We're developing an app that needs to run on a removable device (e.g. USB stick). On Linux, we're using Gnome launchers to place a shortcut to the app on the root of the device. However, we need to use relative paths for the executable and icon since we don't know in advance where the device will mount. In the .desktop file I have something like:
Exec=../myapp/myexecutable
Icon=../myapp/myicon.png
Neither the executable or icon is found. I read the spec on icon lookup in .desktop files (http://standards.freedesktop.org/icon-theme-spec/icon-theme-spec-latest.html#icon_lookup) but it didn't enlighten me.
Is there a way to get launchers to use a relative path? If not, is there another approach to achieve what I want (i.e. a shortcut with an icon and an executable, specified using relative paths)?
Relative paths are not supported*.
One solution is to have an installer. This script updates the desktop file according to the location the script is run from. Make the script executable, and the user can click it to install. The script requires the desktop file to be writable.
This was done with Linux in mind. The file is named autorun.sh; but that is just a convention, it usually won't run automatically. If you deploy this on something other than Linux, then name the file something else(autorun.linux), or adapt it to do different things according to the platform.
#! /bin/sh
#### Fixup $APPNAME.desktop.
APPNAME=xvscatter
ICONNAME=xv_logo.png
cd $(dirname "$0")
APPDIR="$PWD/$APPNAME"
EXEC="$APPDIR/$APPNAME"
ICON="$APPDIR/$ICONNAME"
sed -i -e "s#^Icon=.*#Icon=$ICON#" \
-e "s#^Exec.*#Exec=$EXEC#" "$APPNAME.desktop"
*The convention for the freedesktop is to have the icons in $HOME/.icons, /usr/share/icons or /usr/share/pixmaps. Under those directories are subdirectories for different icon sizes and types. When using one of those directories to store the icon, only the icon name(without the directory) is listed in the desktop file; otherwise record the full path to the file.
The executable, if in the path, can be listed with no pathname(unsafe). It's best to list the full path. Imagine the wrong program getting launched because the full path isn't specified.
Another possibility is to copy the desktop file to the user's desktop or to /usr/share/applications, and edit it there. Do this when the program is on read-only media.
Because none of the above results in a true install, if possible, use the platform's native installer and packaging tools(rpm,dep,portage, etc.). Those tools provide a framework for complete installation including the proper file permissions(think selinux), and desktop menus. They also provide for easy uninstall.
If the program has to run from the removable media, consider using the system install for just installing symlinks, maybe to /opt/vendor/progname.
What I did and worked perfectly was:
Exec=sh -e -c "exec \\"\\$(dirname \\"\\$0\\")/.sh/server.sh\\";$SHELL" %k
Explaining the command:
The snippet below will get the dir name of who is executing that, therefore the launcher dir name
$(dirname \\"\\$0\\")
So appending the desired path, will make this execute relative path.
Ref: https://askubuntu.com/questions/1144341/execute-shell-on-a-relative-path-on-ubuntu-launcher
Related
I am installing a package manually on my own system because I need to make some changes to it that aren't available in the basic version in my package manager. I also am trying to keep packages installed locally if possible, so I'm installing it with prefix=$HOME/.local instead of the more common prefix=/usr/local.
When I do this, I have no problem executing the program from my terminal, because I added ~/.local/bin to my PATH and the package was installed with relative paths to its shared libraries (i.e. ~/.local/lib/<package>). Executing from the command line is no problem, but I want to be able to access it from the favorites menu in gnome, and for that I need to make use of the <package>.desktop file.
I could hard-code the path to the executable in the .desktop file itself, but when I pull a later version down and re-install it, I'll have to redo those steps. I was wondering if there's a way to avoid that.
I've tried symlinking the executable to a directory where .desktop files do have included in their path, and the application is correctly treated as a GUI option, but launching the executable results in an error trying to find a shared library. I think this has to do with how cmake handles rpaths, which to my understanding is a way of relatively linking executables with their required libraries.
