#Threads #fediverse O uso de hashtags no Threads é diferente do Mastodon ela se transforma em um tópico, porem quando a postagem chega no Fediverso ela aparece como hashtag mas você não consegue clicar nela.
This made me LOL: While doing some research on the MAGA wingnut Laura Loomer (for no particular reason) I found a document indexed by Constella Intelligence that shows that in 2022 someone sent Loomer a delivery from shitexpress[.]com, a real business that will let you pick from a bunch of different kinds of animal feces to ship to someone's doorstep.
The accompanying message was curiously signed, "Sorry about your pay pal 🙁 " The sender in this case appears to have chosen to send horse shit.
Having been the recipient of one of these highly odiferous packages, you somehow immediately suspect what is inside based on the stink, heft, and squishiness of the package, even before you notice the sender is shitexpress.
France : Le Ministre de l’Intérieur veut intensifier la répression contre les Free Party secoursrouge.org/france-le-min… "Après la free party qui a réuni jusqu’à 2 500 personnes ce week-end dans les Corbières, les policiers ont saisi du matériel et effectué plus de 1000 verbalisations. Quelques jours après, le ministre de l’Intérieur Bruno Retailleau s’est dit favorable à réprimer plus durement ces événements en
New from Special Correspondent Mary Tuma: House Bill 7 attacks one of the last remaining options for reproductive care in #Texas and declares legal war on blue states. texasobserver.org/texas-legisl…
Meta says says that the 10,000 characters give people more room to express themselves, but also allows them to promote their work and drive others to "wherever it lives," even if not on Threads itself.
Eux : "Bernard Arnault est un capitaine d'industrie valeureux qui crée de l'emploi en FR"
La réalité : des ouvriers chinois payés 4 euros de l'heure qui font 90 heures par semaine pour fabriquer des sacs à 4.000 €, dont LVMH tire des profits gigantesques et Arnault sa richesse
Attaché : 1 image
Eux : "Bernard Arnault est un capitaine d'industrie valeureux qui crée de l'emploi en FR"
La réalité : des ouvriers chinois payés 4 euros de l'heure qui font 90 heures par semaine pour fabriquer des sacs à 4.
If you are looking for a gift for a new home or for a gardener or cook. I have a selection of fruit and vegetable drawings available, including carrots! You will find them in my Folksy shop... folksy.com/items/8530252
The image depicts a serene coastal scene with a prominent rainbow arching across the sky. The rainbow, with its distinct bands of red, orange, yellow, green, blue, indigo, and violet, is set against a backdrop of partly cloudy skies. The clouds are a mix of dark and light gray, with some areas illuminated by sunlight, creating a dramatic contrast. Below the rainbow, the ocean stretches out, its surface reflecting the colors of the sky. In the distance, a range of mountains is visible, adding depth to the landscape. The overall atmosphere is calm and picturesque, with the rainbow serving as a focal point in the composition. ― The image depicts a stunning natural scene featuring a vibrant rainbow arching over a body of water, likely an ocean or sea. The rainbow is prominently displayed in the center of the image, with its colors transitioning from red on the outer edge to violet on the inner edge. The sky is partly cloudy, with large, fluffy clouds scattered across the backdrop, allowing glimpses of blue sky in some areas. The ocean appears dark and calm, with gentle waves visible in the foreground, reflecting the colors of the rainbow. In the distance, a landmass with rolling hills and a small settlement can be seen, adding depth to the scene. The overall composition captures the beauty of nature, with the rainbow as the focal point, set against a dramatic sky and tranquil water.
Provided by @altbot, generated privately and locally using Ovis2-8B
Das Bild zeigt eine ruhige maritime Szene am Abend. Im Vordergrund sind mehrere Boote am Steg vertäut, darunter ein größeres Fischerboot, das mit Lichterketten geschmückt ist. Die Lichterketten verleihen dem Boot eine festliche Atmosphäre. Einige Boote sind weiß, während eines in einem dunkleren Farbton gehalten ist. Ein Laternenpfahl mit einem warmen Licht leuchtet auf der linken Seite. Der Himmel ist in tiefen Blautönen gehalten, mit einigen Wolken, die den Übergang zum Abend hervorheben. Die Wasserfläche spiegelt die Lichter der Boote und der Laternen wider, was zur friedlichen Stimmung beiträgt.
Bereitgestellt von @altbot, privat und lokal generiert mit Ovis2-8B
EMERGENCY drive 🚨 but the ONLY support today for the giving circle—not one but SIX families AND a displacement camp in the heart of the occupation's advancement into Gaza City—came from a handful of folks who clicked "weekly" on £5–20.
