@djc@hachyderm.io
@djc@hachyderm.io avatar

djc

@djc@hachyderm.io

Fan of Rust, CTO Instant Domains, open source maintainer. Please consider sponsoring my OSS work on GitHub.

This profile is from a federated server and may be incomplete. Browse more on the original instance.

ekuber, to rust
@ekuber@hachyderm.io avatar

Request for feedback: how would you change this compiler error? Can you tell what's going on? What the problem is? Do you get a sense of how you might be able to solve it?

djc,
@djc@hachyderm.io avatar

@ekuber I would make it less verbose. I think it is clear enough (especially like the explicit callout that two different versions of the same dep are involved). I would skip the second block of suggestions and focus on a span pointing to the involved type (ideally the type’s definition, not just the import?) and a span pointing at the involved trait (same).

djc, to random
@djc@hachyderm.io avatar

Okay, xilem-web is pretty cool.

Edent, to random
@Edent@mastodon.social avatar

Turning to Mastodon for tech support in an effort to save my sanity.

Gmail is rejecting mail forwarded to it. But only some of the time.

I own a domain - example.com
It has SPK, DKIM, etc.
It has a catch-all email address.
Any email sent to it is forwarded to a Gmail address.
This works 99% of the time.
But some of the emails aren't delivered - with the reason:

> SMTP error from remote mail server after end of data: 550 5.7.1 [ESA] Sender blocked.

Is that a me problem or a Gmail problem?

djc,
@djc@hachyderm.io avatar

@jwz @glyph @Edent @ldodds hmm, sender rewriting (I use postsrsd) seems to work pretty well for me.

djc, to random
@djc@hachyderm.io avatar

Update your rustls dependencies today! Fortanix employees discovered a DoS issue in ConnectionCommon::complete_io(). This is not used by tokio-rustls, so most folks using that should be safe.

https://rustsec.org/advisories/RUSTSEC-2024-0336.html

(If you're on rustls 0.20 or older, you're out of luck -- those versions are no longer being maintained.)

mcc, to random
@mcc@mastodon.social avatar

I have literally implemented SRP at both the client and server side but I am still unable to figure out, if I were to purchase or set up a "Passkey", what exactly I would have, or how it would work, or which computers, web browsers or web sites I should expect it to work with

djc,
@djc@hachyderm.io avatar

@glyph @filippo @irenes @mcc I think the hardware is usually referred to as "security key" and "passkey" is pretty universally understood to be a cloud-backed credential.

Also pretty sure Google is very onboard with passkeys as a way of making something that's easier to handle for more inexperienced users to get closer to the level of security that security keys provide without some of the downsides (recovery).

See for more context https://www.imperialviolet.org/2023/07/23/u2f-to-passkeys.html (from a Google engineer).

djc, to random
@djc@hachyderm.io avatar

Writing a faster parser is cool. Writing it in C++ with Rust bindings (instead of the other way around) makes zero sense in 2024.

ekuber, to random
@ekuber@hachyderm.io avatar

Every time there's a CVE affecting some fundamental part of modern computing that Rust provides a dot-release for, it seems multiple publications find out first from the Rust blog and publish titles implying that Rust is the only affected thing. It's not only mildly annoying hearing the echos of "har, har, I thought it was 'safe'", it does a complete disservice to anyone that doesn't use Rust because they won't find out they have to update or mitigate the issue too!

djc,
@djc@hachyderm.io avatar

@ekuber also just because the Rust blog posts are better.

fasterthanlime, to random
@fasterthanlime@hachyderm.io avatar

lmao i'm in the middle of writing a video about self-hosting and vercel just 10x their prices for most people lol

thanks for making my point <3

djc,
@djc@hachyderm.io avatar

@fasterthanlime do you have a reference for this?

djc,
@djc@hachyderm.io avatar

@fasterthanlime I suppose it's https://vercel.com/blog/improved-infrastructure-pricing, but a little hard to see how you get 10x from that.

fasterthanlime, to random
@fasterthanlime@hachyderm.io avatar

is there beauty in horror? clearly yes

djc,
@djc@hachyderm.io avatar

@fasterthanlime tokio-rustls for rustls 0.23 hasn't been released yet -- there's still time to fix it!

seanmonstar, to rust
@seanmonstar@masto.ai avatar

Newest release of reqwest v0.11.25

  • A fix for Apple linker issues around system proxies

(Next release should be with hyper upgraded to v1!)

https://github.com/seanmonstar/reqwest/releases/tag/v0.11.25

djc,
@djc@hachyderm.io avatar

@seanmonstar will the hyper 1 release be on 0.11.x?

ekuber, to rust
@ekuber@hachyderm.io avatar

What would you take out of this diagnostic? Is this too much information? Would you prefer to have a shorter message at the cost of needing multiple cycles of the compiler telling you you're missing a change?

djc,
@djc@hachyderm.io avatar

@ekuber seems pretty good actually. I think there's enough moving parts here that I'd keep it like this.

ekuber, to rust
@ekuber@hachyderm.io avatar

Regular reminder that if a #rust compiler error isn't clear, we consider that a bug. File a ticket! Worst case scenario it gets closed as a duplicate. Second worst it takes us a long time to get to them. But I assure you I've read every single open A-diagnostics ticket at least twice.

djc,
@djc@hachyderm.io avatar

@ekuber is there some structure or roadmap? I do get discouraged from filing a somewhat complex to MRE diagnostics issue given the amount of open issues I’ve reported that don’t seem to get much movement. (No knock against the people working on diagnostics, but it would be nice to boost visibility on progress more!)

fasterthanlime, to random
@fasterthanlime@hachyderm.io avatar

me: don't go out of scope for this video, don't tell the history of the universe
also me: HEY REMEMBER TRAC https://en.wikipedia.org/wiki/Trac

djc,
@djc@hachyderm.io avatar

@fasterthanlime Trac was pretty good, I still miss some of the features it had on GitHub today (notably, the Timeline).

gabrielesvelto, to random
@gabrielesvelto@fosstodon.org avatar

I was reading on some notable software bugs such as the Android 14 storage SNAFU and the M3 MacBooks incapable of auto-updating.

