Location, Updating, Testing, VLC, Cables, SimpleX
GPS and Navigation problems with modern browsers. It was unfortunately quite painful, several hours of work and unreliable operation. gpsd, gpspipew, chromium - How brittle can GPS things be? Nightmarishly bad process overall. For some reason the SHIM module with gpspipew doesn't seem to work. gpspipew itself works fine. The user experience with browser is just bad. I got several thoughts how the process could be improved. Now when it's at least somewhat working. Fun times. I'm sure that most of people won't get it ever to work, because it's so tricky and requires so much tuning. But after all, thank you so much for this packet and this time I mean it and it's not sarcasm. - gpsd-chrome-polyfill - https://github.com/micolous/gpsd-chrome-polyfill/tree/master - (@ github.com). After getting sick'n'tired with this stuff, I just made program which generates OSM links with coordinates, which you can click to get "gps fix". Yes, I'm very well aware how ridiculous that is, but that's just who bad things are. Browsers and GPS integration isn't just working combination, even if there's nothing special there.
Updating software. I've done this several times. Someone complains about issue, I'll make update. And then they'll claim that they've updated the software and then whine that it's not working. My current solution to that is often that I'll make version, which isn't obviously intended to work at all. Why? Because I've so often noticed that people do not update and still keep lying about the problem. Purpose of the broken update is to make it obvious that they've even updated the software. And then I know I can try to deliver fixed version. Again, so frustrating and pointless. But that's the most efficient way of doing it. New version, first line of code just says: raise EnvironmentError("Congratulations, you've actually managed to update the software!"). How cute is that. - And I've done this several times, since writing this post and between publishing.
Testing procedures. I just can't stand it, when people test things in totally pointless fashion. They can't explain what they're testing and why, and why it would be meaningful, but lots of resources are used to test things that if it would work. It's even painful to watch them doing it. ANd it's even more painful, when they are eager and try to ask for help with that testing. Without understanding that there's absolutely no point to test those things, because it just won't work. And I already know it darn well. - Got so deeply annoyed again because people don't get difference between bytes and strings. If some things are known to be broken, don't even bother to try, it'll be just massive waste of time and effort.
VLC still doesn't handle colors correctly. Some videos look extremely bad when using VLC but luckily there are other players which have correct color handling. It's kind of extra annoying that they've been fixing this issue for years and no progress has been achieved. Yet as example Parole player does another legendary things, like places subtitles outside the image area (which is great), but then it crops the image area based on aspect ratio information, practically hiding most of the subtitles. Sometimes I really wonder who writes this software, because so many things do not make any sense at all. Also when placing subtitles below image, it sometimes would be beneficial to move the image upwards, because there's also empty space above the image. But yeah, that's clearly some dark science which goes way above average developers understanding. - Update: 24.04 LTS Ubuntu updated fixed this
Watched long documentary about submarine cables, related security threats, global politics, finance and corruption. - Power grab is just about everywhere, no news I guess?
SimpleX Chat v5.4 released with remarkable improvements, as well as ability to link devices, which is nice. If you need to post something longer and you don't want to use mobile devices keyboard. Yet I often use secondary channel to post test and or use physical keyboard with the mobile device, which both work fine. Yet the secondary channel option naturally got it's security drawbacks, but usually that doesn't really matter.
One older Python integration projects used lots of loops and lists to check for cross references. I replaced the lists with sets, and now the whole program runs about 40% faster. Nice!
Something different? Boing & Saab T-7 Red Hawk - https://en.wikipedia.org/wiki/Boeing%E2%80%93Saab_T-7_Red_Hawk - (@ Wikipedia) possibly F-7 in future.
2024-09-29