What is Hackers' Pub?

Hackers' Pub is a place for software engineers to share their knowledge and experience with each other. It's also an ActivityPub-enabled social network, so you can follow your favorite hackers in the fediverse and get their latest posts in your feed.

tags.pub clarification questions

julian @julian@activitypub.space

<p>Hey <a href="https://cosocial.ca/@evan">@<bdi>evan@cosocial.ca</bdi></a>, I'm watching your lightning talk at FOSDEM! I'm simultaneously glad it's less than 10 minutes, but sad it's not longer too :stuck_out_tongue_closed_eyes:</p> <p>(Everyone else, want to watch it? <a href="https://ftp2.osuosl.org/pub/fosdem/2026/h2215/WNDQUQ-tags-pub.mp4" rel="nofollow ugc">Here it is</a>)</p> <p>Some questions I'm jotting down while I'm watching it [...]</p>

Read more →
0
0
0
0
4
0
0
0
0
0
3
0

よくねたのでねむい :saba:

7970eb392a (upstream/main) Add share dialog for collections (#37986)
bca57020a0 Profile redesign: Fix fields in Chromium (#37996)
9c4d11f927 Add `featured_tags` to `GET /api/v1/profile` (#37932)
5026bf6ac7 Use validation matchers for `DisallowedHashtagValidator` spec (#37636)
b09e63da87 Federate activity when remote account is added to a Collection (#37992)
951a42f491 Add `max_note_length` and `max_display_name_length` to `configuration.accounts` in `Instance` entity (#37991)
d18a47b6a7 Profile editing: Utilize new API (#37990)
51b81b3ce9 Allow remote accounts in Collections (#37989)
970ac04be7 New Crowdin Translations (automated) (#37985)

0
0
0
0
0

Our newest print: 'Starlight at Roseisle', watercolour, 29x38 cm. The Milky Way shines its cool light above the Moray Firth as we look down a sandy draw to the beach. The original painting is sold but prints in 3 sizes are available at robwighamwatercolours.com/nair gallery. I really, really enjoyed painting this one which makes a great partner to our other print 'The Highlands by Starlight!'

The Mily Way rises above the sea (The Moray Firth in Scotland), with mountains visible on the far side and WW2 beach defences on the beach. We're looking down a sandy draw on to the beach with silhouetted sea grass and bushes on the dunes.
0

LLM에서 마크다운이 널리 쓰이게 되면서 안 보고 싶어도 볼 수 밖에 없게 된 흔한 꼬라지로 그림에서 보는 것처럼 마크다운 강조 표시(**)가 그대로 노출되어 버리는 광경이 있다. 이 문제는 CommonMark의 고질적인 문제로, 한 10년 전쯤에 보고한 적도 있는데 지금까지 어떤 해결책도 제시되지 않은 채로 방치되어 있다.

문제의 상세는 이러하다. CommonMark는 마크다운을 표준화하는 과정에서 파싱의 복잡도를 제한하기 위해 연속된 구분자(delimiter run)라는 개념을 넣었는데, 연속된 구분자는 어느 방향에 있느냐에 따라서 왼편(left-flanking)과 오른편(right-flanking)이라는 속성을 가질 수 있다(왼편이자 오른편일 수도 있고, 둘 다 아닐 수도 있다). 이 규칙에 따르면 **는 왼편의 연속된 구분자로부터 시작해서 오른편의 연속된 구분자로 끝나야만 한다. 여기서 중요한 건 왼편인지 오른편인지를 판단하는 데 외부 맥락이 전혀 안 들어가고 주변의 몇 글자만 보고 바로 결정된다는 것인데, 이를테면 왼편의 연속된 구분자는 **<보통 글자> 꼴이거나 <공백>**<기호> 또는 <기호>**<기호> 꼴이어야 한다. ("보통 글자"란 공백이나 기호가 아닌 글자를 가리킨다.) 첫번째 꼴은 아무래도 **마크다운**은 같이 낱말 안에 끼어 들어가 있는 연속된 구분자를 허용하기 위한 것이고, 두번째/세번째 꼴은 이 **"마크다운"** 형식은 같이 기호 앞에 붙어 있는 연속된 구분자를 제한적으로 허용하기 위한 것이라 해석할 수 있겠다. 오른편도 방향만 다르고 똑같은 규칙을 가지는데, 이 규칙으로 **마크다운(Markdown)**은을 해석해 보면 뒷쪽 **의 앞에는 기호가 들어 있으므로 뒤에는 공백이나 기호가 나와야 하지만 보통 글자가 나왔으므로 오른편이 아니라고 해석되어 강조의 끝으로 처리되지 않는 것이다.