This stuff demonstrate perfectly one issue we have with computers today: we don't need big ticket items or new features. We need polish ffs. Modern OS stacks are huge castles of cards. Stop and make them work reliably instead of ramming in new junk or redesigning the UI.

djc,
@djc@hachyderm.io avatar

@gabrielesvelto I fear that this is incompatible with VC-backed software development.

jhpratt, to rust
@jhpratt@mastodon.social avatar

Anonymous enums in are generally viewed to have a problem — what traits should be implemented for the resulting type? Why not make it explicit?

type T = U | V | W where Self: Debug + Display + Error

Surely something along this line of thinking could work?

djc,
@djc@hachyderm.io avatar

@jhpratt @ekuber now I’m thinking about enum Trait which can only be used as input in non-public APIs.

kornel, to random
@kornel@mastodon.social avatar

I'm annoyed by the rants about HTTP/2 being "bad" for having one (already mitigated) issue, as if we foolishly left a better protocol behind.
HTTP/1 had a long history of request smuggling attacks due to its deceptively tricky to parse syntax. H/1 got hit by slow loris DoS. H/1 pipelining is forever unreliable — browser vendors have given up on it after years of trying, and getting silent data corruption. The "simple" protocol has created complexity of domain sharding and cookieless domains.

djc,
@djc@hachyderm.io avatar

@kornel clearly QUIC (and thus H3!) is the right solution because it implements stream resets at the proper protocol level!

kornel, (edited ) to random
@kornel@mastodon.social avatar

Dear contributors:

do
not
reformat
the
code.

It takes a second to run the formatter, but it requires hours to review the massive patch to ensure there aren't any bad changes hiding among all the spaces and brackets moved around.

djc,
@djc@hachyderm.io avatar

@kornel enforcing canonical formatting in CI is the One True Way.

yosh, to random
@yosh@toot.yosh.is avatar

Calling people you disagree with "melodramatic", "unserious", and questioning their technical acumen is not a way to navigate technical disagreements. Those are rhetorical tactics meant to make people feel bad and get them to shut up.

Bully behavior like this has no place in the Rust community, and we need to call it out when we see it. Disagreeing is fine. Even "bad" or poorly reasoned arguments are fine. But belittling and insulting others absolutely never is.

https://without.boats/blog/thread-per-core/

djc,
@djc@hachyderm.io avatar

@yosh TBH this post feels disproportionately intense to me. The linked article describes "language I would characterize as melodramatic", which I find pretty fair when you actually read the quoted paragraph which has things like "Original Sin", "curses all your code", "kills all the joy of actually writing Rust". It then makes a claim that "our industry is rather unserious", which also doesn't refer to any person in particular.

djc,
@djc@hachyderm.io avatar

@jntrnr @yosh but the post does not call them unserious or melodramatic! It calls the expression itself melodramatic, which is a rather important distinction IMO.

Migueldeicaza, to random
@Migueldeicaza@mastodon.social avatar

I tried reading some Rust code today - nothing makes any sense, it is densely loaded in magic.

Is this the Universe’s way of telling me it’s game over and I should retire?

djc,
@djc@hachyderm.io avatar

@Migueldeicaza it’s the Universe’s way of telling you you spend a bit more time with it before you give up.

fasterthanlime, to random
@fasterthanlime@hachyderm.io avatar

oh god, rustls has the most cursed tagging scheme

https://github.com/rustls/rustls/tags

djc,
@djc@hachyderm.io avatar

@fasterthanlime why is this particularly cursed? I don’t love it (and it’s not what I use when it’s my choice) but it doesn’t seem too bad.

djc, to random
@djc@hachyderm.io avatar

“rustls does not support RSA key exchange.”

https://github.com/rustls/rustls/pull/1504

djc, to random
@djc@hachyderm.io avatar

Are you, or can you recommend, a Rust contractor (with a somewhat proven track record)? Looking for someone who has some experience doing network protocol work.

djc, to random
@djc@hachyderm.io avatar

@wez in reference to https://github.com/bluejekyll/trust-dns/issues/1708#issuecomment-1701841568, is your MTA open source/is there more context somewhere? Would be interested to potentially participate.

djc,
@djc@hachyderm.io avatar

@wez both rustls and trust-dns have Discord and I'm in your Discord server too...

  • All
  • Subscribed
  • Moderated
  • Favorites
  • •
  • anitta
  • khanakhh
  • mdbf
  • InstantRegret
  • Durango
  • Youngstown
  • rosin
  • slotface
  • thenastyranch
  • osvaldo12
  • ngwrru68w68
  • kavyap
  • cisconetworking
  • DreamBathrooms
  • megavids
  • magazineikmin
  • cubers
  • vwfavf
  • modclub
  • everett
  • ethstaker
  • normalnudes
  • tacticalgear
  • tester
  • provamag3
  • GTA5RPClips
  • Leos
  • JUstTest
  • All magazines
    •