I'm trying to write my first Shell extension and - as every blog post out there - I am too struggling with the lack of up-to-date documentation...
Namely I am trying to grab all key strokes from the keyboard, and - looking at code online and old mailing list messages - it seems that until recently the correct way of doing was:
const Shell = imports.gi.Shell;
global.set_stage_input_mode(Shell.StageInputMode.FULLSCREEN);
global.stage.connect('key-press-event', myCallbackHere);
However this doesn't seem to work. Both the function set_stage_input_mode and the property StageInputMode seem to have gone, at least on my Fedora 24 running GTK 3.20.6.
What is the correct, current way to have my extension grabbing all keystrokes performed by the user?
I don't think there is a correct way to do this at the moment. You could hack something together that is either X specific or Wayland specific or maybe you could patch Gnome Shell to provide this information.
For Wayland you could start at https://github.com/MaartenBaert/wayland-keylogger and for X you could start by looking at reusing the key detection from autokey (https://github.com/autokey/autokey/blob/master/src/lib/interface.py).
I can guess this is not the answer you were hoping for, because on IRC you wrote that you want to write a replacement for autokey. For getting input from an external program you can use GLib.spawn_async_with_pipes (https://people.gnome.org/~gcampagna/docs/GLib-2.0/GLib.spawn_async_with_pipes.html).
I wrote some time ago some Shell extension code to poll xinput test and log roughly what's happening (think any alphanumeric key vs left ctrl vs return key). You can look at that code here: https://gist.github.com/daniellandau/7679741bf8bbc5c345591593ca05e9f6. It's not robust enough for doing any kind of macro expansion so I'd recommend reusing the detection code from autokey.
In general the docs at https://people.gnome.org/~gcampagna/docs are the place to find API references for libraries usable from Gnome Shell extensions. I hope you get ahead with your project. Starting out writing extensions is frustrating but eventually very fun once you get the hang of it.
Related
Instead of recording my own keystrokes to create an asciinema tutorial I do want to script them, so I can easily improve my tutorial over time without being being stressed about redoing it each time I want to make a change, or spending a lot of time trying to correct typos in the recording.
Mainly I want to be able to do somethign like:
type: "df"
wait: 2s
type: "echo foo"
This would enable me to easily rebuild the recording from scratch when I change the script.
How can I do this?
I've searched for a lot of solutions for my own projects, and this is the solution that I've come up with:
My solution:
What I'm planning on doing for a project I'm in is to use both asciinema as well as demo-magic.sh.
Specifically, I'll be using asciinema for the recording with the -c flag (see here), and demo-magic.sh for the automated typing, since it supports waiting for commands to finish executing (or not) and custom wait times on top of that. The command would look something like asciinema rec -c "./mydemoscript.sh" myrecording. See the projects for proper usage.
Other things I've come across:
doitlive -- I'm not using this because it's more of a fake-typing automated thing where you actually have to "type like a madman", which could be useful if you're doing a talk in front of an audience but you don't want to mess up your keystrokes, or if you don't want to forget to do certain commands.
asciiscript -- it's written in Go and it works, but you need to compile it yourself, and it also doesn't support waiting for the previous command to finish.
spielbash -- a Ruby project designed specifically to automate asciinema recordings using tmux. I'm not using this, though, because it's not as portable since you need an existing Ruby installation, and more importantly, it keeps corrupting the active console I'm in, and there's also unpredictable corruption in the recordings. Also, the project hasn't been committed to since 2019.
termscript -- it just doesn't work for me, and it freezes my console. It runs on python 2.
UPDATE: I have recently come across terminalizer. It seems to provide support for an window-like border around the actual terminal, and it has a built-in GIF renderer, but I have not tried it.
There is a project svg-term which allows to record a screencast from a command, and I've seen a script that simulates commands being typed letter after letter.
For the moment I ended up using AppleScript to automate iTerm but this approach is buggy and locks you to MacOS and iTerm... and if you happen do to something during recording, you are doomed.
I am still looking for a better approach.
TL;DR: asciinema-automation should do what OP wants. Its dependencies are asciinema and pexpect.
