpn joined 08:42 jpn left 08:51 jjido left 08:55 discord-raku-bot left, discord-raku-bot joined 09:01 abraxxa joined 09:02 abraxxa left 09:03 abraxxa joined 09:05 sena_kun joined 09:10 dakkar joined 09:55 discord-raku-bot left, discord-raku-bot joined 10:13 jjido joined 10:29 jpn joined 10:39 jjido left 10:46 jpn left 10:50 jpn joined 10:55 discord-raku-bot left, jpn left, discord-raku-bot joined 11:04 jpn joined 11:11 razetime left 11:16 Manifest0 joined 11:17 razetime joined 11:20 razetime left 11:30 razetime joined 11:31 razetime left 11:46 PolarBearXL left 11:47 jjido joined 12:00 reportable6 left, reportable6 joined 12:55 discord-raku-bot left 12:56 discord-raku-bot joined 13:09 ispyhumanfly left 13:11 Xliff joined 13:14 jjido left 13:52 rf joined 13:53 razetime joined 13:55 discord-raku-bot left 13:56 discord-raku-bot joined
ugexe Well, the error about META6.json possibly missing wasn't something I was really on board with 14:02
and yeah its wrong much of the time
github.com/rakudo/rakudo/pull/3792...-657565566 14:04
here is a comment where i explain what is wrong with that
lizmat m: say "Foo" ~~ / o :my $foo; o / # so what *is* allowed between : and ; in a regex? 14:06
camelia 「oo」
lizmat m: use Test; say "Foo" ~~ / o :pass("foo"); o / # so what *is* allowed between : and ; in a regex?
camelia ===SORRY!=== Error while compiling <tmp>
Unrecognized regex modifier :pass
at <tmp>:1
------> use Test; say "Foo" ~~ / o :⏏pass("foo"); o / # so what *is* allowe
lizmat so it's apparently only my ? 14:07
oddly enough, if you put anything else in there with RakuAST, it will work fine 14:08
the deparse just won't parse again :-(
Nemokosch To be fair, the situation is still very peculiar 14:25
there is still this schizophrenia about modules and distributions
so if you depend on "Digest", you might get a version of the distribution that doesn't even contain this module 14:26
which is the worst of both worlds
tbh regardless what we think about dependencies, modules and distributions, I assume we can all agree about one thing: 14:27
it was never a good idea to mix the names of modules and distributions
leont Mix? 14:28
Nemokosch Mix
ugexe you wouldn't depend on "Digest" and not "Digest:ver<...>:api<...>:auth<...>" in a production environment where you pin dependencies 14:32
s/and not/and instead/
there seems to be this disconnect where people don't want to pin their dependencies but they want all the benefits 14:33
Nemokosch that's neither an excuse, nor even a proper solution for all the intermixing of distributions and modules
tbh it's borderline annoying that this chaos can always be just strawmanned into pinning versions and auth - simply because hard dependencies work, no matter what resolution system there could be 14:35
ugexe i mean whats really annoying is when someone comes up with concrete assertions about various design decisions and tradeoffs after days of thought on it, as if there is no merit to the original designs. and i don't just mean about distributions 14:36
Nemokosch like yeah, and if you bundle your code with the dependencies zipped, it will always just work /s 14:37
ugexe its hard to want to go into details about some of the things you ask because you come off as insisting you are correct and invalidate the thought that likely went into the original design
lizmat ugexe: agree 14:38
Nemokosch I mean it's hard to ignore the factual problems with mixing distributions and modules
ugexe i think if you were to try to come up with why it actually is the way it is, and what things are good about that, you might answer some of your own questions and be able to make a better argument for doing things the way you think
Nemokosch and it also doesn't help that these are habitually just ignored
I mean, just try to relate for a second. I'm bringing up factual problems and you are basically saying "well, just think hard enough and you will realize why it has to be like this". These are just not on par. 14:40
ugexe lets be real: you have a pattern of doing this
lizmat ugexe agree 14:41
Nemokosch That doesn't make it a bad thing
ugexe so yes i expect *you* to think hard enough about it
Nemokosch Also, there is a pattern of dodging issues.
