<FromGitter>
<straight-shoota> @Rx14 I've added the issue with user32.lib to the wiki
<RX14>
ok
<FromGitter>
<straight-shoota> until we find a proper solution
<FromGitter>
<straight-shoota> to build gc without it
<crystal-gh>
[crystal] lukeasrodgers opened pull request #5525: Define `new(JSON::PullParser)` on BigDecimal so it can be deserialized (master...json-parse-bigdecimal) https://git.io/vNe7X
<crystal-gh>
[crystal] RX14 opened pull request #5526: Remove unused sprintf function in the compiler (master...refactor/compiler-sprintf-cleanup) https://git.io/vNe5q
<hightower3>
Hey, I want to have something like: enum Levels; One; Two; Three; end. And then later I need some macro work like {% for level in %w{one two three} %}. What's my best approach to unify this so that I only specify levels in one place and use them for both purposes without don't repeating them?
<hightower3>
Got it handled
illyohs has joined #crystal-lang
<FromGitter>
<jwaldrip> What did `macro def` do?
<FromGitter>
<jwaldrip> nevermind found the old docs.
rohitpaulk has joined #crystal-lang
rohitpaulk has quit [Ping timeout: 248 seconds]
hightower3 has quit [Remote host closed the connection]
rohitpaulk has joined #crystal-lang
ashirase has quit [Ping timeout: 265 seconds]
ashirase has joined #crystal-lang
rohitpaulk has quit [Ping timeout: 252 seconds]
ashirase has quit [Ping timeout: 256 seconds]
ashirase has joined #crystal-lang
aroaminggeek has joined #crystal-lang
<crystal-gh>
[crystal] opiation opened pull request #5527: Added support for secure web sockets when protocol is HTTPS (master...playground-wss-front-end-support) https://git.io/vNvqx
<FromGitter>
<jwaldrip> Can someone explain the .dwarf outputs I am seeing in 0.24.1
aroaminggeek has left #crystal-lang ["Textual IRC Client: www.textualapp.com"]
illyohs has quit [Ping timeout: 240 seconds]
rohitpaulk has joined #crystal-lang
rohitpaulk has quit [Ping timeout: 250 seconds]
flaviodesousa has joined #crystal-lang
claudiuinberlin has joined #crystal-lang
mark_66 has joined #crystal-lang
<crystal-gh>
[crystal] jhass closed pull request #5526: Remove unused sprintf function in the compiler (master...refactor/compiler-sprintf-cleanup) https://git.io/vNe5q
dragonkh has quit [Remote host closed the connection]
<livcd>
i lost track on what happened
<livcd>
glad i do not care about infosec
rohitpaulk has joined #crystal-lang
dragonkh has joined #crystal-lang
<FromGitter>
<bararchy> livcd basiclly, bug in Intel\ARM\AMD CPU's that allows local attacker (on same machine via VM or app) to get info he should not get.
<FromGitter>
<bararchy> the fix will disable a feature that will cost around 5-30% CPU performance, mainly select operations and shared memory access
<FromGitter>
<bararchy> this is why VMs are primery target for this performance hit
<livcd>
ah yeah this I read but did not know it affects ARM and AMD as well
<FromGitter>
<bararchy> yeha, it's a CPU global issue
<FromGitter>
<bararchy> every CPU from 95 to 2018
<livcd>
does it affect sparc and powerpc ?
<FromGitter>
<imonmyown> @bararchy so if I run a local VM, I am also affected?
<FromGitter>
<bararchy> yeha, if the attacker get control of the local VM (even as "guest") he can steal info from the host machine
<FromGitter>
<mjago> ouch 😱
<FromGitter>
<drosehn> There are two different issues. One seems to be Intel-only, the second one shows up on AMD and ARM chips. The fixes which are rolling out (and which will cause a performance hit of up to 30%) are for the Intel-only bug. There is no quick fix for the second problem.
<FromGitter>
<drosehn> I should say that the second one shows up on AMD and ARM -- in addition to Intel.
<FromGitter>
<asterite> jwaldrip: dwarf file (only on Mac) contains debug info so you can see nice stack traces on exceptions
<FromGitter>
<mjago> @asterite on a mac have you any idea why I get about 20 of the following warnings when compiling on a mac whithout --no-debug ? ⏎ ⏎ ```ld: warning: unknown dwarf DW_FORM_strp (offset=0x00000080) is too big in E-xception.o``` [https://gitter.im/crystal-lang/crystal?at=5a4e059129ec6ac311b932a5]
faustinoaq has quit [Quit: IRC client terminated!]
<RX14>
@sdogruyol not that I'm aware of
<FromGitter>
<sdogruyol> I'm just doing that
<RX14>
Might want to bring it up in the manas chat
<RX14>
Oh ok
<FromGitter>
<sdogruyol> I'll do so after
faustinoaq has joined #crystal-lang
faustinoaq has quit [Remote host closed the connection]
hightower3 has joined #crystal-lang
hightower2 has quit [Ping timeout: 248 seconds]
faustinoaq has joined #crystal-lang
alex`` has quit [Quit: WeeChat 1.9]
<Vexatos>
Now, here's a question: I've seen PR #5058 where @asterite states that LLVM 5.0 should not be the default LLVM version used, but I would still like to add it to the makefile so that llvm-config-5.0 is automatically detected on debian-based systems. Which priority should it have? For instance, should it be right below llvm-config-4.0, or at the very bottom?
