The bitmap visualizer has been simplified considerably. The previous version was designed to work with the TIM format which has some peculiarities that are not general enough. The current implementation has the following specifications.
. Whether colors are in a lookup table or part of the image itself they are always 32 bit R8G8B8A8.
. If using a color LUT the image then has indices as its element. Indices can have 16(32000 colors), 8 (256 colors) or 4(16 colors) bits each.
.For the cases 0f 16 and 8 bits, the data should be an array of N*M elements of the given size where N is the number of rows and M is the number of columns of the image.
. For the 4 bit case use an array of N*M/2 bytes so that each column contains two indices.
ToDo: Documentation, sample patterns and unit tests.
The 3-d visualizer can now handle textures from both the command line or the user interface and things should work as expected. A command line entry will be automatically displayed in the user interface, but changes will be applied immediately as you type or use the file picker. If the user interface text is deleted, then the command line texture will be used again. If a texture is invalid for any reason, then the previous one, if any, will be still in use and an error message will be displayed until the problem is cleared. Valid textures are image files that the stb library can open.
<!--
Please provide as much information as possible about what your PR aims
to do.
PRs with no description will most likely be closed until more
information is provided.
If you're planing on changing fundamental behaviour or add big new
features, please open a GitHub Issue first before starting to work on
it.
If it's not something big and you still want to contact us about it,
feel free to do so !
-->
### Problem description
This is a fix for the ImHex bug "ImHex crashes when analysing any PE
file #2221"
### Implementation description
This is a fix for the ImHex bug "ImHex crashes when analysing any PE
file #2221". The fix requires changes to the Pattern Language and the
ImHex UI. Revision would be wise. We want to avoid collateral damage.
It's a big code base and I'm new to it and the compilers/build-systems
used. And the bug is complex and low-level. I suspect this will fix
other random crashes. The problem is caused by two issues:
- The std::sort algorithm conjuring up garbage due to the sorting
criteria not being a strict weak ordering. See
[this](https://github.com/Voultapher/sort-research-rs/blob/main/writeup/sort_safety/text.md)
link.
- We sort shared_ptr<ptrn::Pattern> by pointer value, and the
object is a clone. In essence we're changing the values as we're
sorting.
Fixes#2221
### IMPORTANT
I'm not sure how "plugins/builtin/source/content/data_formatters.cpp"
got into the PR. Been trying for an hour to rectify. I'm not a Git
expert (the last time I used source control seriously SourceSafe was a
thing) please ignore that file. It's a fix for another PR I submitted. I
suspect I stuffed up the branching and merging.
---------
Co-authored-by: Nik <werwolv98@gmail.com>
The recent update that made importing patterns undoable had the side
effect of undoing tab changes as well. When working on a fix for that it
became clear that the undo/redo stacks were being shared by all
providers so that you could undo changes done in a separate file. The
problem was fixed by making the text editors (and the console editors as
well) to be per provider which gives better control on things like
breakpoints and selections and a more robust focusing system that
preserves the cursor positions and removes the need to click on a
pattern to start editing. There are a lot of changes, but they are
mostly all the uses of text editors being now coded differently to
manage the providers. File imports are still undoable, but switching
providers is not.
Further changes suggested by reviewer were implemented namely a mull provider was being used and there was a get function missing which prevented the use of the preferred syntax.
The recent changes to the text editor to fix the longest line length
problems broke the console horizontal scrollbar. The code that displays
the console editor was more complicated that it needed be, and it had
the bad side effect of resetting the cursor which prevented horizontal
scrolling. Adding a function that appends lines to the text editor fixes
all problems and makes the code clearer. To accommodate for strings
containing zeros, the code that inserts text was changed to print a '.'
when zeros are encountered thus keeping the line length the same.
The implementation of subpixel rendering using draw call lists with
callbacks prevents call list merging and their associated clip rects in
tables. As a result clip rects become as narrow as the columns of the
table which can clip previously renderable primitives. The hex editor
has several draw calls that render outside their column so if subpixel
rendering is selected those primitives cease to be displayed.
To fix this issue, and to verify that this was indeed the cause behind
the issue, we simply push an adequately sized clip rect before the draw
call command and pop it right after.
A problem with segment vertical separators not being rendered in the
first tow was also fixed.
If you tried to collect the indices using addressof like stl pattern
collects vertices you get a small square for visualizer and no error
message. The changes here are able to extract the indices if they can be
extracted and give an error message if they can't.
Proof of concept for implementing subpixel processing in ImGui. This is
work in progress, and it is bound to have problems.
What it does:
1) Uses freetype own subpixel processing implementation to build a
32-bit color atlas for the default font only (no icons, no unifont) . 2)
Avoids pixel perfect font conversion when possible. 3) Self contained,
no ImGui source code changes.
4) Results in much improved legibility of fonts rendered on low dpi LCD
screens that use horizontal RGB pixel layouts (no BRG or OLED or CRT if
they even exist anymore)
What it doesn't:
1) Fancy class based interface. The code is barely the minimum needed to
show it can work. 2) Dual source color blending. That needs to be
implemented in shader code, so it needs to change ImGui source code
although minimally. This will result in some characters appearing dimmer
than others. Easily fixed with small fragment and vertex shaders. 3)
subpixel positioning. If characters are very thin they will look
colored, or they can be moved to improve legibility. 4) deal with
detection of fringe cases including rare pixel layouts, non LCD screens,
Mac-OS not handling subpixel rendering and any other deviation from the
standard LCD. 5) tries to be efficient in speed or memory use. Font
Atlases will be 4 times the size they were before, but there are no
noticeable delays in font loading in the examples I have tried.
Any comments and code improvements are welcome.
---------
Co-authored-by: Nik <werwolv98@gmail.com>
