T-strings, Croc, Firefox Units, 802.11[k,r,v], Transmission
Python t-strings, aka templates PEP-750 (@ peps.python.org) - t"Hi! {user!r}, balance: {balance:.2f}" - I especially would love if pyodbc would support this syntax. Because currently it doesn't support dynamic lists (at least in all cases), which is a big problem, or not... I've created a few ugly helper functions for that. Like function which produces '?,?,?' string to be placed inside the statement and it's length is auto defined by the list being processed. lph(stuff) aka list_place_holders generates just ','.join('?' * len(input_list)) today
Croc - Thought about the bad UX with Croc I wrote a few months ago. It was so annoying to explain to the users the flaws over and over again, that I simply created using AI and PowerShell a light script that makes Croc practically usable! croc_share.ps1 opens UI for user to select a file, then configures firewall, assigns local port and shares the file using croc and gives user the code. After the transfer, it removes the created firewall rules. Users tend to forget those, if not automated! The croc_receive.ps1 prompts for the server_name and code, then it fetches the file from the host directly, by doing the DNS lookup the croc is missing (!) and using predefined port number. Now crock works as I would prefer programs to work, and is also 10x+ faster than with the slow default config. - Alternate configuration would be using data center specific IP addresses and per data center installed private gateway, it would also provide speed benefits. But again UX would be bad, users would need to know if the systems are in same or different data centers, etc. Of course that could be also scripted, no problem there. But this solution is simple and works, even if the systems are in same or different data centers, as long as the systems got public IPs (and usually DNS records which are preferred over IPs) which is the norm in these cases.
Firefox Units - So is it helpy or helpful, if I tell that I find it utterly annoying that Firefox configuration has UNIT problem? Some fields use different units than other fields. And if there's a field, why it doesn't allow specifying the unit. Current solution with integers of "some unit" is just enraging. This field seemingly uses seconds? I assume, it was hard to confirm. No I didn't bother to read the source code for it. 'browser.sessionstore.idleDelay' and this field uses milliseconds 'browser.sessionstore.interval' I assume. I find this really bad approach. Using a single unit would be fine, or allowing user to define unit. But just mixing "random unit" with integers, and without specifying what the unit is, is enraging. Yeah, it size is 8234 ... desibits of course!!!!
Still about previous post and heply. It's sometimes baffling to find out that some serious bug has been in production for years, and nobody has reported it, even if it's obvious that is has not ever worked as specified and designed. - How's that possible? - If anyone whom encountered the issue would have reported it ...
WiFi / WLAN Roaming - Finally some standards (when and if supported) that would make WiFi networking practically working: 802.11k (Neighbor Reports), 802.11v, 802.11r (k,v,r). These are by design features which I find meaningful and useful. But mostly what people talk about WiFi roaming is totally pointless disinformation. When devices and access points do support those, then I say it sounds much better and actually working solution. It's good to remember, that the support is not mandatory! Ref: 802.11 category (@ Wikipedia), 802.11k (@ Wikipedia)
Transmission 4 support for infohash (@ Wikipedia) directly (without URL formatting) finally works! I did hear so many people complaining that Transmission doesn't support infohashes. Of course it did, but you had to enter it in right format. Some users just weren't smart enough to remember and correctly add the necessary prefix to the infohash. magnet:?xt=urn:btih:0123456789012345678901234567899012345678Â yet now the client handles the hash itself correctly without needing the prefix. - Talking about user friendliness and generic it works, users are just stupid attitude. - Very similar to the enraging design with many apps, they provide links, but the app doesn't consume the links generated by the app correctly, and this problem is wide spread! - Their engineers aren't even them selves smart enough to understand what their own generated URLs mean. How do they assume that users can do it. Lack of support for magnet:?xt=urn:btmh:12201234567890abcdef1234567890abcdef1234567890abcdef1234567890abcdef0123 v2 with SHA-256 hashes is also slightly duh.
2026-06-11