lizmat and there's a pattern of sealioning developing
ugexe i have a github sponsors, and if you pay me i'll handle any issue you want
Nemokosch Or blaming them on whoever. Not even necessarily me; just whoever in general. The author, the user, whoever. 14:42
lizmat: this is not sealioning whatsoever
lizmat you buy a car: you put in diesel instead of gasoline. The car doesn't work very long
is it a design issue of the car ?
Nemokosch this is a bad analogy 14:43
there is a car that takes diesel AND gasoline on an arbitrary basis
lizmat not the car you bought
Nemokosch well then we aren't talking about "the car I bought" because this car does work on an arbitrary basis, as illustrated again... 14:44
"Digest" resolves to a distribution that doesn't contain the "Digest" module 14:45
this, after being told that dependencies are modules
this goes way beyond whether I like that dependencies are modules or not
(btw I think diesel and gasoline are the same thing so it's hard to roll with the analogy) 14:46
ugexe if you have a solution to fix that without breaking anything i'm all ears
Nemokosch So we can acknowledge that this is a problem that demands a solution? 14:47
ugexe uh that isn't what i said
i'm telling you i'd entertain your solution if it doesn't break anything 14:48
Nemokosch but that does imply that there is an issue to solve, so I ask you to be explicit about it
ugexe no it implies you have an issue
Nemokosch it's hard to solve issues if they aren't even acknowledged to be issues
ugexe i'm not saying it is or isn't an issue, i'm saying that doesn't really matter for what my comment meant 14:49
Nemokosch so you think it's perfectly fine the way it is?
ugexe as i literally just typed into irc moments ago: i am open to alternative solutions that don't break anything
thats how tradeoffs work 14:50
but you seem to think there is an absolute perfect answer to everything
Nemokosch there is a very specific behavior. Can you just say if that behavior is okay or not?
ugexe so, you show me a better solution that doesn't break anything. that will give me something to reevaluate those tradeoffs for myself
i asked first 14:51
Nemokosch that's an excellent way to dig an issue 2 miles under the ground 14:52
ugexe yes you are great at that it seems
Nemokosch is this a "no u" attempt, lol
be sure that once I reach there, I will raise a suggestion about the solution, like I pretty much always do when I have access to it 14:53
the problem really isn't the attitude you are showing towards me; I have accepted this role by now. The real problem is that most of you guys have no real connection to users and people who try to get something done with this language. 14:54
These, let's say, ecosystem issues, aren't even problems I am facing, I just try to get some progress done on behalf of people who don't have infinite resources to invest into this one language. 14:55
ugexe you only show how little you know. i've spend multiple thousands of dollars on this language, countless hours writing zef for the sole purpose of having something i could confidently use in a production environment so I could try to get raku used in large companies
and that is just me 14:56
14:56 linkable6 left, evalable6 left
ugexe the problem seems to be that you think "user error" can't exist 14:56
14:56 linkable6 joined
Nemokosch I mean, good on you but that definitely shouldn't be used to downright neglect all the actual problems that a user faces. 14:57
14:57 evalable6 joined
After all, this is not meant to be a theme park of a couple of people. 14:57
ugexe you think all problems are ignored. that shows a level naivety
Nemokosch I don't think that - but I do think that those people who have the most immediate possibility to change things, or give pointers at the very least, have way too little interest in the actual users. 14:59
ugexe the people with the most immediate possibility to change things are people with a lot of free money. who are these people, and can you put them into contact with me?
