🦋 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.
00:24 Manifest0 left 02:11 jpn joined 02:15 jpn left 02:24 hulk joined 02:26 kylese left 02:47 festerdam left 03:15 hulk left, kylese joined 03:50 justache- is now known as justache 04:53 jpn joined 04:58 jpn left 06:53 jpn joined
xinming Is it possible to make a raku modules mirror please? 07:21
GFW randomly blocks connections make module install a pain 07:22
08:04 abraxxa joined 08:08 jpn left 08:30 abraxxa1 joined 08:33 abraxxa left 09:09 sena_kun joined, Manifest0 joined 09:10 dakkar joined 09:12 jjido joined 09:22 Sgeo_ left 09:30 jjido left
lizmat xinming: does GFW block Github ? 09:34
09:52 jpn joined 10:04 abraxxa joined 10:07 abraxxa1 left
xinming lizmat: They don't fully block the github, But there will be random packet drop which caused the network doesn't work most time. 10:09
lizmat you could configure zef to just use the REA
xinming lizmat: What does REA mean? 10:10
lizmat Raku Ecosystem Archive
if you run zef --help 10:11
it will tell you the location of the configuration file 10:12
CONFIGURATION ....
that's a JSON file
the "Repository" section contains an array of 2 arrays 10:13
sorry, 3 arrays
if you move the array with the "shortname": "rea" to the first location, then it will first attempt to load modules from the REA, aka Github 10:14
xinming Is it possible to make a whole module mirror? 10:18
lizmat have you tried using the REA ? 10:20
or maybe I'm misunderstanding the question? 10:21
you can also git clone the REA
then you'd have the whole ecosystem locally
xinming lizmat: I saw that REA uses githubusercontent that site will be blocked randomly too. 10:22
lizmat and a git clone ? 10:23
xinming I'll dig deeper about the git clone way.
I uses proxy to install modules for now, It's working, But proxy connection is too slow, most of time wasted when I install module is caused by slow proxy link, That's why I want to make a local mirror, and I can update the mirror in the background. 10:25
lizmat a git clone / git pull would do that for you
git clone github.com/Raku/REA.git 10:37
or ssh://github.com/Raku/REA.git
xinming Thanks, will check after I go home. :-) 10:50
11:30 jpn left 12:04 Bits joined
Bits Hi, guys. 12:05
12:05 zups left 12:06 Bits left 12:15 jpn joined 12:21 jpn left 12:26 Xliff_ joined
Xliff_ \o 12:26
Can someone tell me why this is happening?
Method absolute (from Distribution::Resource) seen at:
/home/cbwood/Projects/p6-GLib/lib/GLib/Raw/Distro.pm6 (GLib::Raw::Distro), line 165
Please use %?RESOURCES<key> directly instead.
In a specific piece of code that ALL of my later code depends on, I need to get the absolute location of a Distribution::Resource. 12:27
Using %?RESOURCES<key> has proven to be problematic in this use-case. 12:28
Why is .absolute being deprecated?
Also, since this message started appearing, my parallel compiles are now failing. Compunits are now hanging while waiting for a file lock to release. Said release is not happening. 12:29
12:39 jpn joined
lizmat ugexe ^^ 12:45
I wonder whether the deprecation message is interfering with dependency recording
hmmm I guess it can't as the deprecation message is on stderr, not stdout 12:46
12:51 jpn left
Xliff_ I think the issue is that compunits are being recompiled that don't NEED to be. I have compunits I compiled 15 minutes ago being recompiled again. In the case of parallel precomp, this can be devastating since those earlier compunits are now locking files that do not need to be locked. 13:06
I'm convinced that the hangups are due to errant raku processes not releasing their file locks. 13:07
13:07 abraxxa left
Xliff_ Since I do compilings every week, this issue can't be older than the last compilations I did. 13:07
That should be 8 days ago.
13:15 abraxxa joined
lizmat could 47fdc20d73c7236f593078 be the reason? perhaps revert that 13:16
linkable6 (2024-01-10) github.com/rakudo/rakudo/commit/47fdc20d73 Deprecate the use of PERL6LIB
lizmat otherwise 225533d6f2a207f8b913 13:17
linkable6 (2024-01-07) github.com/rakudo/rakudo/commit/225533d6f2 Merge pull request #5507 from rakudo/ugexe/deprecate-more-resources-stuff
ugexe Why exactly can't you use %?RESOURCES<key>? You do exactly that a few lines down from the deprecation message reference here - github.com/Xliff/p6-GLib/blob/a491...o.pm6#L176 13:21
and why do you *need* the absolute location of a Distribution::Resource? 13:24
github.com/rakudo/rakudo/issues/5504 contains the why and links supporting it 13:27
14:17 abraxxa1 joined 14:19 jpn joined 14:21 abraxxa left 14:47 epony left 14:55 Sgeo joined 15:20 epony joined 15:23 abraxxa1 left 15:41 zups joined 15:42 guest912 is now known as wh 15:43 wh is now known as wlhn 15:49 MasterDuke joined 16:51 donaldh joined 17:37 dakkar left
Geth Raku-Steering-Council/main: 06f534eee7 | (Elizabeth Mattijsen)++ | minutes/20240113.md
Add RSC meeting minutes for 2024-01-13
 17:38
Raku-Steering-Council/main: e38ba54620 | (Elizabeth Mattijsen)++ | papers/cat_membership.md
Add Richard Hainsworth to the Community Affairs Team
 17:39
18:19 discord-raku-bot left, discord-raku-bot joined 19:10 abraxxa-home joined 19:14 jpn left 19:18 jpn joined 19:23 jpn left
tbrowder__ Hi. I would appreciate any critique of the new README on my Date::Utils module. I have been wordsmithing it for many iterations in the last week. The latest is at github.com/tbrowder/Date-Utils/, and I want to make a new release--but not until someone else puts eyes on it. A PR for correction would be swell! Thanks. 19:57
20:04 sena_kun left 20:05 sena_kun joined
antononcube @tbrowder I will get an LLM to do it using this prompt: resources.wolframcloud.com/PromptR...s/CopyEdit 20:30
20:32 jpn joined 20:36 jpn left
Actually, using the prompt "GrammarReport" is probably more interesting. resources.wolframcloud.com/PromptR...marReport/ 20:47