I think what I want to do is have PATH inside a .desktop file include ~/.local/bin, without changing the .desktop file itself. Can I alter the 'default' path used in accessing a .desktop file?
The answer to my question was found in the Archwiki:
Specifically, I needed to add ~/.local/bin to my path in ~/.xinitrc. Now my graphical programs work as expected.
According to Apple's Developer Docs the Library global allows one to import compiled scripts so they can be used as a library in one's current script. This works just fine if you were to do something like the below code with myLibName.scpt located at ~/Library/Script Libraries:
myLib = Library('myLibName');
myLib.myLibMethod() // Works just fine
But, the docs also claim that one can export an environment variable — OSA_LIBRARY_PATH containing a string of : delimited paths — and Library() would then defer to that list of paths before proceeding to it's default path: ~/Library/Script Libraries. Ya know, like the bash environment variable Path. Here's the relevant piece of documentation below; it describes the path hierarchy:
The basic requirement for a script to be a script
library is its location: it must be a script document in a “Script
Libraries” folder in one of the following folders. When searching for
a library, the locations are searched in the order listed, and the
first matching script is used:
If the script that references the library is a bundle, the script’s
bundle Resources directory. This means that scripts may be packaged
and distributed with the libraries they use.
If the application running the script is a bundle, the application’s bundle Resources
directory. This means that script applications (“applets” and
“droplets”) may be packaged and distributed with the libraries they
use. It also enables applications that run scripts to provide
libraries for use by those scripts.
Any folders specified in the environment variable OSA_LIBRARY_PATH. This allows using a library
without installing it in one of the usual locations. The value of this
variable is a colon-separated list of paths, such as /opt/local/Script
Libraries:/usr/local/Script Libraries. Unlike the other library
locations, paths specified in OSA_LIBRARY_PATH are used exactly as-is,
without appending “Script Libraries”. Supported in OS X v10.11 and
later.
The Library folder in the user’s home directory, ~/Library.
This is the location to install libraries for use by a single user,
and is the recommended location during library development.
The
computer Library folder, /Library. Libraries located here are
available to all users of the computer.
The network Library folder,
/Network/Library. Libraries located here are available to multiple
computers on a network.
The system Library folder, /System/Library.
These are libraries provided by OS X.
Any installed application
bundle, in the application’s bundle Library directory. This allows
distributing libraries that are associated with an application, or
creating applications that exist solely to distribute libraries.
Supported in OS X v10.11 and later.
The problem is that it doesn't work. I've tried exporting the OSA_LIBRARY_PATH variable — globally via my .zshrc file — and then running a sample script just like the one above via both the Script Editor and the osascript executable. Nothing works; I get a "file not found" error. I found this thread-where-the-participants-give-up-hope online; it doesn't explain much. Any thoughts?
On a somewhat related note, the Scripting Additions suite provides two other methods — loadScript and storeScript — that seem like they might be useful here. Unfortunately, when you try to use them, osascript gives you the finger. Though, I did manage to return what looked like a hexadecimal buffer from a compiled script using loadScript. Anyway, any insight you guys can shed on this would be much appreciated. Thanks.
The OSA_LIBRARY_PATH environment variable is ignored by restricted executables when running with System Integrity Protection enabled.
To workaround this limitation you can either turn off SIP, or you can use an unrestricted executable.
For instance, to make osascript unrestricted, you should first make a copy, and then re-sign it with an ad-hoc signature:
cp /usr/bin/osascript ./osascript
codesign -f -s - ./osascript
Once you have the unrestricted osascript, you can run it with the OSA_LIBRARY_PATH environment variable set like this:
OSA_LIBRARY_PATH="/path/to/libs" ./osascript path/to/script.scpt
As a lousy alternative, you can put a symlink at one of the "Script Libraries" folders that osascript would look at and point it to the folder you want. Note that the symlink must be a replacement for the entire folder, it can't just exist inside of it.
rm -rf ~/Library/Script\ Libraries
ln -s "/Your/Custom/Path/Goes/Here/" ~/Library/Script\ Libraries
Tested on 10.13.2
I have a c++ command line application that I have already compiled into an executable and have added it into my Xcode project. I have also added the "Copy Files" section to the Build Phases tab of the project properties and added my executable with the "Executables" destination. When I build my application I see it in the test.app/Contents/MacOS folder when I View package contents on the test.app that is built.
