Usually when I translate texts from one language to an other I use DeepL to do it, it makes better results than Google translate in the most cases (at least if german is the target language). It also has an API to use and the first 500.000 chars are free, so I created a little program for Amiga to use that API.
It’s stupidly named AmiTranslate I know very unimaginative but hey seems the name is not used already, so why not. Of course the program needs an Internet connection and SSL (OpenSSL3 on MorphOS; AmiSSL on all other systems) because it talks directly with the DeepL server (via https) and it needs the texteditor.mcc.
Of course for some of the languages there is the problem of showing the chars (e.g. russian) for that I included a PDF writing function and a plain UTF-8 Text output then you can use other programs to read the actual text. of course nicer would be to make a real TTF output maybe later.
I also removed some languages which does not work at all (even the PDF output, like japanese, korean and chinese)
If you have an own DeepL key you can add the ToolType „DEEPLKEY=<key>“ to the icon, then even my 500.000 chars are used you can still use your own, it’s free so why not.
Of course it’s perfectly possible to use any text 2 image webpages like getimg.ai with AmiFox on Amiga. But of course a better way is to make a specialized app for it using hugging face with stable diffusion.
It also contains Magic prompt which can produce more sophisticated prompts or even a complete random prompt. Gives really nice pictures sometimes.
Introducing AREXX commands for AmiFox, so you can send a URL to AmiFox to view. If the Browser is idle it will directly start to load the new URL. If still loading another URL it will remember the URL and start it after the current one is finished loading, maybe later also stop the current task, atm its easier that way.
Wit this new AREXX command of course one can use an AREXX script but even better you can use programs like openurl to directly send urls to your favourite Web-browser… or AmiFox 😀
A new version of AmiFox is ready for release, I fixed some more bugs the last days. The 0.3 Version is a big step forward for usability of AmiFox. Most of these features I already introduced in the blog entries here so just a short recap.
Copy Text from selected webpage (AmiFox-Server 0.2+)
Copy Clipboard from Server (AmiFox-Server 0.2+)
IFF Format (AmiFox-Server 0.2+)
send longer texts to page (needs texteditor.mcc)
Mouse wheel support
Shift cursor up/down send as Page up/Down (AmiFox-Server 0.2+)
Error message when Datatype could not load image
send Youtube URLs to AmiTube
send Google Maps and OpenStreetmap positions to Mapparium
AROS Version atm i386 and 64bit ABIv11
Read URL at Mouse position
direct download of catched URL at mouse position
save Webpage Image
Copy/Paste URL
Downloads as always on the AmiFox Page or with the Update check inside AmiFox.
If you using an own AmiFox Server please note that you must update to the latest Version to get all the Features, The Current Version should have the Version Number „AmiFox-Server 0.3“ visible in the AmiFox about (after at least one page was loaded) or in the footer of the AmiFoxServer webpage, next to the WRP Version number.
A new release of AmiTube is ready. The biggest enhancement is the Search by ID window for easier handling of YouTube URLs.
A second improvement is the self made AREXX Port for AROS which make it possible to send URLs from AmiFox to AmiTube when you are on a Youtube Video page. (all other platforms had it before already)
Bugfix for Get Original Window when no entry is selected
Search by ID Window added for easier URL handling
Support for Youtube shorts (movies in portait format)
I worked a bit on AmiTube, I guess soon we can make a new release. For the next release I improved the handling of shorts and lists. Added a special window where one can paste a youtube URL into and it extracts the IDs from it.
A big problem with shorts is that they are saved in portrait mode means the height is much bigger than the width. until now I always only scaled by the width. I worked nicely for the small sizes, OCS and AGA but failed spectacular for AGA+. So now I read out the sizes from the move and scale differently when there is a portrait video. On that I notice that the agablaster has a curious restriction, the CDXL movie must be at least 128 pixels wide. is that a general restriction in CDXL or just for agablaster? I’m not sure.
Also the VCD-MPEG1 showed some strange effects because it tried to scale it, resulting in fat faces, no I added a padding.
There was a miracle happened… I improved the LCL a bit after a long long time, investigated some unusual behavior at MorphOS with LazSokoban, especially because it was slower as expected. I found the culprit and recompiled the MorphOS binary, so have fun testing it.
Also, I did compile for OS4 but I did not not try it, seems I should have because the EXE does not work, at all, also no error message or so, I’m not sure what goes wrong there. It seems it happens when I try to compile a program from Lazarus for OS4 that the program does not work, but when I compile it in the command line it seems work. Interestingly the command line one is much bigger (for LazSokoban it’s double size) so that should be the reason, but I’m not sure whats the culprit. Needs investigation.
Someone asked if EdiSyn could be compiled for AmigaOS4 and yes I tried to compile it but the Executable crashes. I also tried to compile EdiSyn to compile for AmigaOS3 68k and on my new PiStorm Emu68 it works, but I would say it’s still a bit to slow to really use it.. the LCL is too thick of a layer for 68k
A long time before when I was working on the LCL for AROS I tested out some 3rd Party LCL application code to compile for AROS with the newly created LCL. It worked quite nicely. On of them was LazSokoban a Sokoban clone written by a russian guy which worked nicely on AROS.
Some Texts in the GUI where in russian and I translated them to english for the AROS release, some I had forgotten or mistyped a bit. Now I got a request to fix that problem, first there was the big problem to find such and old source code, but luckily it’s included in the archive I published (it’s GPL2, so I had to 😉 ) but I also found in on my Harddisk in the backup.
The changes where not to difficult to make, but the curious thing was if it would still work with the latest LCL. One thing I remember that I need to add the AThreads unit (like Linux CThreads) to make any LCL application work, back in the days that was not needed. And it still works, I checked out the original source and I noticed that they added some kind of skin system (different graphics for the tiles) and also supplied some which looks much nicer than the original one. So I added that as well the the AROS system, it will be available on the next Version of AROS One or here on my page of course.
The other thing that changed since then, we can now compile for MorphOS and OS4 as well (and for Amiga68k, but it will be much to slow for that, I tried) and it also seems to work, not everything, the Buttons images are missing but as far as the game goes not too bad. So I created a Version for all 3 Platforms, AROS, MorphOS and AmigaOS4
„Alex R“ found a AmiTube problem on Amiga OS4 and reported it to me. (some days before already) I tried to implement that AmiTube remembers the splitter position. Sadly I made a big error there resulting in a very small barely recognizable List on the left side of the window. I only tested it on m68k-Amiga where it did not happen, but after a bit of trying I found the culprit and fixed it.
Originally I planed to release that fix with the next version, but now I got some more bug reports about the exact same thing (hmm, nobody of the beta testers reported it). It seems that it is not only an OS4-problem but also happens on at least MorphOS and some 68 Amigas. Therefore I will release this new version now, with bugfixes only, no functional changes.
Sorry for the trouble, if you are affected by this bug.
Thanks to „Alex R“ for the report and the members of the AmiTube Discord channel and of course Amix for the constant support and feature ideas (even if I can’t implement them all)