(disclaimer: I developed it for my own usage)
Some time ago, I was also looking for a way to automate asciinema recordings and I saw the very nice answer of a-la-linuques. I chose asciiscript because it used asciinema, but it is not maintained anymore. I tried to keep on using it via a fork adding new small features, but being not very familiar with go, I finally decided to rewrite everything in python in this repo.
It reads bash files, where comments can give special instruction like adding time between command or key stroke, or wait for an expected output (see examples). Of course, this is very much in the spirit of asciiscript.
I am very new to making applications, and using Xcode, so I don't know much about it... I'm using Xcode 5.0.2 to attempt to create a basic osx application which will allow the user to search through a list of built in scripts, select the ones they would like to run, and push a button to run the scripts. I have several scripts already, all of which have the extension .scpt. i wrote the scripts using applescript editor.
From searching this website, and the rest of the internet, i haven't found a clear, step by step answer of what to do. I know that in order to run the shell script, there must be an interface button linked to an action in the implementation file, which is what will execute the script.
What is the action i have to put in the implementation file?
How do i link the button to the action?
where do i put my .scpt files?
I have absolutely no code written in Xcode, it is nothing but a blank new project.
I don't have much experience with Cocoa apps, if you have experience with Visual Studio this is nothing like it, I started by looking some videos from the link below, they're really helpful, he also provide a full course for Objective-C.
http://www.youtube.com/playlist?list=PLE83F832121568D36
I hope this helps!
There is a huge amount of aspects to the problem you are trying to solve. Honestly, connecting the NSButton to an Action is just taking the first (small) step! Just a few of the aspects you may encounter:
How are the files executed? (your main question)
How is the selection interface set up?
How to read from the files? How to process the data you read?
What kind of actions will your scripts perform? What is in fact allowed from within a Cocoa app? (I have no idea myself to be honest)
Each one of these suggests potential SO questions in itself.
I would suggest that you familiarise yourself with the environment and the language a bit before taking big leaps. Take some time (a few days at least) and follow some tutorials, program some simple apps. The link provided in the other answer is in fact a good one, AppleProgramming tutorials are easy to follow. You may also find Apple's Objective-C intro useful.
I am quite familiar with MATLAB, although not so much with writing GUIs in it. Now my boss gave me a MATLAB program with GUI which was developed on a Windows machine and runs there without problems. I copied it to my mac (osx 10.6.5 with MATLAB R2009b) using a USB memory stick and tried to work on the files on my computer.
When starting the GUI, some of the callbacks are executed, some not. I was able to open the figure file containing the GUI in GUIDE, but there cannot click on the callback fields in the Property inspector. Also some components are not shown in GUIDE which are visible when running the GUI. First I suspected it might be corrupted by changed newline characters, but learned that .fig files are binary, so this should not be the case. I opened the binary .fig file with a text editor and found the following ASCII line followed by binary data:
MATLAB 5.0 MAT-file, Platform: PCWIN, Created on: Tue Aug 10 17:11:57 2010
Are the GUIs created with MATLAB OS specific? Wouldn't this be really stupid?
I opened the files on the windows machine again and tried to export them or save them by specifying that I want them to be compatible with other operating systems but could not find an option like this.
Of course I researched this on google, and wondered that I did not really find something related. If an incompatibility exists, I think this would have been discussed, right?
Anyone knows about it or had the same problem before?
Hmmm, ok. Well, the answer is quite embarrassing but still bears some potential to learn from it. The folder that my boss gave me contained a lot of files and figures with similar names, like:
bla1.fig
bla.fig
bla0.fig
the simple problem was, the first time I opened the figure with GUIDE I opened figure bla.fig instead of bla1.fig. After this I always used the reopen function of GUIDE and did not see that it was always the wrong file. this caused the problems.
==> so there is no incompatibility between Mac and PC, although I wonder why the platform is written in the file then
==> it could have been a problem with backwards-compatibility as the original figure was created with MATLAB 5
==> as #Adrian and #Mikhail pointed out: it would be very very useful to use a Version control system in order to prevent to have hundreds of files with different numbers where most of them are of no use anymore
There seems to be some incompatibilities between different platforms and also between different versions, from my experience (MATLAB 5, 6, and 7) also -- I've encountered similar situations many times.