I also have App Sandbox enabled on the Capabilities tab of the project (so that I can distribute my application through the mac app store.
How can I expose this command line executable that is bundled with my application to the user so that they can run it from the command line (terminal)? I have not been able to find anything on search engines or on StackOverflow about how to get this file (or a symlink to this file) into the users PATH. I tried using an NSTask to create a symlink, but that only works if I disable the App Sandbox (which makes sense). Has anyone done this before? How did you get it to work? Or can these executables only be executed by code within your application?
I don't see a good way to do this. First, a clarification: the PATH is a list of directories that contain executables, not a list of executables; there's no way to add a single executable to the PATH. Instead, what you'd need to do is either put your executable into one of the directories in the user's PATH, or add the directory your executable is in into the PATH.
On OS X, the default PATH is /usr/bin:/bin:/usr/sbin:/sbin:/usr/local/bin. The first 4 directories shouldn't be modified from the system default, so only /usr/local/bin is a possibility. But creating it (it doesn't exist by default) would require admin (actually root) rights, which isn't allowed by App Store policies. So that's out.
That leaves modifying the user's PATH. The "right" way to do that system-wide is by placing a file in /etc/paths.d, which requires admin (/root) rights, so that's out too. Technically modifying the /etc/paths file would work, but that has the same permissions problem plus it's the wrong way to do customization.
The next possibility is to modify (/create) the user's shell initialization script(s). This'll work, but doing it at all right is going to be messy, because there are several shells the user might use, each with several different possible initialization scripts that the user might or might not have created...
Let's take a very simple case: a user who only ever uses bash, and who doesn't already have any initialization scripts. When a "login" instance of bash starts, it looks for ~/.bash_profile, ~/.bash_login, and ~/.profile (in that order), and runs the first one it finds. But your app doesn't know which shell he uses, so you'd better create ~/.profile so zsh and ksh will use it as well. So, your app creates ~/.profile, and puts this in it:
PATH="$PATH:/Applications/MyApp.app/Contents/Helpers"
Great, right? Yup, great, until the user runs something else that wants to set their PATH, it creates ~/.bash_profile, and this overrides your setup. After that, your executable will be in the PATH of zsh and ksh, but not bash. Whee.
And then one day the user decides to use tcsh instead, and it (and csh) have a completely different but equally messy pile of possible init files...
I have installed SenchaSDKTools-2.0.0-beta3-windows on my windows Xp 32-bit.
following is the command I am running on command prompt.
C:\Program Files\SenchaSDKTools-2.0.0-beta3>sencha app create MyApp c:\xampp\htdocs\sencha\myapp\
Giving following error
[ERROR] the current workind directory (C:\Program
Files\SenchaSDKTools-2.0.0-beta3>) is not valid SDK directory. Please
'cd' in to a SDK directory before executing this command.
environment variable has been set following way.
Variable name : Path
Value : C:\Program Files\SenchaSDKTools-2.0.0-beta3
can anyone figure out what exactly wrong in this process ?
*****Sencha Touch setup Guide steps(in Window)******
Download Sencha Touch SDK Tool. (http://www.sencha.com/products/touch/download/ -->SDK Tools Beta for Developers). Run .exe file. It will install tool to default path. (i.e. C:\Program Files\SenchaSDKTools-2.0.0-beta3)
Download Sencha Touch SDK. (Download openSource version file licensed under GPL.).
Extract SDK you just download in step 2.
Open Command line terminal (start->Run->type cmd). Change directory path to SDK path as you just extracted in step3.
Verify that Sencha Command is working properly on your machine. So Type Sencha.
