idoubtit

That's interesting. I haven't used fail2ban for a long time, but reaction is worth evaluating. Unfortunately, that post does not describe their full configuration. Maybe it's on purpose, so that attackers can't adjust to fit.

My experience is that modern web scraping had no obvious pattern, since it is proxied through many IPs. The last time a server was failing to handle the pressure, we decided to temporarily ban IPs from some Asian regions. How does the FSF decide to ban an IP?

Why do they use iptables + ipset instead of nftables? Is there a technical reason or is it just legacy? AFAIK, Nftables is more performant, and IMO simpler. And it has native sets, see https://wiki.nftables.org/wiki-nftables/index.php/Sets

After a cursory read of the PDF, my impression is that the methodology is sound, but the results are blown out of proportion. Of course, if the title was "Boko Haram's internal hearsay about their use of AI", it would draw much less attention.

The weak part is that the interview were with only 15 persons that had knowledge about AI. But, from what I understand, but they never used it themselves. Only the top commanders and the specialized units could send prompts. So it's hard to guess what is the real AI use from a few indirect statements. For example, the commanders could have decided to spread the rumor they were using AI a lot, even if they mostly used plain web search, because they thought it would boost the morale.

For instance, why would anyone pay an AI service to get basic help like that:

AI provided both immediate technical fixes by teaching “how to uncouple the gun by washing it with diesel” and tactical guidance, in terms of “how to change the military formation so that fighters with jammed guns move to the back and others take their positions until the problem is solved.”

BTW, the paper does explain that Boko Haram was initially just a plain sect, rather living peacefully. Then "following a violent government crackdown and Yusuf’s death in police custody in 2009, the movement turned into a jihadist insurgency". And the last time I read a report by Amnesty International about the conflict, it estimated that 55 % of civilian casualties were caused by the terrorist group, and 45 % by the security forces. The Nigerian army sometimes razed whole villages. Like always, the world is not black and white, good guys and bad guys.

If you opened the OP link, you missed the section "Featured in The New York Times".

your definition of dead is elms definition of stability

If Elm's definition of stability is keeping bugs and runtime errors for years, then I'm glad I stopped using Elm long ago.

Not only were the issues unaddressed, but for the past years the PR got no human response. For instance this one¹ fixes infinite loops in the core. [¹]: https://github.com/elm/core/pull/1137

I also use OsmAnd~ for hiking, and even more for cycling. I tried CoMaps after a friend recommended it. In my experience, OsmAnd is much more powerful, and if you can tolerate its complexity and battery-drain, it's worth it.

- CoMaps is much simpler, but lacks features. OsmAnd has (too many) features and requires heavy configuration, e.g. to disable some features I never use and optimize battery life.

- Loading a track (GPX) in CoMaps is cumbersome (you have to "Open with..." from a file navigator). OsmAnd can load tracks from within the app, and just display them or turn one into a route. With CoMaps, I couldn't load more than one on a map. My GPX routes are prepared with BRouter on my desktop computer, since both apps are unable to build routes of 150 kms

- CoMaps is better at energy saving. OsmAnd drains my battery in 5-6 hours when the screen is on (the screen roughly accounts for half of the drain). YMMV. For long rides, I switch the phone off as soon as I don't absolutely require its help for navigation.

- OsmAnd has features that I would never give up. Notably, displaying markers on "drinking water" POI, so that I know when I'll be able to refill, or decide if I must deviate my route toward a drinking tap.

- There are more maps and data in OsmAnd than in CoMaps, and more visualisation options.

- Last time I tried CoMaps (from F-Droid), at some point the screen went blinking, alternating between home and CoMaps. The only way out was to shutdown the phone.

[edit: BRouter for routing]

podman can work with kube files, which is the YAML format from Kubernetes. That's more "platform agnostic" than docker-compose. And it can be read as a documentation in the same way.

Of course quality is subjective, but my own stats on AniDB show no decline.

I've watch at least one anime produced every year from 1977 to now. For 2000-2025 I've watched 24 to 62 anime of each year. My average vote by year is surprisingly stable at 5 ±0.6.

My top votes also don't show any significative tendency. Out of my 25 favorite anime, 6 were produced in 2020 or later. Notably the film "Kaguya-hime no monogatari", but also seinen series like "Nami yo Kiite Kure", "ACCA" or "Eizouken".

BTW, "Eizouken" (2020) is wonderful, and it's about a young girl wishing to become an animator, and how she creates short animes with her friends. I strongly recommend it!

