Implemented basic sorting and search function including replace, very basic but working, also the sorting close to the LibreOffice sorting, multi key sorting (up to 3 but in principle could make more, but is more than 3 really useful?) Together with the Format options this should be enough for a new LEU Version release Leu 0.08.
Fixed a little bug in Leu regarding number format. If no number format was selected the output format was one char too long, resulting in very ugly rounding errors. Already long time before I started to implement a solution to set the number format for a cell. But it was very crude and not ready to release, so I deactivated it for Leu 0.07. But this bug got me back to work. 😉 I started to implement a format settings window, a little bit inspired by the LibreOffice one. Finally its working, not perfect but already usable.
Also fixed some small problems with the color blending on selection for RTG systems (for <15bit screens it will just use a fixed color for the selection) and new now the fixed col/row will show where the selected cell is. much easier to find, especially if it is out of scroll area.
Updated Online Amiga FPC Compiler again with the latest CRT unit and also tried a bunch of console based games using CRT units, all work more or less (some need some adjustments in the code) some of them are even single file pascal source files, so can be used in the online compiler.
I also tried DeadCold (and with it the new CRT implementation) on my Amiga 1200 (68030/50Mhz, no GFX card) but the colors did not show up. everything stayed black. But the ConUnit is there and the Pen allocation worked, so what went wrong. I looked again through the documentation and found a little hint. For speed reasons, CON: tries to use as little as possible bitplanes for the output to speed up any scrolls. That was a new information for me. That also means this colorful prompts slowed down the console speed. But when I just write the pens to the ConUnit structure, how he can know that he needs more bitplanes seems it does not detect that automatically. But there is a field in ConUnit named cu_mask, it is not described in the includes but it seems to be exactly controlling that, if I change that, it started to show more colors. And the result is that now the DeadCold also works on a native Amiga 1200, not actually very fast, but it’s ok.
Working on the CRT colors. The problem is that the setting of Front and Background color in Amiga Console is very limited. You can only set the colors 0-7 of the Workbench palette as colors using the ANSI api. which are usually not very useful. Even on RTG that is very limiting. I found that you can set the color directly via the ConUnit, which is available for the standard console window (but not with ViNCed for example). But on AROS it works nicely and on plain shell in OS3.1 also, so at least there is a way to use it.
On MorphOS the ConUnit is never available, but thanks to some hints in the morph.zone forum I learned that one can use one of the 256 ANSI colors directly in the console. so I searched for closest possible colors of the standard VGA colors, sadly that only works on MorphOS.
Next days I will also try OS4… I hope that results not in an other case.
Someone might ask how the crt is going, and how useful this is already. I want to answer it with a little screenshot:
That is a nice little rouge-like game called DeadCold, written in FreePascal for the console. As you can see it runs and the picture looks right. (it’s not perfectly right, but somehow usable). And because a little video tells much more about how this game runs here is also a little video about it:
There are some drawing problems, especially on the messages and sometimes some of the map disappear but not bad I would say. There are also some messages about removing critter failed… but that seems some error in the game code.
The colors is badly hacked for now, and only works for AROS, I’m not sure if I can upload that this status to the official repository, I will work on it a bit more.