I am not exactly sure when the GUIDE became available with MATLAB -- I first noticed about it when I was using the first version of MATLAB 7. (It used to be very unstable at first, but is very usable now. For example, I vaguely remember a critical issue with MATLAB 7 Student version. You had to download a patch to make the GUIs work.)
Does your boss know how the *.fig file was created and in which MATLAB version? Can you open the *.fig on the old version of MATLAB and display the GUI correctly?
Some functions have possibly become obsolete since the earlier version of MATLAB that was used to create the figure, and may be causing some minor problem with it. If that is the case, you can troubleshoot it by checking all the error and warning messages that appear on the Command Window (or the MATLAB Console.)
Also, you should be able to load the *.fig files on the workplace using the following command from the Command Window, assuming that the files are in MATLAB's search path.
load your_file.fig -mat
You should then see a structure dataset, and see all of the figure properties, etc. in that structure.
You can use the information to create a new figure without the GUIDE (GUI Development Environment / Tool).
I recommend you writing your own GUI app. It's not that difficult, with/without the GUIDE.
I recently opened a site with some Tips on MATLAB Programming for beginners and intermediate users, and update it in my spare time. If you have any specific question about GUI building, I may (or may not) be able to provide answers there...
I'm a newbie to Mac and trying to use AppleScript to ease my daily job. One of the things I want to achieve is to automatically create Smart Mailbox based on the groups in my Address Book. While I have no problem automate Address Book, there's no much article about how to automatically create Smart Mailbox. After some searches on the web I realized it can only be done through GUI scripting. But most of the answers are about how to bring up the new Smart Mailbox dialog but no further information about e.g. how to select different rules. I tried to use AppleScript Editor to record my activity so that I can learn from that but found out after recording, there's nothing get recorded! I tried with other apps like Finder the recording does work! If anyone knows how to record it, or how to get more information on this type of GUI scripting would be very appreciated!
Unfortunately Mail's Applescript support is more than a little maddening at times due to being so limited. Control of smart mailboxes is but one example. I'd hoped SL would have fixed this but Apple's thinking about Applescript can be pretty schizophrenic at times. I'm not sure they have decided what to do with it.
I'd got a Smart Mailbox creation script working but it never worked reliably so I never deployed it. If you really want it I can give you the code - although it's written in Python + Appscript rather than pure Applescript.
Were I you I'd submit this as a bug to Apple and hopefully it'll be fixed with Lion.
The problem with GUI scripting as I'm sure you discovered is that figuring out how to get the name of each particular element can be maddening. There are a few utilities out there that will help, but unless you are doing a lot of GUI scripting they probably aren't worth the cost.
Regarding Applescript recording this was a really nice feature often supported under the old MacOS. However with the switch to OSX almost no application supports it so it's basically a dead feature for all practical intents.
Terminals and shells are very powerful but can be complicated to learn, especially to get the best out of them. Does anyone know of a more GUI based command shell that helps a user or displays answers in a more friendly way? I'm aware of IPython, but even the syntax of that is somewhat convoluted, although it's a step in the right direction.
Further to this, results could be presented graphically, e.g. wouldn't it be nice to be able to pipe file sizes into a pie chart?
Hotwire is an attempt to combine the power of the traditional command line interface with GUI elements. So it has a GUI side, and tries to be helpful in suggesting commands, and in showing you possible matches from your history. (While there are keyboard shortcuts to do this in bash and other shells, you have to know them ...)
You can use all your common system commands, but a number of key ones have new versions by default which use an object pipeline, and are displayed with a nice GUI view. In particular ls (aka dir) shows lists files and shows them in columns. You can sort by clicking on the column headers, double click on files to open, or double click on directories to move to that directory. The proc command allows you to right click on a process and one of the options is to kill it.
The object pipeline works in a similar way to Microsoft Powershell, allowing commands in the pipe to access object properties directly rather than having to do text processing to extract it.
Hotwire is cross platform (Linux, BSD, Windows, Mac), though it is at an early stage of development. To learn more, install (click on the link for your platform) and work through the simple getting started page.