My father once received a letter from Algeria, with 3 words on the envelope : his first name, "Créteil" (the town where he lived, ≈100k inhabitants), and "France". Of course, in the 70s there was no Internet nor central database to find him, yet the postal service managed to deliver the letter. He was a very active social worker, managed a youth football team, etc. which made him locally well-known by his first name.

Nowadays, many people can't find anyone or any place unless their phone helps them. And postmen never stop to chat. Such a letter would not pass through the technology process, and probably not through the human network.

I also think these are harmful simplifications, for several reasons.

First of all, I'm skeptical about the study that proves that people seeing three trees have better mental health. There are so many factors that it's hard to separate one. A solid study would compare families living in the same building, roughly at the same floor, and with similar parameters (family size, income, education, street noise, etc.). Comparisons from different buildings induce too many side factors. I think that collecting this sample would be very hard. I can't access the full-text behind the paywall, so I don't know their methodology, and their abstract is vague, so I fear the paper is meaningless.

Then do people really watch that much through their windows? I'd be surprised that having a glimpse of a few trees at home once a day could change anyone's life.

Even if trees did has a positive impact on mental health, I suppose inciting people to bike or walk (at least partly) to workplaces and stores would dwarf that impact, for mental and physical health.

Lastly, the 30% of tree cover seems arbitrary. For the same percentage, would covering every street with trees have the same impact as keeping trees inside parks? I think the goal to provide places where people go for a walk requires different solutions than the goal to reduce the heat in a concrete jungle.

There were reason to believe they were less subject to US juridiction: their Subscriber Agreement is for "Sectigo Limited, a limited company formed under the laws of England and Wales". See https://www.sectigo.com/uploads/backgrounds/Certificate-Subs...

Sadly, their United Terms and Conditions in section 8.2 are even more restrictive than LE's. They reject any entity "located in, incorporated under the laws of, or owned (meaning 50% or greater ownership interest) or otherwise, directly or indirectly, controlled by, or acting on behalf of, a person located in, residing in, or organized under the laws of any country sanctioned under the laws of the U.S. or E.U." See https://www.sectigo.com/uploads/backgrounds/United-Terms-and...

From a layman point of view, it could even mean that the ICC and the UN are prohibited from using Sectigo. The Customer must have no "affiliates, officers, directors, or employees" that are on sanction lists, and the US have sanctioned some high-profile members of the UN and the ICC that spoke about the genocide in Gaza.

Couldn't LE have a branch in Europe or anywhere outside the USA and its minions?

Because they're betraying their own goals, as stated in their About page: “It is a service run for the public’s benefit. [...] Anyone who owns a domain name can use Let’s Encrypt to obtain a trusted certificate at zero cost. [...] Let’s Encrypt is a joint effort to benefit the community, beyond the control of any one organization.” Now they own they are under the control of a political organization.

Here is the paragraph Let's Encrypt added to their Subscription Agreement on 2026-06-04:

You are not a person or entity that is:

(a) located in, organized under the laws of, or ordinarily resident in any country or territory that is the target of comprehensive U.S. sanctions;

(b) a prohibited or restricted party under U.S. or other applicable sanctions and export control laws and regulations;

or (c) owned or controlled by or acting on behalf of anyone described in (a) or (b).

You agree to use Let’s Encrypt Certificates and any services provided by or on behalf of ISRG in compliance with applicable U.S. export control and sanctions laws and regulations.

The problem with this article is that the benchmark method they use is flawed. The documentation of zshbench explains why: https://github.com/romkatv/zsh-bench

Even with a low grade laptop, my zsh config grants me a sub 5ms prompt and a sub 1ms input lag, and that's far more important than the exit time.

     ./zsh-bench
    ==> benchmarking login shell of user XYZ ...
    creates_tty=0
    has_compsys=1
    has_syntax_highlighting=0
    has_autosuggestions=0
    has_git_prompt=1
    first_prompt_lag_ms=54.942
    first_command_lag_ms=57.069
    command_lag_ms=4.275
    input_lag_ms=0.669
    exit_time_ms=26.522

     hyperfine --warmup 3 'zsh -i -c exit'
    Benchmark 1: zsh -i -c exit
      Time (mean ± σ):      26.5 ms ±   0.5 ms
      Range (min … max):    25.5 ms …  27.6 ms

Not too bad? A hidden procedure with ten clicks, which the user has to repeat for each web browser. And it may break at any time if the browser changes some details. Or if KDE changes. And it's specific to KDE, with no alternatives in most Wayland WMs.

All that for _one_ feature which works out-of-the-box with Xorg, and which Wayland removed for security reasons. From what I've seen, sharing the screen is another common feature which was broken with Wayland and is still painful.