E.g. sdkPath>sencha
you will see "Sencha Command v2.0.2" message with other sencha command detail.
Create app in your web directory by typing following.
E.g. sdkpath> sencha app create firstSenchaApp "path/ to/ www"
(If you have tomcat 7 installed in your computer, give path to \apache-tomcat-7.0.30-windows-x86\apache-tomcat-7.0.30\webapps)
Note: There must not be space in directory path name. _(underscore),- (desk) are allowed.
If app is not generated in step 6. There might be an error. There are as follow:
If error message appear is "sencha' is not recognized as an internal or external command, operable program or batch file.”, follow these steps to troubleshoot:
The path to SDK Tools directory is prepended to your system's PATH environment variable.
From the terminal, run
echo $PATH or echo %PATH% .
The SDK Tools directory should be displayed in part of the output. If this is not the case, add it to your PATH manually.
The environment variable SENCHA_SDK_TOOLS_{version} is set, with the value being the absolute path to the installation directory mentioned above.
For example: If the installed version is '2.0.0-beta2', a SENCHA_SDK_TOOLS_2_0_0_BETA2 must be set.
From the terminal, run
echo $SENCHA_SDK_TOOLS_2_0_0_BETA2 or echo %SENCHA_SDK_TOOLS_2_0_0_BETA2% on Windows.
If the output is empty, set the environment variable manually.
Wrong Current Working Directory
A common mistake is not running Sencha Command within either a valid SDK directory or an application directory. If the current directory is not a SDK or application directory, "sencha" command will fallback to backwards-compatible mode. As of SDK Tools release "2.0.0-beta2", you should see a clear warning in such case:
"The current working directory (...) is not a recognized Sencha SDK or application folder. Running in backwards compatible mode." So in this case follow proper step as mentioned above.
Reference:
http://www.sencha.com/forum/showthread.php?192169-Important-SDK-Tools-Sencha-Command-Update
There is an invisible file that you need to copy, called .senchasdk. Make sure it is in the folder.
I think you're having the same problem I had; I had attempted to run 'sencha app create...' from the folder containing sencha.bat. Instead, you need to run 'sencha app create...' from the sencha-touch-2.0.1.1 folder (which contains the .senchasdk file).
See this previous SO question!
Save my days.
By the way, if you can't extract the sencha-cmd to sdk tools, try to extract on other location and copy the folder [3.0.0.250] to your sdk tools > [bin] folder.
hope this helps.
If I try to build an application with the application class outside the default package, so the application file path is /app/AppClass.mxml instead of /AppClass.mxml (as would normally be the case), Flash builder cannot launch the application for debugging because it is looking for the SWF in debug/app/AppClass.swf and the SWF is being output to debug/AppClass.swf instead. Changing the output folder to debug/app makes it put the swf in debug/app, but then it puts the application configuration file "AppClass-app.xml" in /debug/app/app and then that can't be found.
Is there a way to change only the SWF output folder, or the location of the xml configuration file in the run-configuration?
You may use symbolic link to created swf file - http://en.wikipedia.org/wiki/Symbolic_link
for example for Windows :
cd project/path/bin-debug/package/path/
MKLINK ClassName.swf project/path/bin-debug/ClassName.swf
and it's work
or you can use symbolic link for folder:
cd project/path/bin-debug/package/
MKLINK path project/path/bin-debug/ /D
I think I remember this worked for me. But it was long time ago. And, yes, it is a known problem, I also recall Adobe people mentioning it as a limitation of FB.
In my Ant script, you'll need to do the adjustments to reflect your actual file names and directory structure. Also note that it will make it more cumbersome to debug it from FB. You'll need to use the debugging target in Ant, and then connect the debugger to the running application (so that some info, especially on the startup) will be lost. The only way you would be able to debug it, though I've never tried it, is with the commandline tools (I'm not sure of adl syntax for breakpoints / printing / stack frames, so idk how to do it.
Also, for the released application you will probably want to change the signing mechanism.