Nemokosch also, in my opinion, to acknowledge an issue as something "we will have to live with for an indefinite amount of time", is still a much better perspective than either not caring, or some sort of counter-sealioning so that the pinpointable topic gets lost 15:02
ugexe you want us to care about a solution that doesn't yet exist 15:03
and i don't mean the code doesn't exist, i mean a solution at all
one that doesn't break stuff
Nemokosch I want you to care about a *problem*
"care about", as in acknowledge it at least 15:04
15:04 Sgeo joined
lizmat well, yes, It is unfortunate that the Digest distribution does not include a Digest module 15:04
stevied what is the supposed problem, exactly? "mixing" of modules and distributions? What exactly does that mean? 15:05
ugexe i do, but communicating/arguing with you takes soo many brain cycles by itself that makes me think you haven't put enough thought into it to be worth it
im telling you what you can do to get me to "care" in the way you want
lizmat but that does not have to be a problem if people would indicate :auth in their use statements or install instructions
so: now the problem is acknowledged 15:06
maybe a note should be added to the documentation about this: PRs are welcome
Nemokosch auth and ver didn't help here - they _could have_ if one pointed the exact version 15:07
lizmat well, I'm a firm believer in pinning your dependencies
it's one of the USPs of Raku over other programming languages
if you do not do that, then you're just perpetuating bad habits from the past 15:08
Nemokosch lizmat: I know that and... I hate to say that but that's also a reason your modules tend to break when another module of yours gets updated
lizmat well, if they do, I'd like to know about it
because that shouldn't happen *because* they are pinned
if nobody tells me about that, then I can't fix it 15:09
Nemokosch here's one: github.com/lizmat/Rakudo-CORE-META/issues/2
yes, you are right; perhaps it's not the version pinning per se but there is this tendency with this approach that all modules need to be updated once one is updated, and when that happens, all kinds of funky errors start happening at installation 15:11
lizmat % zef install Rakudo::CORE::META
===> Searching for: Rakudo::CORE::META
===> Testing: Rakudo::CORE::META:ver<0.0.9>:auth<zef:lizmat>
===> Testing [OK] for Rakudo::CORE::META:ver<0.0.9>:auth<zef:lizmat>
===> Installing: Rakudo::CORE::META:ver<0.0.9>:auth<zef:lizmat>
Nemokosch well, that's also a way to solve issues, I suppose 🙂 15:14
lizmat sorry, I have limited time in which I want to do a lot of things
if an issue is basically "it doesn't work" 15:15
and then when I try, it does, then I cannot draw another conclusion that either the problem has been solved
or the problem was caused by something I don't have control over
I'm *NOT* going to be spending time trying to figure out whatever the people meant when they made an issue 15:16
I'd rather spend that time on other, well described issues
or work on new developments like RakuAST 15:17
or work on rak or the irc logs
or write blog posts on top of the weekly I already do every week
Nemokosch I also have a PR for rak for that matter. Although the thing with rak for me is that I usually use it *while working* and therefore I often just don't follow up an issue if I can find a way around. 15:18
for Ecosystem: I wanted to get something done regarding the "citation index" and similarly to your way of thinking, when I saw it didn't work, I shifted towards something else 15:20
now that it turned out that it can be installed, after all, fyi: github.com/lizmat/Ecosystem/issues/3 15:24
15:26 Nemokosch joined 15:29 Sgeo left 15:43 Sgeo joined 15:52 Xliff left
[Coke] HEAD from a few days ago on my windows box just segfaulted. 15:56
(downgrading to 2022.12) 15:57
16:02 Averna left 16:34 abraxxa left
ugexe 2022.12 is regressed on windows as well from what i understand 16:38
