Opening a new terminal window from so x application - macos

I have a Mac OS X project (a game using SDL), and for debugging purposes when not running in xcode, I'd like to open up a terminal window that I can send text to and get line input from the user.
Is there a quick way to do this that doesn't involve creating a custom window (which is slightly problematic since the game uses SDL and I don't directly create windows)?
Opening up the standard OS X terminal would be fine, or even connecting to a separate terminal process and then sending output and reading input would also be OK.
The one criteria is that it needs to work when the application is run outside of xcode, or even on machines that don't have xcode installed.
I've spent the last few hours trying to Google this, but my searches are all returning unhelpful results. I'm clearing not seaching on the right keywords.
Thanks.

Actually every app has ability to interact with console, your just need to run it in a proper way (from a terminal or console)
Locate folder with your application. For XCode you can do it in a "Project Navigator" folder "Products", then for your app in context menu press "Show in Finder". For installed application it probably will be in the folder /Applications/. Let it be, for example, /Applications/MyBigProgramm.app.
Open console or terminal make cd to your app folder /Applications/MyBigProgramm.app/Contents/MacOS
There will be an executable file of your app.
Run it from console. After it stdin and stdout will be linked to your app.
Work with stdin and stdout as you want.
UPD 1
A bit of automation.
Create applescript with the following content:
tell application "Terminal" set currentTab to do script
("/Applications/MyBigProgramm.app/Contents/MacOS/MyBigProgramm")
end tell
Save your applescript either as app or script.
Run your base application ("MyBigProgramm") by starting your applescript.
Work with stdin and stdout as you want within your base application.
Watch output in the terminal, type sth in the terminal as input to your base application.

Related

OSX running app acts differently from running on terminal

I have written an app for OSX (Sierra) in Go (lang) with the GUI powered by the Golang QT binding therecipe/qt. When I build the app and run it from the command prompt with command open $(pwd)/myapp.app I get a terminal output while its running, and everything works fine.
When I double click on the app in finder to run it, I obviously don't have access to the console to see debugging, but in parts of the app that work fine in the GUI when I run from console (clicking on things, etc), it crashes the app when run by clicking on it.
I realise people don't know what the app is, and without seeing where in the app it crashes, they can't help me. What I am after is some ideas about debugging it, for instance:
Outputting logs to a text file and tailing the text file to watch as it is written to
Connecting a debugger to the application by the process, perhaps I can discover where it crashes. This doesn't sound straight forward
Is there a way to attach to a console of an application run like this to see what it is outputting?
Any ideas appreciated.
Thanks

Mac OS X: interacting with an application programmatically

I am working on a project where I need to call methods on an existing application (my own) and use some of its functionality. For e.g. my application ThunderBolt runs on Mac OS X 10.10. It also provides a dictionary of events that can be called externally through Apple Script or some other way that I don't know yet.
My question is what are the different (and better) ways of interacting with an application programmatically on Mac OSX? If I use something like the following code in Apple Script Editor:
tell application "ThunderBolt"
set open_file to (choose file with prompt "Choose the file you wish to parse")
set theContents to read open_file as data
set retPict to (image convert theContents)
end tell
then it is going to launch ThunderBolt with a splash screen and then call "image convert". This can be done via NSAppleScript but still it would launch the application and call methods/events on it.
Is it possible to somehow create an instance of (or get a pointer to) one of the class inside the application and use that? Something similar to COM or automation on a Windows system?
If you're working on OS X 10.10, you might consider taking a look at JavaScript for Automation (JXA).
With it you can apparently build methods into your app that can be invoked from client scripts written in JS (although I'm not yet familiar with the particulars of how to handle implementation of such a thing on the app side). But many of the apps that ship as part of OS X Yosemite have such APIs built in (e.g. iTunes and Finder).
Here's a great tutorial on JXA written by Alex Guyot: http://www.macstories.net/tutorials/getting-started-with-javascript-for-automation-on-yosemite/
The JXA-Cookbook repo also appears to be a nice resource: https://github.com/dtinth/JXA-Cookbook/wiki
Here's a brief example - this script makes iTunes go back one track. Try it while iTunes is playing (by putting the text into Script Editor, with the language option set to JavaScript, and hitting the Run button):
iTunes = Application('iTunes')
state = iTunes.playerState()
// Console msgs show up in the Messages tab of the bottom view:
console.log("playerState: " + state)
iTunes.backTrack()
Alternatively, you can place the code into a .js file and run it on the command line:
$ osascript itunes-backTrack.js
playerState: playing
The way you specify the 'tell application' is the best way, in my opinion.
What do you do with your app that needs to be called? Maybe some of the functionalities can be done with Applescript? It would simplify things a lot.

