After moving a window, the cursor selects buttons slightly higher then were the cursor is drawn. I'm talking about moving an ImTui window, not the whole browser window. Having the issue on Firefox 71.0 (64-bit).

Currently, Imtui seems to support Imgui v1.81 whose demo has a bunch of table related stuff, which Imtui seems to be missing.

ImGui Demo v1.81 (has lots of table related demo items):

ImTui (master) (ncurses_demo) (doesn't have majority of table related items):

I'm unable to click on anything in ncurses example. It just selects text (even title says that, it changes from <path_to_exe> to Select <path_to_exe>). Right-click or pressing some key resets select mode, but there is no way for me to make example application work unfortunately.

I'm compiling with MSYS2 64-bit with MinGW-w64 64-bit compiler on Windows 10. I had issue with compilation as mentioned in issue #1, maybe it is relevant to this problem in some way.

It might be more like an issue of tmux. Both hnterm and imtui-example-ncurses0 worked well in my terminal, but when I run them using tmux, I didn't see anything.

That happened on macOS Catalina with any kind of terminal. I used default .tmux.conf.

If we could say farewell to the 70's. Why we are still fixated to the idea that everything is a tty and we need to use ANSI escape codes to hack around the fact the 70's is over and you don't have a line-printer in your living room? Why nobody (literally, nobody) attempts to make any real changes, like you know, to create hi-res modern command-line (ow, you can't fallback then to runlevel1 to run our superduper (TM) program! You would actively exclude users!!!! NONONO). This is all bullsit, to ride on the existing terminal, and you know what, the silly thing is even the terminal emulators doing things differently, so at the end of the day libs like this and the retarded ncurses will be full of hack and cases and still it will provide some features which won't be supported by all term emulators. But noooooo, we can't do anything new, we have to sit in this swamp created by made-up scientists in the 70's, not because they were evil or something, but because they were employed by telecommunication. But no, linux, unis, bsd, mac and everybody thinks nooo, U CANT TOUCH THIS!! You know what? I would like to beat everybody up who is responsible for the sorry state of the modern personal computing, and for the available operating systems. There was nice ideas, but NOOOO, we had to settle down with the worst possible option. unix, were we bolt and glue layers above layers and above layer.

Fck you for continuing this. But really. Fuck you! Your are ALSO the reason why we can't have nice things.

Now go and close this ticket without spending any minute to think about why the "Beat up a programmer, he will know why he got it" movement exist and why they beat up programmers.

So again: fuck you.

Hello,

I am currently trying to include ImTui to my current console project. I followed instructions and it seems there is some linking problems (probably trivial, I am not too familiar with c/c++ world). I am using CMake and linked all ImTui .a files, I also have ncurses linked to my project. Result when I try to build and run my project:

/usr/bin/ld: /home/shardine/projects/neonlit/client/../lib/libimtui-ncurses.a(imtui-impl-ncurses.cpp.o): in function `ImTui_ImplNcurses_Init(bool, float, float)':
imtui-impl-ncurses.cpp:(.text+0xd2): undefined reference to `initscr'
/usr/bin/ld: imtui-impl-ncurses.cpp:(.text+0xd7): undefined reference to `use_default_colors'
/usr/bin/ld: imtui-impl-ncurses.cpp:(.text+0xdc): undefined reference to `start_color'
/usr/bin/ld: imtui-impl-ncurses.cpp:(.text+0xe1): undefined reference to `cbreak'
/usr/bin/ld: imtui-impl-ncurses.cpp:(.text+0xe6): undefined reference to `noecho'
/usr/bin/ld: imtui-impl-ncurses.cpp:(.text+0xed): undefined reference to `curs_set'
/usr/bin/ld: imtui-impl-ncurses.cpp:(.text+0xf4): undefined reference to `stdscr'
/usr/bin/ld: imtui-impl-ncurses.cpp:(.text+0xfe): undefined reference to `nodelay'
/usr/bin/ld: imtui-impl-ncurses.cpp:(.text+0x105): undefined reference to `stdscr'
/usr/bin/ld: imtui-impl-ncurses.cpp:(.text+0x10c): undefined reference to `wtimeout'
/usr/bin/ld: imtui-impl-ncurses.cpp:(.text+0x116): undefined reference to `set_escdelay'
/usr/bin/ld: imtui-impl-ncurses.cpp:(.text+0x11d): undefined reference to `stdscr'
/usr/bin/ld: imtui-impl-ncurses.cpp:(.text+0x127): undefined reference to `keypad'
/usr/bin/ld: imtui-impl-ncurses.cpp:(.text+0x25b): undefined reference to `stdscr'
/usr/bin/ld: imtui-impl-ncurses.cpp:(.text+0x313): undefined reference to `mouseinterval'
/usr/bin/ld: imtui-impl-ncurses.cpp:(.text+0x31f): undefined reference to `mousemask'

so all the link problems seems to be with ncurses. Do you know what I am doing wrong here?

Best regards.