lizmat clickbaits rakudoweekly.blog/2023/01/16/2023-...ent-radux/
ugexe skaji shows it in github.com/ugexe/zef/issues/491
i looked at the changes in between rakudo releases but nothing stuck out unfortunately 16:39
a nice thing about having an intel macbook was i could run x86 windows VMs 16:41
16:58 melezhik joined
p6steve you "can" run Docker Desktop (with Rosetta) for that .... just dog slow 16:58
16:59 jpn left
tonyo .tell ispyhumanfly can you add a key value to `$HOME/.fez-config.json`: `"bundlers": ["Fez::Util::Tar"]` and give it another shot? 17:00
tellable6 tonyo, I'll pass your message to ispyhumanfly 17:01
p6steve or my solution (what I wrote) is to go zef install CLI::AWS::EC2-Simple and then raws-ec2 --nsu launch and raws-ec2 connect 17:03
(you will need to configure your favourite Windows AMI in the yaml) 17:05
17:12 SqrtNegInf joined
SmokeMachine bisect: say "bla".encode.subbuf: ^10 17:17
bisectable6 SmokeMachine, Will bisect the whole range automagically because no endpoints were provided, hang tight
SqrtNegInf Just opened an issue regarding a segfault involving grep, GH5160. Observed on macos/linux, probably occurs on Windows too. Possibly related? 17:18
bisectable6 SmokeMachine, Output on all releases: gist.github.com/4c673d7093a027c2af...432c46ca7f
SmokeMachine, Bisecting by output (old=2018.09 new=2018.10) because on both starting points the exit code is 0
SmokeMachine, bisect log: gist.github.com/924560cc25b8aaabb1...daa18e4bb2 17:19
SmokeMachine, (2018-09-29) github.com/rakudo/rakudo/commit/cc...612996e5ec
SmokeMachine, Bisecting by output (old=2018.06 new=2018.08) because on both starting points the exit code is 0
SmokeMachine I'm testing an old code (from 2017), and it seems this behaviour ☝️ has changed... is that expected?
bisectable6 SmokeMachine, bisect log: gist.github.com/4f758539ad0cc0431c...15c595a421
SmokeMachine, (2018-08-04) github.com/rakudo/rakudo/commit/4e...931337274a
SmokeMachine, ⚠ New output detected, please review the results manually
SmokeMachine, Output on all releases and bisected commits: gist.github.com/c713f1d9ac8cf038e9...9063103b6a
SmokeMachine it used to add zeros to complete the number of elements I've asked for... 17:20
I think the original behaviour was more useful... 17:22
SqrtNegInf Actual link to the issue: github.com/rakudo/rakudo/issues/5160 17:23
melezhik El_Che - I ended up moving ci.sparrowhub.io to a new hosting (hetzner) - it is much cheaper and now sparrow ci now has even more powerful VM, let's see how build times increase ))) , but I keep your idea with oracle in mind, eventually I will probably move some of the infra component to free oracle hosting, I am just not ready to tackle with 17:24
oracle hardware at the moment ... now I will pay 24.70 euro a month for 4 VCPU / 16 GB RAM vs $53 I paid for 4 CPU / 8 GB in digital ocean, not bad ...
tellable6 2023-01-16T20:55:22Z #raku <rf> melezhik You can look at vultr too
melezhik rf - thanks - already decided to move to hetzer - thanks 17:26
Geth rakubrew.org: masukomi++ created pull request #18:
added fish shell path instructions
17:35 razetime left 17:49 dakkar left
ugexe SmokeMachine: that sounds useful for like .raku, but for .gist? 17:52
SmokeMachine ugexe: sorry, what do you mean? 17:53
ugexe adding 0s that aren't otherwise needed doesn't sound useful for .gist (the method you linked) which is just a human readable representation 17:54
SmokeMachine ugexe: the Buf generated does not have thew zeros...
ugexe github.com/rakudo/rakudo/commit/cc...612996e5ec
i'm not even sure how the linked commit has anything to do with 0s 17:55
SmokeMachine ugexe: that was not only not printing the zeros... the Blob itself does not have that
me neither...
ugexe ah the second commit looks like it could be related though 17:56
SmokeMachine m: say "bla".encode.subbuf(^10).elems # IMHO it should print 10...