Here are the results I got over that README: 1. Error: The copyright year is in the future (2024). Correction: The copyright year should be the current year or a past year. 2. Error: The phrase "New since the original release is a more general routine" is awkwardly phrased. Correction: It should be rephrased for clarity, such as "A more general routine has been added since the original release". 3. Error: The
phrase "Lists of Date days stay in the proper order as shown here in code used herein" is awkwardly phrased. Correction: It should be rephrased for clarity, such as "The lists of Date days remain in the correct order, as demonstrated in the following code". 4. Error: The phrase "So we must get one of the above sequences in a first week of one to seven days" is unclear. Correction: It should be rephrased for clarity, such as
"Therefore, we must obtain one of the above sequences during the first week, which can range from one to seven days". 5. Error: The phrase "Note also each sequence is defined by its first day number, but it does not have to have its full set of days (as occurs in a partial first week)" is awkwardly phrased. Correction: It should be rephrased for clarity, such as "Also note that each sequence is defined by its first day
number, but it doesn't necessarily include a full set of days, as seen in a partial first week". 6. Error: The phrase "Those remaining days divided by seven (and rounded up by one for any partial week) yield the remaining weeks so we get the desired result" is awkwardly phrased. Correction: It should be rephrased for clarity, such as "Dividing the remaining days by seven (and rounding up by one for any partial week) gives us
the remaining weeks, leading to the desired result". 20:48
Obviously, the 1st one indicates the training date range of the model (I used OpenAI's "gpt-4".)
20:50 jjido joined
Here is how my chatbook code for those edits looks like: 20:59
cdn.discordapp.com/attachments/633...6d8b1&
Xliff_ ugexe: Why do I need .absolute? Because my code crashes without it and that particular resource isn't documented well enough for me to figure out why. 21:00
One use case is if the end user is expected to edit said resource and will need to know where in the repository it is located. 21:03
That is the long and short of it. 21:04
ugexe: So why shouldn't end-users know where their resource is located?
If this is now the expected behavior, then please tell me why I have to handle Distribution::Resource differently than IO? 21:10
They both represent a [file location] so why can't they be handled in a similar manner?
If you expect Distribution::Resource to be handled in a specific way, then could you please tell me how I can get a usable file reference from it? 21:11
ugexe because it doesn't represent an IO path, it could represent a socket for all you know 21:31
so you can only expect to consume data from it, you shouldn't expect to it to be a file or do file like things 21:32
that is why it is called Distribution::Resource and not IO::Path 21:33
if it could just be an IO::Path, then it already would be
otherwise if a user really needs to know where some resource is located then it can be accessed through CompUnit::Repository::Installation or CompUnit::Repository::FileSystem 21:35
you can expect a file from those CompUnit::Repository::* implementations
a CompUnit::Repository::GitHub on the other hand would not return a file for resources (more likely a socket or process handle), yet you could still *read* the data with %?RESOURCES<foo> 21:36
librasteve what about copy %?RESOURCES<apply/Cargo.toml>, "$apply-dir/apply/Cargo.toml"; ??
ugexe sure, but then you wouldn't be using %?RESOURCES<apply/Cargo.toml> in your code 21:37
librasteve "sure" as in don't worry that will continue to work? 21:38
ugexe right 21:39
librasteve later I go chdir<apply>; run<cargo build> so the Build.pm co-locates all the bits via copy 21:40
good - just checking!
ugexe %?RESOURCES should be viewed like IO::Handle, not an IO::Path 21:41
a path-less IO::Handle
%?RESOURCES<foo>.Str/.absolute/.relative etc also ends up actually being a security issue. Staging occurs in a tmp directory, and precompilation also occurs during this step. The precompiled is "installed" in the next step, which involves moving all the files from the tmp directory to the install location. If someone has a string representation of a resource precompiled in their code it will still 21:54
point at the tmp directory, and will run fine until whenever that tmp directory is cleared. A clever person could use this to install a malicious library to that tmp location once it gets cleared, potentially having a privileged user run their library
21:57 donaldh left
ugexe "if the end user is expected to edit said resource and will need to know where in the repository it is located." - but the user is expected to *not* do that, particularly for CompUnit::Repository::Installation 21:57
files in most repositories are considered immutable
the only repository where that isn't the case, you wouldn't need a special way to find the resource (because it will be the literal string you would otherwise use to look it up) 21:58
guifa tonyo++ for behind the scenes help 22:16
tonyo ++
tellable6 2024-01-14T13:04:15Z #raku <melezhik> tonyo: I've my comment on your grant proposal
tonyo melezhik: i do have some thoughts around that, i need to think through how to move ahead with those before i have any meaningful questions 22:17
tellable6 tonyo, I'll pass your message to melezhik
22:29 jpn joined 22:35 jpn left 22:56 sena_kun left 22:57 lizmat_ joined 23:01 lizmat left 23:03 abraxxa-home left 23:16 lizmat_ left 23:17 lizmat joined, jjido left 23:21 jjido joined 23:27 jjido left 23:36 tejr left
tbrowder__ antononcube: thanks so much. looks nice test of LLM (but I will see 23:43
23:43 jpn joined
tbrowder__ how it really fits). 23:43
23:49 tejr joined, jpn left

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!