Launching app in foreground with Qt under MacOS X

I am having problems with a little Qt 5.0.1 program under Mac OS X 10.8.
(I have not tested any other platforms yet.)
I am launching an external Mac OS X program with this line of code:
QDesktopServices::openUrl(QUrl::fromLocalFile(fullpath));
Where fullpath contains a path to an application like:
/Users/schube/QTWorkspace/HelloWorld-build-Desktop_Qt_5_0_1_clang_64bit-Debug/HelloWorld.app/Contents/MacOS/../../../Aptus.app
(Aptus.app is a random app I've chosen, could be any app. I placed it in this path for testing purposes).
The application starts correctly but always in background; or at least, behind a Finder window. Really strange!
How can I force the new launched app to be sent to the foreground?
Use QProcess instead, but make sure that rather than using the path to the executable as the object to run, pass it to the open command as an argument. Something like this: -
QString cmd = QString("open %1").arg(fullpath); // may need QUrl::fromLocalFile(fullpath)
QProcess::startDetached(cmd);
Without using 'open', it will also open up behind other applications.
Note that you could also use the execute function, if you want to wait for the program to finish.
Also, with the open command, I think you only need to pass the path to the app bundle, rather than full path to its executable in Contents/MacOS. Either should work.
This might be a feature of the Mac OSX's window manager so that it does not steal focus.
You might need to alter your application to minimize.

Is it possible to create a Mac OS X gui app which is not a bundle?

I want to make an executable file (not the Mac .app bundle) which when run with a specific option (e.g. -gui) will pop up the gui.
For example, say I'm writing wget. I could do: wget www.google.com and that would print the result to the console, but if I instead do: wget -gui www.google.com that would pop up a gui window with the render of the html.
Is this possible in OS X (it is in windows)?
Note: I mean that the gui code is contained in the executable, calling open on another app is not acceptable.
Simple: Yes
Your second question will be how I guess.
Sometimes people ask me for applications that when launched by the Finder they should behave normal but when started by another process or terminal it should behave different. So, just like firefox, you can start the application the normal way or you can start it with special option.
Another way is creating agents. They are applications but don't appear in the dock and are designed to show interface elements when needed.
If you don't want to use a bundle per se you can just create a (cocoa) command line utility that loads an NSApplication when needed (depending on the command line options).

Calling a command line in MONO on MAC OS X

I want to be able to call the automator or unix commands like ls from a mono app and ge the results back.
This can be accomplished on windows easily. The question is how is this done on the mac??
caveat: I've never written a char of mono in my life.
I imagine it's a matter of redirecting stdout and firing up a process. this linux forum shows that you can do pretty much that - OSX will behave mostly as a UNIX-like system for you, I reckon.
Oh by the way, if you want to fire up an OSX application, have a dig around inside the ".app" bundle. OSX shows these as a file, but they're actually directories. In the finder you can right-mouse click and "show package contents", or you can open up a terminal / command prompt and cd into them. For instance, you can launch the Automator like this from the terminal:
/Applications/Automator.app/Contents/MacOS/Automator
I don't know if you would want to go down this route, but if you're going to be interfacing with OSX (gui) apps, you might want to look at using Applescript as some "glue" between Mono and the app.

Resources