If you don't like hotwire, you could also look at the list of related projects and ideas maintained on the hotwire wiki.
fish is a Unix shell that focuses on user-friendliness, such as by providing colored highlighting and extensive tab completion.
For a different kind of blend of textual and graphical interface, there's Quicksilver, as well as similar/inspired tools like Launchy, GNOME Do and ENSO.
Is this for Python in particular, or are you just interested in any command shell that has a GUI interface?
If the idea of piping file sizes into a pie chart interests you, you might try PowerGUI, a GUI layer on Microsoft's PowerShell command shell. PowerShell also lets you pipe data from commands into XML, CSV, and other formats that are understood by GUI programs.
GUI-based command shell seems like an oxymoron to me.
Not really? A command shell is just an encapsulated environment in which to execute commands. Why can't they have GUI extensions? We are in the 21st century! :)
Check out http://hotwire-shell.org/
This is along the lines of what I was thinking. It's a shame it uses PyGTK, I'd have preferred PyQT (perhaps a licensing issue?). There look to be some interesting related links from the project as well.
If the idea of piping file sizes into a pie chart interests you, you might try PowerGUI, a GUI layer on [...]
PowerGUI looks like a hobby project I've been working on that organises regularly used tasks. It looks like it organises frequent jobs and formats the output for you. The formatting I see as the end result of the data flow. But it would be nice to be able to tinker with data and then continue to use it.
PowerShell as a command shell is very forgiving for new users and is easy to learn. There is an add-on product (it is a commercial product) called PowerGadgets that would let you pipe file sizes into a pie chart or other types of displays
PowerGadgets looks very interesting. It would be interesting to have things like system monitors so that you could say, read the CPU usage per second and pipe it into a graph.
Is this for Python in particular, or are you just interested in any command shell that has a GUI interface?
Any really, currently, but I like the idea of cross platform, easy to edit, no compiler setup. I use Windows at work and Windows/Linux (Ubuntu)/OSX at home. Python is just an easy solution, and for writing stuff like this is has a lot of libraries already.
Thanks for all the links. Keep them coming. :)
I'm not sure whether you're asking for a shell as in bash/csh, or a shell as in ipython. If it's the later, then I'd recommend looking at Reinteract. While it's still very alpha, it's already a great tool for rapid prototyping in python, and allows embedding of plots, widgets, etc.
GUI-based command shell seems like an oxymoron to me.
The key-word here is Graphical.
If I want a GUI, I want a full-featured GUI. But if I want raw performance, I want a command line.
I'm not exactly sure what you're asking for. You can either have a GUI or a command line. What do you need from a graphical command shell that you couldn't get from a straight GUI?
Also, if you want graphical information about file sizes there are a few applications that do that. One example is WinDirStat.
Also not related to Python, but Ubiquity (a firefox extension) is a graphical command-line-like tool for the web, with a similar spirit to Quicksilver/Launchy/GnomeDo.
I know that Automator in Mac OS X is not a shell but it is the best graphical tool I have ever used to do batch tasks. I think it is worth mentioning here as even I (self-titled as a power user) use it from time to time to rename files or other routines. Although these could be done in a few lines of shell script, the Automator's graphical interface makes me feel like I am not working and it just works.
Check out http://hotwire-shell.org/
PowerShell V2 is developing a graphical command shell, but I don't think that is what you are looking for.
PowerShell as a command shell is very forgiving for new users and is easy to learn. There is an add-on product (it is a commercial product) called PowerGadgets that would let you pipe file sizes into a pie chart or other types of displays. Information about that can be found here.
As for ease of use, PowerShell command follow a Verb-Noun pattern (along with aliases for ease of use from the command line) and is very discoverable. Check out some screencasts I did on using PowerShell at PowerShell Basics.
I found POSH, a GUI for MS PowerShell. This is pretty much what I intended. You have a command-line backend with a WPF GUI frontend. You can pipe results from command to the next and show graphical output.
Maxima provides a mathematical shell (screens) . It is nice that you type in a C-like syntax and receive graphically formatted output.