We're running this drive so they can get supplies & shelter urgently in this catastrophe. 3296/4100 by Saturday and please join in by clicking "weekly." chuffed.org/project/hope-givin…
We are a team of volunteers in the US and UK, led by two sisters passionate about helping our friends on the ground keep their families and neighbors alive with food, water, & shelter.
In 1988, Domus explored Armani’s approach to designing garments: a true design process, from sketch to prototype, culminating in a model that was repeatable yet at the same time inimitable.
🎞️ J’ai vu le film de Christian Petzold Miroirs n°3, un film lumineux sur le deuil et la dépression. Son itw vers la 20ème minute. radiofrance.fr/franceculture/p… @franceculture
Aujourd'hui, on se raconte des histoires dans un bar-tabac avec Vincent Maël Cardona, et on prend la place d’une morte dans la campagne allemande avec Christian Petzold ; avec aussi Charlotte Garson.
Das Bild zeigt eine Schüssel mit einem hellbraunen, gut aufgegangenen Teig. Der Teig hat eine glatte, leicht fettige Oberfläche und ist in einer weißen Schüssel auf einem dunklen Hintergrund platziert. Die Oberfläche des Teigs ist nicht perfekt glatt, sondern weist kleine Unebenheiten und Luftblasen auf, was auf die Fermentation hinweist. Die Schüssel scheint aus Porzellan oder Keramik zu bestehen und hat eine glatte, glänzende Oberfläche. Um die Schüssel herum sind kleine Teigreste zu sehen, was darauf hindeutet, dass der Teig möglicherweise vorher in der Schüssel gemischt wurde. ― Das Bild zeigt ein großes, runder Brot mit einer dunkelbraunen, knusprigen Kruste. Die Kruste ist mit feinem Mehl bestäubt und weist eine unregelmäßige, gebräunte Oberfläche auf, die durch natürliche Risse und Kratzer geprägt ist. Diese Risse verleihen dem Brot ein handwerkliches Aussehen. Das Brot liegt auf einem hellen Untergrund, der einen Kontrast zur dunklen Kruste bildet. Die Oberfläche des Brotes ist nicht glatt, sondern zeigt eine Mischung aus glatten und rauen Bereichen, was auf eine traditionelle Backmethode hinweist. Die Farbe und Textur des Brotes deuten auf eine herzhafte, hausgemachte Backware hin.
Bereitgestellt von @altbot, privat und lokal generiert mit Ovis2-8B
Iedereen boos op PVV'er Markuszower, die een debat aanvraagt en mensen beschuldigt en vervolgens niet komt opdagen. #PVV youtube.com/watch?v=VU1NUj0Xpr…
Profitez des vidéos et de la musique que vous aimez, mettez en ligne des contenus originaux, et partagez-les avec vos amis, vos proches et le monde entier.
Joy of harvest personified: It is not just the colors and emotion of my piece, “Girl and Flame Tokay,” that has always been compelling. It also tells a story: Of a grape that for over 100 years was the most widely planted crop in Lodi. A Vitis vinifera tailor-made for the climate and deep, rich sandy loam soils. It is still cherished today, mostly as an heirloom by longtime Lodi familes.
Taking a stab at making a privacy-preserving analytics thing for my blog.
I'm hashing the user's IP and User-Agent, so they're never stored directly. Not 100% accurate, but good enough for me to work out whether a page-view is unique or not.
Country is very rough - using GeoIP.
Referer is either from HTTP or if the linking site uses UTM.
Seems *broadly* accurate compared to other plugins.
I'm not sure of the performance if this hits a million rows. Might need further DB tables.
Plausible Analytics is an open source project dedicated to making web analytics more privacy-friendly. Our mission is to reduce corporate surveillance by providing an alternative web analytics tool which doesn’t come from the AdTech world.