<crystal-gh>
crystal/master c663f52 Chris Hobbs: Add Process.exit stub on windows (#5524)...
<FromGitter>
<asterite> We could remove support for older llvm versions that we know aren't used in anything we distribute (or arch, freebsd, etc.), specially because they are buggy
<RX14>
IIRC 3.5 support is gone
<RX14>
and we know 3.8, 3.9 have the --no-debug bug
<crystal-gh>
[crystal] Vexatos opened pull request #5531: Support LLVM 5.0 postfixed versions (master...llvm-5.0-support) https://git.io/vNvKJ
<Vexatos>
Here we go then
<crystal-gh>
[crystal] asterite opened pull request #5532: Allow nested constants in type lookup for generic arguments that resolve to integers (master...bug/5426-constant-resolution) https://git.io/vNvK7
faustinoaq has quit [Quit: IRC client terminated!]
faustinoaq has joined #crystal-lang
faustinoaq has quit [Client Quit]
<crystal-gh>
[crystal] RX14 closed pull request #5532: Allow nested constants in type lookup for generic arguments that resolve to integers (master...bug/5426-constant-resolution) https://git.io/vNvK7
<crystal-gh>
[crystal] RX14 closed pull request #5527: Added support for secure web sockets when protocol is HTTPS (master...playground-wss-front-end-support) https://git.io/vNvqx
rohitpaulk has quit [Ping timeout: 248 seconds]
<crystal-gh>
[crystal] RX14 opened pull request #5533: Implement snprintf on win32 (master...feature/windows-snprintf) https://git.io/vNvQE
<crystal-gh>
[crystal] bew opened pull request #5537: Fix typo in benchmark spec (master...benchmark-spec-typo) https://git.io/vNfZ0
aroaminggeek has left #crystal-lang ["Textual IRC Client: www.textualapp.com"]
<oprypin>
bew, i dunno, in times when CI is struggling, maybe changing 1 character at a time is not the best idea
<FromGitter>
<bew> lol you're talking about typo PR is not a good idea right?
<crystal-gh>
[crystal] RX14 opened pull request #5538: Workaround llvm.powi.* intrinsics not available on win32 (master...feature/windows-powi) https://git.io/vNfcd
<RX14>
but the majority of compiler time is still spent in non-llvm parts of the crystal compiler
<FromGitter>
<bew> is there a simple way to replace the first char of a string by another char (ascii or unicode)? I know strings are immutable but I'm not sure how to build the new string efficiently with only 1 char replacement
<Groogy>
ah
<RX14>
@bew not sure if there's a sane way to do that
<FromGitter>
<asterite> those `sub` and `gsub` names were copied from ruby... maybe we should do like Elixir and choose good names for everything, even if that means looking less like Ruby
<RX14>
@asterite yeah i've thought that a bit
<RX14>
I think keeping a few of the most common ones is good
<FromGitter>
<bew> oh perfect, I missed the sub with index 👌
rohitpaulk has joined #crystal-lang
rohitpaulk has quit [Ping timeout: 240 seconds]
<RX14>
"Finished in -2116114 microseconds"
<RX14>
there's some kind of bug here
<FromGitter>
<asterite> or you optimized it A LOT
<jhass>
well or it's really really really slow
<RX14>
lol
<RX14>
it's intermittent
<RX14>
it's usually about
<RX14>
2.2 seconds
<RX14>
i.e. 2200000 us
<RX14>
so somehow it's get it's negative sign flipped
bazaar_ has quit [Quit: Reconnecting]
<RX14>
no idea how thats possible
bazaar has joined #crystal-lang
<FromGitter>
<Fryguy> a signed 32bit integer has a max of 2_147_483_648, which happens to be 2.14 s in nanoseconds
<FromGitter>
<Fryguy> wonder if something is using nanoseconds, going over that a little, and setting the top bit, which is being interpreted as signed
<FromGitter>
<Fryguy> although I would expect a small negative number
greengriminal has joined #crystal-lang
<FromGitter>
<elorest> I like sub and gsub. They make about as much sense as the methods in other languages and they're short.
<RX14>
nice
<RX14>
so I can ssh into WSL
<RX14>
and from there I can run all I need to to cross-compile windows
<RX14>
it's nice not having to deal with windows proper
claudiuinberlin has joined #crystal-lang
<RX14>
actually I can make cross-compile and run all one command...
<FromGitter>
<yxhuvud> Sub and gsub *are* good names *if* you have a certain background. Now, is it possible to make good readable and terse names that is more inclusive? Beats me.
jnyw has joined #crystal-lang
<oprypin>
first of all, sub is not useful
<oprypin>
`replace(limit: 1)`
greengriminal has quit [Quit: Leaving]
<FromGitter>
<elorest> <oprypin> (sub)stitute makes a lot of sense in english. `sub` is used in almost all professions to mean to fill for or replace something else.
<FromGitter>
<elorest> The guy fixing my heater just said yesterday that he subs in coleman parts when he can't order bryant in time. My sons teachers have been out of town this week so he has had subs.
<robacarp>
man it's good to have line numbers in traces again