camelia 3
ugexe github.com/rakudo/rakudo/commit/4e...931337274a
SmokeMachine ugexe: now that's more like it! :) 17:57
but should it add the zeros or not?
m: my $a = "bla".encode.subbuf(^10); say Blob.new: |$a, 0 xx (10 - $a.elems) # I'm having to do something like this to fix my code 17:59
camelia Blob:0x<62 6C 61 00 00 00 00 00 00 00>
18:00 reportable6 left
SmokeMachine and I'm ok with that... I just think that shouldn't be that "hard"... 18:00
18:02 reportable6 joined
ugexe kind of like .subbuf needs a way to also use docs.raku.org/type/Blob#method_allocate 18:05
SmokeMachine I have tried "bla".encode.allocate: 10; helping it would allocate and then copy... but it doesn't... :( 18:07
m: "bla".encode.allocate: 10
camelia Cannot resolve caller allocate(utf8:D: Int:D); none of these signatures matches:
(Blob:U: Int:D $elements, *%_)
(Blob:U: Int:D $elements, Int:D \value, *%_)
(Blob:U: Int:D $elements, Mu:D $got, *%_)
(Blob:U: Int:D $elements, in…
18:10 Sauvin left, Sauvin joined
ugexe m: say "bla".encode.subbuf(^10).Buf.reallocate(10) 18:11
camelia Buf:0x<62 6C 61 00 00 00 00 00 00 00>
SmokeMachine ugexe: that's good! 18:12
m: say "bla".encode.Buf.reallocate(10) 18:13
camelia Buf:0x<62 6C 61 00 00 00 00 00 00 00>
SmokeMachine any reason for Blob not have a .realocate too?
ugexe Blob is not mutable
SmokeMachine .reallocate 18:14
ugexe changing the size is a mutating operation
SmokeMachine but, couldn't it create a new Blob with that data and that size?
ugexe yes it could try to do the right thing by emulating mutating operations on a non-mutating data structure, but that also makes thing not easy to reason about 18:15
why not have Blob.push/pop etc do that? 18:16
SmokeMachine that may be only for me (and my bad english), but when I read reallocate, I intend to guess that's creating a new structure...
18:17 melezhik_ joined
ugexe so what then would be the difference between "allocate" and "reallocate" 18:17
allocate:"Returns a newly created Blob object", reallocate: "Change the number of elements of the Buf, returning the changed Buf." 18:19
SmokeMachine allocate allocate the memory and works only for :D, reallocate works only for :D and allocates the memory copping the data
I don't know, but in my mind that sound to me like C's malloc and realloc... 18:20
ugexe "realloc() is a function of C library for adding more memory size to already allocated memory blocks" 18:21
18:21 melezhik_ left
SmokeMachine but if it has no enough continuous memory and you are trying to make it bigger it will allocate the memory on a different place (at least that's how I remember it... but I could be wrong...) 18:23
and copy the content
18:24 notna joined
ugexe my C is rusty but i would think it would only be copying if it needed a bigger e.g. continuious memory block or something, and if it didnt it would just expand the memory to the new size 18:26
SmokeMachine yes, that's my understanding also... 18:28
18:48 melezhik left
[Coke] rolling back my windows box to 2022.07 ... will try to track down the windows issue this week 18:51
rf SmokeMachine: I think it reallocates a new buffer of a different size, bufs are a fixed size in Raku 19:02
19:41 Sauvin left 19:42 notna left 19:43 Sauvin joined, jjido joined
[Coke] would it make sense to consider making --/rea the default for zef? 19:43
ugexe once enough of all modules are in fez, yeah 19:49
but for now rea is the source of the p6c and cpan distributions 19:51
19:54 discord-raku-bot left, discord-raku-bot joined
[Coke] Ugh, ok, not the default: but I have to use --/rea to install File::Directory::Tree on windows, but then have to allow rea to install the rest of these deps. 19:54