if performance is your concern, DuckDB is incredibly fast. I had Claude write me an AWS Cloudfront to Parquet 1 converter with IPinfo lite MMDB lookups, which you can then import into a DuckDB. It takes 41 milliseconds to find the top 10 countries over 8.8M rows, about 5 years' worth of logs:
``` zulfiqar ~/lognami>duckdb blog.duckdb DuckDB v1.3.2 (Ossivalis) 15004e53 Enter ".help" for usage hints. D .timer on D select count(*) from web; ┌────────────────┐ │ count_star() │ │ int64 │ ├────────────────┤ │ 8820061 │ │ (8.82 million) │ └────────────────┘ Run Time (s): real 0.008 user 0.006334 sys 0.000867 D select country, count(*) from web group by 1 order by 2 desc limit 10; ┌─────────┬──────────────┐ │ country │ count_star() │ │ varchar │ int64 │ ├─────────┼──────────────┤ │ US │ 3942963 │ │ FR │ 727561 │ │ DE │ 570468 │ │ CN │ 456912 │ │ NL │ 418870 │ │ SG │ 408658 │ │ GB │ 279144 │ │ CA │ 218756 │ │ IN │ 175326 │ │ RU │ 157289 │ ├─────────┴──────────────┤ │ 10 rows 2 columns │ └────────────────────────┘ Run Time (s): real 0.041 user 0.181591 sys 0.648902 ```
@cisene I'm going to be cranky with the Unicode folk if adding an offset to the characters of a two letter country code doesn't yield the characters to make up the flag 😂
Bring back the old ways, convert numbers to characters by adding 48 to them.
SET NAMES utf8mb4;<br><br>-- helper: safe code point to utf8mb4 string using JSON unescape<br>DELIMITER //<br><br>DROP FUNCTION IF EXISTS unichar//<br>CREATE FUNCTION unichar(cp INT)<br>RETURNS VARCHAR(4) CHARSET utf8mb4<br>DETERMINISTIC<br>BEGIN<br> DECLARE v INT;<br> DECLARE hs INT;<br> DECLARE ls INT;<br> IF cp < 0x10000 THEN<br> RETURN JSON_UNQUOTE(CONCAT('"\\u', LPAD(HEX(cp), 4, '0'), '"'));<br> END IF;<br> SET v = cp - 0x10000;<br> SET hs = 0xD800 + FLOOR(v / 0x400);<br> SET ls = 0xDC00 + (v % 0x400);<br> RETURN JSON_UNQUOTE(<br> CONCAT('"\\u', LPAD(HEX(hs), 4, '0'),<br> '\\u', LPAD(HEX(ls), 4, '0'), '"'));<br>END//<br>DELIMITER ;<br><br>-- idempotent create for the flag function<br>DELIMITER //<br>DROP FUNCTION IF EXISTS flag_emoji//<br>CREATE FUNCTION flag_emoji(cc VARCHAR(2))<br>RETURNS VARCHAR(8) CHARSET utf8mb4<br>DETERMINISTIC<br>BEGIN<br> DECLARE base INT DEFAULT 127462; -- U+1F1E6 REGIONAL INDICATOR SYMBOL LETTER A<br> IF cc IS NULL OR cc NOT REGEXP '^[A-Za-z]{2}$' THEN<br> RETURN NULL;<br> END IF;<br><br> RETURN CONCAT(<br> unichar(base + ASCII(UPPER(SUBSTRING(cc, 1, 1))) - ASCII('A')),<br> unichar(base + ASCII(UPPER(SUBSTRING(cc, 2, 1))) - ASCII('A'))<br> );<br>END//<br>DELIMITER ;<br><br>-- examples<br>SELECT flag_emoji('GB') AS gb, flag_emoji('us') AS us, flag_emoji('de') AS de;<br>
mysql> SELECT flag_emoji('GB') AS gb, flag_emoji('us') AS us, flag_emoji('de') AS de;<br>+----------+----------+----------+<br>| gb | us | de |<br>+----------+----------+----------+<br>| 🇬🇧 | 🇺🇸 | 🇩🇪 |<br>+----------+----------+----------+<br>1 row in set (0.00 sec)<br>
I apologize for the JSON abuse, but that was easiest.
U+1F1E6 is the unicode hex value of the character Regional Indicator Symbol Letter A. Char U+1F1E6, Encodings, HTML Entitys:🇦,🇦, UTF-8 (hex), UTF-16 (hex), UTF-32 (hex)
I want to draw a graph of visitors. I don't want to call a 3rd party graphing service. I don't want to use a big library from someone else. I am not going to write my own graphics generator.
So I think I'm going to build a Unicode sparkline generator for PHP.
Let me know how you get on. My page is getting a bit bloated, and needs a revamp. What I did not want to to was generate and include loads of external images for each (daily) rebuild. A possible solution for me is in-line SVG which my HTML minifier could do a good job on in passing...
Might have gone a *little* overboard! I rather like this though. Self-hosted analytics showing a graph of the last 7 days. Today's top referers, pages, and countries.
Super simple and suitable for my needs.
Next up - import historic data and work out if I want to keep countries & referers in it.