CommonMark 명세에서도 설명되어 있지만, 이 규칙의 원 의도는 **이런 **식으로** 중첩되어** 강조된 문법을 허용하기 위한 것이다. 강조를 한답시고 **이런 ** 식으로 공백을 강조 문법 안쪽에 끼워 넣는 일이 일반적으로는 없으므로, 이런 상황에서 공백에 인접한 강조 문법은 항상 특정 방향에만 올 수 있다고 선언하는 것으로 모호함을 해소하는 것이다. 허나 CJK 환경에서는 공백이 아예 없거나 공백이 있어도 한국어처럼 낱말 안에서 기호를 쓰는 경우가 드물지 않기 때문에, 이런 식으로 어느 연속된 구분자가 왼편인지 오른편인지 추론하는 데 한계가 있다는 것이다. 단순히 <보통 문자>**<기호>도 왼편으로 해석하는 식으로 해서 **마크다운(Markdown)**은 같은 걸 허용한다 하더라도, このような**[状況](...)**は 이런 상황은 어쩔 것인가? 내가 느끼기에는 중첩되어 강조된 문법의 효용은 제한적인 반면 이로 인해 생기는 CJK 환경에서의 불편함은 명확하다. 그리고 LLM은 CommonMark의 설계 의도 따위는 고려하지 않고 실제 사람들이 사용할 법한 식으로 마크다운을 쓰기 때문에, 사람들이 막연하게 가지고만 있던 이런 불편함이 그대로 표면화되어 버린 것이고 말이다.

* 21. Ba5# - 백이 룩과 퀸을 희생한 후, 퀸 대신 **비숍(Ba5)**이 결정적인 체크메이트를 성공시킵니다. 흑 킹이 탈출할 곳이 없으며, 백의 기물로 막을 수도 없습니다. [강조 처리된 "비숍(Ba5)" 앞뒤에 마크다운의 강조 표시 "**"가 그대로 노출되어 있다.]

As Markdown has become the standard for LLM outputs, we are now forced to witness a common and unsightly mess where Markdown emphasis markers (**) remain unrendered and exposed, as seen in the image. This is a chronic issue with the CommonMark specification---one that I once reported about ten years ago---but it has been left neglected without any solution to this day.

The technical details of the problem are as follows: In an effort to limit parsing complexity during the standardization process, CommonMark introduced the concept of "delimiter runs." These runs are assigned properties of being "left-flanking" or "right-flanking" (or both, or neither) depending on their position. According to these rules, a bolded segment must start with a left-flanking delimiter run and end with a right-flanking one. The crucial point is that whether a run is left- or right-flanking is determined solely by the immediate surrounding characters, without any consideration of the broader context. For instance, a left-flanking delimiter must be in the form of **<ordinary character>, <whitespace>**<punctuation>, or <punctuation>**<punctuation>. (Here, "ordinary character" refers to any character that is not whitespace or punctuation.) The first case is presumably intended to allow markers embedded within a word, like **마크다운**은, while the latter cases are meant to provide limited support for markers placed before punctuation, such as in 이 **"마크다운"** 형식은. The rules for right-flanking are identical, just in the opposite direction.

However, when you try to parse a string like **마크다운(Markdown)**은 using these rules, it fails because the closing ** is preceded by punctuation (a parenthesis) and it must be followed by whitespace or another punctuation mark to be considered right-flanking. Since it is followed by an ordinary letter (), it is not recognized as right-flanking and thus fails to close the emphasis.

As explained in the CommonMark spec, the original intent of this rule was to support nested emphasis, like **this **way** of nesting**. Since users typically don't insert spaces inside emphasis markers (e.g., **word **), the spec attempts to resolve ambiguity by declaring that markers adjacent to whitespace can only function in a specific direction. However, in CJK (Chinese, Japanese, Korean) environments, either spaces are completly absent or (as in Korean) punctuations are commonly used within a word. Consequently, there are clear limits to inferring whether a delimiter is left or right-flanking based on these rules. Even if we were to allow <ordinary character>**<punctuation> to be interpreted as left-flanking to accommodate cases like **마크다운(Markdown)**은, how would we handle something like このような**[状況](...)は**?

In my view, the utility of nested emphasis is marginal at best, while the frustration it causes in CJK environments is significant. Furthermore, because LLMs generate Markdown based on how people would actually use it---rather than strictly following the design intent of CommonMark---this latent inconvenience that users have long felt is now being brought directly to the surface.

* 21. Ba5# - 백이 룩과 퀸을 희생한 후, 퀸 대신 **비숍(Ba5)**이 결정적인 체크메이트를 성공시킵니다. 흑 킹이 탈출할 곳이 없으며, 백의 기물로 막을 수도 없습니다. [The emphasized portion `비숍(Ba5)` is surrounded by unrendered Markdown emphasis marks `**`.]
14
2
1

Can you run Signal with basically no iOS or Android devices right now and still do the (mandatory) prove-your-phone-id steps?