[MSVC 15 2017 Win32], Windows 10 Command Prompt, using PDCurses I had to fiddle a bit to make it work ( tweaking small things in the CMakeLists.txt to link PDCurses properly ). Then, even in release build, the refresh is very slow in the imtui-example-ncurses0 demo executable. I experience approximately 10fps, even the Web Assembly demo is faster and more responsive. It looks like it is spending a lot of time in ImTui_ImplNcurses_NewFrame > wgetch > sleep. Disabling the std::this_thread::sleep_for did not help either. I can provide more information if needed.

The CMakeLists file sets C++17 as the necessary C++ standard to compile imtui (https://github.com/ggerganov/imtui/blob/418552cbd35a2b468bcac5026903a0b7d3997815/CMakeLists.txt#L85) but the project compiles and runs correctly on C++11 (tested with gcc 4.8.5 on rhel7).

Is there a real need for that?

Repro:

• Open https://imtui.ggerganov.com/
• Open subsection "Popups & Modal windows"
• Open subsection "Popups"
• Click "File"
• Move cursor down from "(dummy menu)" past "Save As.." to "Options"
• Move cursor one entry down

Expected:

"Options" menu closes

Actual:

"Options" menu stays open

Unrelated, it would appear as though there is a small rendering issue with the menu itself (screenshot attached)

First of all apologies for this newbie question, but I am struggling with too many things at the same time: imtui, c++, cmakefiles, static vs dynamic libraries, etc. so I am not sure where the error may be, or whether my approach may not be the better one.

Let me try to break out my problem in parts to make it easier for you to understand and advice

First of all, what I actually want to do I want to create a dummy hello-world imtui example from scratch, and build it in my own ~/projects/imtui-hello folder, with as minimal cmake/dependencies as possible. This is, ideally I'd like to have a hello-imtui.cpp file so I can do something like g++ -Wall [-I SOME_INCLUDE -DSOME_OPTION -SOMETHING_ELSE] ./imtui.cpp.

Can I do a simple build like that or do I actually need to use a CMakeFileLists.txt? Do I need to have the imgui forked dependency cloned on my hello-world folder (i.e. as in imtui/third-party/imgui? What if I install (make install) the (forked) imgui on its own to my system so I do not need to have a cloned imgui for each dummy folder I want to experiment with?

Now, I know there is an examples/ncurses0 but I want to do something much more minimal (dummy box with "hello world" text) than this. Additionally, I am also totally getting lost with the amount of CMake stuff that I do not know how to "re-use" it for this bare minimum experimentation folder

What I have done I've properly build imtui, and ran the examples that are generated on bin folder. I have also installed imtui by issuing a make install. This is, imtui is working and, in theory, properly installed.

I am on Fedora 35 if that's of any help

Again, apologies if I am missing something basic here...

`-- Performing Test CMAKE_HAVE_LIBC_PTHREAD -- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Failed -- Looking for pthread_create in pthreads -- Looking for pthread_create in pthreads - not found -- Looking for pthread_create in pthread -- Looking for pthread_create in pthread - found -- Found Threads: TRUE
-- Found CURL: /usr/lib64/libcurl.so (found version "7.69.1")
-- Configuring done CMake Error at third-party/CMakeLists.txt:15 (add_library): Cannot find source file:

imgui/imgui/imgui.cpp

Tried extensions .c .C .c++ .cc .cpp .cxx .cu .m .M .mm .h .hh .h++ .hm .hpp .hxx .in .txx

CMake Error at third-party/CMakeLists.txt:15 (add_library): No SOURCES given to target: imgui-for-imtui

CMake Generate step failed. Build files cannot be regenerated correctly.`

It seems that the table demo ("Tables & Columns > Vertical scrolling, with clipping") has bug, the first row is missing when you scroll down

Safari/605.1.15

MacOS Terminal/2.12.7

This is visible in the imtui online demo (so it should be easy to see): if you go under Widgets / Vertical Sliders the sliders have knobs that are 2 rows tall, except when at the very top they are being rendered as one row tall... but clearly with the space where the second row would be being rendered, so it isn't like it is somehow off the top of the slider or anything. Some of the sliders (the ones towards the left) are exhibiting the same behavior at the bottom.

Using imtui from within a tmux session, I get a completely blank screen. Looking in the debugger, the render loop does appear to be running, but nothing is visible. I believe tmux uses the alternate screen itself, so maybe we're toggling out of that mode and sending output to the wrong place? I also tested from screen and see the same behaviour, so I don't think it's a bug in tmux.

Do you have plans to upstream your changes to Dear ImGui? I would like to be able to dynamically choose between a TUI and a GUI (ideally even a remote web GUI) in a single binary but that requires a single version of Dear ImGui that I can use and have different code paths to enable backend-specific configurations within my tool.

When keyboard navigation is enabled, the borders around the selected control are drawn two characters wide, with one character of space between them and the selected control. This is huge in general and in things like combo boxes is incredibly confusing.

This can be fixed by patching ImGui::RenderNavHighlight method to disable both of the if (flags & ... blocks. It would be nice if the back end had some mechanism for telling Dear ImGui not to render these borders.

terminal emulator : windows terminal preview 1.12.2931.0 ncurses : 6.2

Cannot receive mouse movement events

Mouse movement works under ncurses-5.x, but not reporting mouse movements after columns 94