Interestingly, I've started seeing hits come in from `noai.duckduckgo.com`
Small numbers for now, but wonder if they'll grow? I'll probably consolidate it to the 2nd level for my stats. Not sure there's much useful I can do with the information.
information and granularity overload can make for a challenging decision making environment. At my time in the Civil Service our web teams wanted *so many* metrics, but ultimately only listened to the user researchers. Can’t put detailed 30 minute chats into a PowerBI dashboard though.
I suppose you could write anti-AI content to juice those numbers.
I think I've cracked the problem - MySQL makes my brain hurt.
Table 1: Logs *everything*. Date, post ID, user, etc.
Table 2: Is date, view count, user count. When (1) is updated, an AFTER INSERT trigger increments the view count for that day. It also counts the number of unique users for that day and updates (2).
Table 3 is the same as (2) but on a per-post basis. It also updates after insert.
That way, I don't have to use logic in my PHP app - all it does is fire an insert.
if you are updating a "user count" in a trigger, then every single insert is going to be in contention of a lock on that row, and you will get massive concurrency problems that bottleneck your throughput.
My advice if there is even the slightest chance this will scale beyond one blog would be to do the write without any triggers, and the calculate the total per day when you need it, or every 5 minutes, or something.
e.g. in an OrderItemLog table with 680767 rows, it takes 0ms - that's zero - to count how many are on a given day. And 0ms to count how many distinct orders there are.
be careful with 'UPDATE count = count + 1" patterns. They're unlikely to accurately count as you expect due to concurrency issues. Especially in MVCC systems, it will depend alot on transaction isolation levels.
For this kind of thing, just logging the visits as records and aggregating on analysis / periodically is often the better approach. Removes a lot of the concurrent write headache and give you more flexibility.
Arse. My host doesn't support EVENTs on my cheap plan. Fair enough. Looks like I'll have to use cron. Wonder whether it's better to have a stored routine, or just keep the logic in the PHP?
My heart says use the stored routine - but it's probably more flexible and fixable to have the SQL in the PHP.
Friends don't let friends use Stored Procedures. That way darkness lies... (as in weird stuff happens and you've forgotten about the SP and version control is a PITA)
I can now click on a single post and see how many times it has been viewed, how many unique users visited it, which countries they came from, and what referers sent traffic to that page.
So far, it is *broadly* accurate compared to my 3rd party plugin. It over-counts users and under-counts page views. But that's fine.
Next steps: draw some pretty graphs. Ideally with weekly / monthly / yearly groupings.
@jackeric oh, I see. I've emailed both the manufacturer and OEM to see if I can get the SDK. No luck yet. You can upload your own photo to one of the faces. Not exactly LCARS, but a start 🙂
I think initially I wanted to be able to store other types of visual codes than QR - even if 99% of readers expecting Code 128 will take a QR version in stride
You could truncate the domain to a specific number of characters too, with the full domain, as you currently have it, available when hovering over as title text.
I've used charts.css in a recent backoffice project, where you create a table using html templating language and then it's styled like a bar chart. Very handy!
You're very welcome to nick my SVG sparkline generator (It's in js but should convert to PHP fairly easily if you want to do it server-side rather than client-side) Post: kevinmarks.com/svgsparklines.h… code: view-source:kevinmarks.com/sparkline.svg
A million rows shouldn't cause any problems if you have sensible indexes, but you might want to rewrite:
DATE(time) = '2025-09-05'
as:
time >= '2025-09-05' AND time < '2025-09-06'
because DATE(time) *might* cause any indexes on time to be ignored, since you're transforming the column to something other than what was indexed (I had a client a few years ago with a really slow report because of this)
I'm going to be a bit of an annoying nerd, but this is not actually privacy preserving. Firstly, you're building user identifiers, which allows you to track users' behaviour. Secondly, hashing the user IP and user-agent is very easily reversible, so you shouldn't be even publishing that screenshot 😀
we may be using different definitions of words, but you know which pages I've visited at what times and where I came from. is that not tracking? if I went to a book store, and they recorded which coffee shop I came in from, and what books I've looked at in which order, I wouldn't call that privacy preserving
Plausible Analytics is an open source project dedicated to making web analytics more privacy-friendly. Our mission is to reduce corporate surveillance by providing an alternative web analytics tool which doesn’t come from the AdTech world.
🌍🐬 🦋Earthling84 🐞🕊️🌞
in reply to [G.R.K.]©️®️™️👽🏳️🌈🏳️⚧️🖖 • • •il semblerait que les amendes étaient de 750€, donc 750x1000= 750000€ de ramassés !
Sur le dos des danseurs en un weekend !
Ha elle est belle la liberté d'être soi....
Ils pouvaient rien détruire le réchauffement sen était déjà occupé !