I don't think Wayland's security model is very relevant to me since I have faith in Debian for filtering out rogue applications. So I have to reason to drop my smooth UX for a world of "not too bad" workarounds.

No, more like:

    git rebase -i
    # squash all the commits (e.g. in vim with ctrl-v)
    git reset HEAD^
    git add -p
    # interactively pickup the RED hunks
    git ci -m RED

There's one problem with arrays that I haven't seen mentioned here or by the OP: when inserting a key-value, the type of the key may change. For instance ["4" => "four"] === [4 => "Four"]

This can lead to some unexpected behaviors. For example, I've already been bitten by `array_merge()` whose result is different if its parameters are arrays with numeric indexes.

    array_merge(["4 " => "four"], ["5 " => "five"])
    // ["4 " => "four", "5 " => "five"]

    array_merge(["4" => "four"], ["5" => "five"])
    // [0 => "four", 1 => "five"]

You can do that. Of course, PHP's native types are quite limited, but a phpdoc syntax should work with static analysis tools. For instance:

    /** @psalm-type MyobjType = object{mystring: string} */

    /**
     * @param MyobjType $myobj
     */
    function (object $myobj): void

- For Psalm, see https://psalm.dev/docs/annotating_code/type_syntax/utility_t...

- For PHPstan, see https://phpstan.org/writing-php-code/phpdoc-types

It may work in your IDE (autocompletion, etc.) but there is no standard on this side. Some IDE have their own parsers, others use one of the LSPs for PHP.

What is the open standard?

As far as I know, the ISO standard for zip only specifies two compression methods: "store" (no compression) and "deflate". If I follow that, when I create a zip file, I know it's not performant, but at least it's almost universal (except for file ownership, permissions, character encoding and anything modern).

The corporate PKWARE has added other compressions to their original zip software, but those are not in the standard. They will not work for an EPUB, a LibreOffice file, etc. If I want a good compression, I reach for zstd (often through `tar`) or 7z if I want more portability.

I had no heard of ECT, but I'm not impressed. I've just benchmarked it against two others PNG optimizers, and here are the file sizes for default and max levels:

    1985457 oxipng-o6.png
    2030036 oxipng-o2.png
    2125459 ect-o9.png
    2144598 ect-o3.png
    2169351 optipng-o7.png
    2215086 optipng-o2.png
    2218326 original.png

    oxipng 9.1.5
    OptiPNG version 7.9.1
    Efficient Compression Tool Version 0.9.5

I tried to apply ECT to a few .gz files, but it complained it was not compatible, and I did not dig further.

[edited for a typo s/I/it/]

Not the GP, but I also have mixed feelings about Standard Ebooks. They modernise texts for American readers. This means changing the punctuation, merging some words, altering the syntax, etc.

When I read an old novel, written two centuries ago in England, the little differences to modern English are part of the charm, and I certainly don't want any Americanism mixed in. For one of my favorite novels, The Forsyte saga, the author deliberately used some rare forms of words, which SE replaced with the mainstream forms.

As pointed in your link, NetBSD achieved this with some help from Debian. If I understand correctly, it's not that NetBSD tried harder, it's that their problem was easier: fewer packages which change less (they still use CVS, "stability" is an understatement!).

BTW, most Debian packages have reproducible builds. Those which have not (I'd say 5%) are shown in orange in the graph there: https://wiki.debian.org/ReproducibleBuilds

I expected a toy project, but it is a usable library, which required a lot of work. Good job on delivering. A few comments:

After reading "composer.json", I thought that the tests used a custom framework. I'm glad the project does not suffer from NIH syndrome, but the dev dependency on PHPUnit should be declared.

There should a warning that it's only meant for some Western Latin languages. The normalization of the input is built on a character table for a handful of cases. That's not enough for some Latin languages, e.g. Turkish. And any input with Cyrillic, Arabic, CJK and so on, will be ignored.

There is no Unicode normalization or cleanup. Real-life input have many corner cases, e.g. diacritics next to the characters, or invisible characters inside a word to prevent hyphenation. Unless I'm mistaken, this engine would treat the NFD form "fête" as "fe te", instead of the expected "fete", which the NFKD form "fête" produces. I suggest using ext-intl for Unicode normalization, at least as an option.

Lastly, I can't think of a use case for this library. I've always had access to some external service (MySQL, Postgresql, Manticore Search, Solr, etc.) or to a PHP extension for a local Sqlite with FTS. Even for hobby projects, I haven't deployed to a shared hosting for more than two decades.

If you have trouble with the default software on a Kobo ereader, you can install other applications aside it, then switch to them after boot. In my experience, the installation process is innocuous and straightforward.