ugexe why? path length issues? 19:57
19:58 abhinav joined, abhinav left
ugexe even then... i dont understand why --/rea would help 19:58
that module is *only* on rea 19:59
--/rea --p6c would find it, but thats an extra flag 20:00
20:04 vrurg joined, vrurg_ left 20:09 Sauvin left 20:10 Sauvin joined
[Coke] I don't know why, but it works. 20:11
... OK, it's not necessarily a recent change causing my issues today; reinstalled 2022.07 - getting a segfault in a raku script 20:12
20:30 jgaz joined 20:33 Sauvin left 20:39 Sauvin joined 20:40 Sauvin left, Sauvin joined, jpn joined 20:48 jpn left, jpn joined
leont would switch everything over to fez the moment it connects with my keychain manager 20:59
21:04 jpn left
ugexe i too am lazy. i've been waiting till i make a tool to upload to fez from my github releases 21:13
all my repos use circle-ci though, and with that security issue its probably time i do another round of updates on my repos. maybe i'll do it then 21:14
21:26 dogbert17 joined
tonyo i've been speccing out making fez crawl known repos for vX.Y.Z tags to _try_ to automate releases.. the problem is when that tag changes (eg, delete/retagged) and when the tag isn't indexable (eg bad meta) 21:48
that might get delayed more, still trying to decide whether fez should move ahead with becoming a full on dist management tool or remain just an indexing tool 21:49
SmokeMachine unlink "bla"; my $f = "bla".IO.open: :a, :bin; for ^33 { my $a = .Str.encode; $f.write: Blob.new: |$a, 0 xx (256 - $a.elems) }; my $r = "bla".IO.open: :r, :bin; $r.seek: 1 * 256, SeekFromBeginning; say $r.read(256) 21:50
m: unlink "bla"; my $f = "bla".IO.open: :a, :bin; for ^33 { my $a = .Str.encode; $f.write: Blob.new: |$a, 0 xx (256 - $a.elems) }; my $r = "bla".IO.open: :r, :bin; $r.seek: 1 * 256, SeekFromBeginning; say $r.read(256)
camelia Buf[uint8]:0x<>
SmokeMachine m: unlink "bla"; my $f = "bla".IO.open: :a, :bin; for ^34 { my $a = .Str.encode; $f.write: Blob.new: |$a, 0 xx (256 - $a.elems) }; my $r = "bla".IO.open: :r, :bin; $r.seek: 1 * 256, SeekFromBeginning; say $r.read(256)
camelia Buf[uint8]:0x<31 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 …
tonyo that might turn into fez peeling a bit of functionality back and the API being documented so that other tooling can integrated
SmokeMachine that seems very wrong ^^
am I doing something wrong? if I write it 34 times it works, but if I do that 33 times or less it doesn't work... 21:51
it seems to me a very odd bug...
gfldex [Coke]: I don't think the Long Now Foundation will aprove of `<year>=[ \d ** 4 ]`. :-> 21:58
SmokeMachine why when the file is small it reads nothing from it but it reads ok when the file is bigger? 22:04
gfldex SmokeMachine: what is the size of the buffer in bytes? 22:05
also, see: docs.raku.org/type/IO::Handle#method_flush
SmokeMachine hum! flush! that would make sense! 22:06
yes, that was it... a missing flush... Thank you! 22:07
gfldex barely manages to supress an inappropriate joke about the mens rest room 22:09
22:14 Nemokosch left 22:28 jgaz left 22:38 jjido left 22:41 sena_kun left
tonyo in a way, very appropriate 22:45
22:55 loken joined 22:56 epony left 22:58 discord-raku-bot left, discord-raku-bot joined 23:02 loken left 23:03 loken joined 23:17 loken left 23:36 heartburn joined
tonyo CIAvash: i put a logger in the meta updater if you send another request and i'm afk please lmk and if it's not too big an ask to pm me what you're trying to update to (and which auth you're updating if not ciavash) 23:41