Get rid of it! :D

I'm writing this program to manage windows, which makes use of vi-like keystrokes to give the user more control over actions performed on windows.

It's in early stage, though I think I can get it done in a few weeks.

(07-08-2016, 10:58 AM)jkl Wrote: Get rid of it! :D

Yeah, I should. Learning learning sockets would be a good idea I believe.
I recall some guy publishing a book online about sockets in C for free. I've heard it's pretty good -- can't remember the name, though -- but should I instead read up on sockets in C++ instead of C?

EDIT: Found it: It's was Beej's Guide to Network Programming <https://beej.us/guide/bgnet/>

(10-08-2016, 03:37 PM)Tmplt Wrote: but should I instead read up on sockets in C++ instead of C?

It's not that different. The only thing that differs is how C++ will treat the objects (because of RAII and all). That is, if you use a C++ library for sockets.

Or write your own one. Or rely on the coming std::asio ... ;-)

(10-08-2016, 04:47 PM)jkl Wrote: Or rely on the coming std::asio ... ;-)

That's in C++20, I think. Isn't that a while off?

I guess the compiler support should improve really soon. :-)

(10-08-2016, 06:22 PM)jkl Wrote: I guess the compiler support should improve really soon. :-)

Not that soon, but there's a technical specification for a networking library.

I am working on a floating window manager with EWMH support that receives commands through X client messages (like vain's katriawm). The whole program is written with the XCB library.

This is the biggest programming project I have ever worked on, hope it ends up well!

The WM's name is TWM (tudor's window manager or sTrange window manager).

You might want to change that name. Too unclear.

I would second that. People already got confused with swm. If you did STWM you can just say sir tudor's window manager or STrange window manager.

STWM is a great name. Thanks for both of you!

Or go nuts and remove that useless "WM" suffix. Web browser do get names like "firefoxwb" or "surfwb". Text editors don't get names like "vite" or "nanote". Why does this non-sense still apply to window managers?

---

I've also started writing my own window manager, based on libwm (it's more a proof-of-concept, to show some example usage for libwm). It's named glazier. That's how the originals window manipulators were called, right?

I also started to code my file synchronizer: synk. It's supposed to work with inotify, to help keeping files in sync between two machines, in real-time. I'm still not sure about the implementation or model. But at least that's a good reason to start tinkering with internet sockets again!

As my blog is a weekend project and it's not weekend yet, I started working on a modern Gopher server. The de-facto standard pygopherd is annoying.

You should take a look at geomyidae. That's what I'm using personally, and it works pretty well!

AFAICS (without having tested it) Geomyidae won't work on non-POSIX platforms because dirent.h is heavily tied to that API.

Quasi-related to ya'lls Gopher escapades, I'm back to writing my own client... for Emacs. It's not quite interactive yet but it works more or less.

(17-08-2016, 02:36 PM)jkl Wrote: AFAICS (without having tested it) Geomyidae won't work on non-POSIX platforms because dirent.h is heavily tied to that API.

Probably. but that's why we have standards anyway ;)

Why would I want to restrict my software to a platform, just because a language lacks portability?

@rocx Neat. :)

I made a small static site generator in Go.

Right now I'm implementing some logic for defining content types. It should be useful for making blogs.

Version 1 of windowchef should be done.

Features:

• Groups
  
• Controlled via X client messages with
  Code:

  waitron
• Maximize windows vertically/horizontally/fully
  
• Respects ewmh (panel/docks and fullscreen programs)
  
• Can put windows in virtual grids. You define a virtual grid with a width and height and tell the window manager to put the window in a cell at the given coordinates.
  
• Snap windows in corners/middle of the screen.
  

As my first serious program, I'm creating a touch screen gesture maker for linux, using python and Tk GUI.

Imagine a small floating window with a 3x3 grid of buttons taking up a small portion of the screen. These buttons can be tapped, dragged, or swiped. Using one or a combination of taps, swipes, or drags on these buttons lets you execute programs, scripts, or simple commands. You can open your web browser, change the backlight or volume settings, even shutdown your computer with a simple tap and swipe. You can customize the button layout if you dont want a 3x3 grid. You can also change the combos and what commands will be made with those combos.

Another project im thinking of doing is bringing the messagease keyboard to linux, but another time.

(03-09-2016, 11:24 PM)mrmister Wrote: As my first serious program, I'm creating a touch screen gesture maker for linux, using python and Tk GUI.

In the python world there's the kivy project that also aims at applications with gesture based interaction. However that's at the software level not the whole window manager.

Keep us updated.

I have successfully created multiple button objects with its own personal set of commands and combos.

I am now working on getting the program to make buttons, combos, and commands based on a configuration file

Demo video will be posted soon

tryna keep my weight down cant keep goin to the shake shack

Doing some kernel tuning and I just wrote a small patch for st-0.7 to fix a bug that seems to just be on my machine. If anyone else is having issues with line drawing with bitmap fonts and chscale greater than 1, then maybe it can help.

So im now officially stuck on both programs im working on.

On my swipe n execute program, while the configuration file is being read, the buttons seem to end up getting all of the combos and commands that were set in the file.

I took a break on that one and got started on my keyboard program. My first problem was that I had a hard time finding an easy to install library that sent keystrokes virtually (without physical keyboard). I found that I can do that with 'xdotool', so now the program needs that program to work.

Another problem, which im now stuck on, is that when i click on the program, focus is sent to it, and not the window that needs to be typed on. I cant find a way to either prevent the program to take focus, or give focus to a previous window.

Programming is hard ;_;

I'm thinking about getting started on my file tagging project. I use xattr's a lot to classify and tag files for certain projects. I have yet to find an elegant way to find files with specific extended attributes or values of the extended attributes. I hope to make such a program capable of seeking out these files. I also make use of cdargs. I hope to combine the two concepts and extend cdargs by adding a 'selection stack'. Create a stack of files' paths/inodes and keep them in memory as part of the stack. You can then operate on the stack, which recurses to each file in the stack. In case your files you want to chmod/chown, cp, etc are in different paths. It's an idea I took from cdargs' bookmarked directories. I also want to use the stack as a means to quickly backup specific files. Probably useless to most others, but I would greatly appreciate the shortcuts provided.

Just getting started setting up my workflow with 2bwm. Today I configured basic notifications and will work on using notify-send to call `bar` as my default notification system.

Found a workaround for the window switching issue. I used xdotool again to activate a keybind set by the window manager (openbox) to change to the previous window, then type the key. This may not work with other window managers.