I use Koreader: after experimenting with various configuration parameters for a few days, the UI is now stable and tailored to my taste. Once in a while, I switch to another app: Plato is better at handling huge PDF files.

Another bonus point is that I can mount my ereader as a USB mass-storage and rsync the git repository of my ebooks onto it.

The training of one LLM requires as much emissions as 17,000 people over a year. Which, according to the article, is 8 times more than last year, and may be underestimated by a factor 2.

That does not cover the whole usage: the hardware, the bots that collect learning data, the prompts, etc. And there are now many models of this size, and thousands and thousands at smaller sizes. And some of this parameters are increasing.

AI is estimated to emit more than 80e6 tons of CO2-equivalent this year. Much more than whole countries like Austria or Israel. Is that trivial?

The chain of facts makes me sad:

1. The French government announces its digital agency is to write a plan, by the end of the year, so that France could reduce its extra-European dependencies. The communiqué is wrapped up with minor facts (e.g. the digital agency is to switch to Linux on dozens of computers) and big promises from Ministers.

2. Various news sites state that "France is ditching Windows", at least in their titles.

3. On new aggregators, most people react to the titles. Some do read the articles. Very few realize it's about promises to act toward a vague goal, with an unknown calendar, and many political uncertainties.

I would have hoped for more cautious reactions. It's not a leading act, not a reason to be proud, not a example to follow. It's just words.

The French government already made similar promises in the past. Sometimes, it did happen, like the Gendarmerie (rural police) switching to a Linux distribution. Sometimes, it didn't, like the pact signed by the Army Ministry with Microsoft in 2022: many clauses are still secret, even the prices.

The title is very far from the actual public statement that is linked in the article.

The French government announced that its digital agency will switch to Linux during this year. This is about a few hundreds of computers owned by the agency.

The second statement is that this agency is expected to publish, by the end of the year, a plan to reduce the digital dependency on the US. It's not "France to ditch Windows", it should be "French government promises to plan soon for possible ways to decrease digital dependencies, but calendar unknown". Also note that the government (and president) will change next year, so even if the present drive was real, a political u-turn could come soon.

Overall, this statement could be the presage of a major upturn in a few years, but I think it far more probable that the policy change will be minor. There's already a small tendency toward Linux and Free Software in the public sector.

Here is an histogram of the energy mix 1990-2020 that illustrates your answer. https://en.wikipedia.org/wiki/File:Energymix_Germany.svg

The nuclear share (red) is reducing during the 200Os. The wind and solar (light blue and yellow) went over the max nuclear share at the end of the period — it seems there is much more wind than sun in Germany ;-). The fossil fuels (dark colors below red) are still very high.

The media (web or desktop) is irrelevant: a file format must exists for backup and interoperability. I barely use office documents myself, but I work on software that produce and parse many spreadsheets every day.

An open standard is even more very relevant in public administrations where the process follows legal constraints and ISO standards. The Document Foundation's article reacts to an German institutional decision.

I hope that Germany mandating ODF over OOXML will enhance the whole ecosystem.

As a programmer, finding decent ODF libraries is far from certain. Last year I had to output some spreadsheets from a Go program, but I could not find any maintained library for ODS, so I had to output XLSX files. Recently, I was luckier while programming in Rust.

You missed an easier alternative that was in the article: ctrl-u saves and clears the current line, then you can input new commands, then use ctrl-y to yank the saved command.

With zsh, I prefer to use alt-q which does this automatically (store the current line, display a new prompt, then, after the new command is sent, restore the stored line). It can also stack the paused commands, e.g.:

$ cp foo/bar dest/ <alt-q>

$ wcurl -o foo/bar "$URL" <alt-q>

$ mkdir foo <enter> <enter> <enter>

I've contributed a few optimisations to some implementations in these benchmarks, but as I read the code of many other implementations (and some frameworks) I lost most of the trust I had in these benchmarks.

I knew that once a benchmark is famous, people start optimising for it or even gaming it, but I didn't realise how much it made the benchmarks meaningless. Some frameworks were just not production ready, or had shortcuts made just for a benchmark case. Some implementations were supposed to use a framework, but the code was skewed in an unrealistic way. And sometimes the algorithm was different (IIRC, some implementation converted the "multiple sql updates" requirements into a single complex update using CASE).

I would ignore the results for most cases, especially the emerging software, but at least the benchmarks suggested orders of magnitudes in a few cases. I.e. the speed of JSON serialization in different languages, or that PHP Laravel was more or less twice slower than PHP Symfony which could be twice slower than Rails.