It feels like very near future I, and everyone else who cares about computing autotonomy, we're all gonna have to start runnin' the Linux phones, no matter how non-ideal that user experience is right now

@cwebberChristine Lemmer-Webber I once talked with someone who had set up a Signal account while trying not to use any Android or iOS devices. I believe part of what they did was use some emulation, including the Android SDK. There's a step where one has to scan a QR code. To get around that: the SDK includes a sandbox for testing camera functionality, and lets the user "walk" around a little "room" -- and you can put specific images on the "wallpaper".

0

Can you run Signal with basically no iOS or Android devices right now and still do the (mandatory) prove-your-phone-id steps?

It feels like very near future I, and everyone else who cares about computing autotonomy, we're all gonna have to start runnin' the Linux phones, no matter how non-ideal that user experience is right now

0
26
0

Something has changed about the unprompted responses from people seeing me using the MNT (Pocket) Reform in the last month. I'm not just getting a "whoa this is cool and nerdy", I'm hearing more people, unprompted, say "Oh, this seems important, because maybe we won't be able to get computers we control soon."

It's uncomfortable, in a way, that this is starting to become more of the zeitgeist, because I think it's true.

I really do think we need computers we can control and hack on and advance. Because we're in real trouble if we can't.

0
3
0

Something has changed about the unprompted responses from people seeing me using the MNT (Pocket) Reform in the last month. I'm not just getting a "whoa this is cool and nerdy", I'm hearing more people, unprompted, say "Oh, this seems important, because maybe we won't be able to get computers we control soon."

It's uncomfortable, in a way, that this is starting to become more of the zeitgeist, because I think it's true.

I really do think we need computers we can control and hack on and advance. Because we're in real trouble if we can't.

0
3
1
0
10
0
0
0
0
0
0
0
7
0
0
0
0

Our newest print: 'Starlight at Roseisle', watercolour, 29x38 cm. The Milky Way shines its cool light above the Moray Firth as we look down a sandy draw to the beach. The original painting is sold but prints in 3 sizes are available at robwighamwatercolours.com/nair gallery. I really, really enjoyed painting this one which makes a great partner to our other print 'The Highlands by Starlight!'

The Mily Way rises above the sea (The Moray Firth in Scotland), with mountains visible on the far side and WW2 beach defences on the beach. We're looking down a sandy draw on to the beach with silhouetted sea grass and bushes on the dunes.
0
0
17
0
0
0
1
0

Today we made some changes to . Trending posts are now evaluated over 24h to give more visibility to fresh content. We'll also be working on trending links.
Everything is available through an API. The web version makes it accessible to anyone without technical knowledge.

Website: discover.holos.social/

About the project : discover.holos.social/how-it-w

0

I've been seeing some messages saying that "AI outputs not being eligible for copyright means they can't be integrated into FOSS projects because they can't be licensed!" and I don't think that's true. Instead, I will give you several *different* reasons to reject AI generated content in your FOSS projects.

In the US at least, the absence of copyright means the public domain. The public domain itself is considered FOSS compatible with every FOSS license. Except!

The public domain doesn't exist in all jurisdictions in the same way. This is one reason CC0 was made, which is a public domain waiver with a fallback license. (CC0 is not a good idea to use with software though, separate thread about why that is.)

There *is* a related risk: internationally, there isn't agreement about whether or not whether or not AI generated content is subject to copyright yet. AND, not everywhere even has a concept of "public domain". AND, the legal status of this AIgen isn't well settled ANYWHERE yet.

0
5
0
0
0
0
1
0
0
0

I got on the NOVA show's mailing list after I was in the documentary and went to their event. I dunno why they would think it wouldn't be unsettling to read this email:

> We've noticed you haven’t opened our emails in a while, and we miss you! While you may have been engaging with NOVA in other ways, we’d love to reconnect with you through our newsletter as well.

y'all I appeared on your show talking about user empowerment don't send me emails advertising that you're using email tracking antipatterns (which my client isn't susceptible to since it doesn't load images by default btw)

0
0

@cwebberChristine Lemmer-Webber the prove-your-phone-id and ios/android requirement is precisely why i don't use signal 🙃

i really hate how certain parts of society require a phone number at all. i have a google voice number from like 17 or 18 years ago at this point that i got before i even had a phone. i would prefer not to use it, but unfortunately my medical provider requires sms 2fa and only supports sms 2fa. banks are also notorious for this.

0
0
0
0

뻘짓하지 않고 올바른 길을 살아가는것이 뭐랄까 속는것 같고 큰 보상이 없는것 같아도 나중에 수습하기 어려운 업보가 돌아오지 않는다는것이 어찌보면 그것이 큰 보상 아닌가 그런 생각을 요새 자주 함.

0
0
1
0
0
0