Find the path of an application, and copy a file to that directory in Inno Setup - matlab

I'd like to install a file into a user's MATLAB folder in Inno Setup. But depending on the version of MATLAB, the directory can change.
In the Windows command line, it is possible to get the path of the MATLAB executable like so:
where matlab
Which will output
C:\Program Files (x86)\MATLAB\R2015b\bin\matlab.exe
I'd like to copy a file in the following folder
C:\Program Files (x86)\MATLAB\R2015b\toolbox\local
How can this be done?

The where command searches the file in the path specified by the PATH environment variable.
In Inno Setup Pascal Script, you can implement that using FileSearch function, like:
FileSearch('matlab.exe', GetEnv('PATH'))
Though I'd say, that there's must be a better way to find installation folder of MATLAB.
Anyway, you can resolve the path using the above method to a global variable in InitializeSetup event function. It will also allow you to abort the installation, when MATLAB is not found.
And then you can use the variable as an installation path using a scripted constant.
[Files]
Source: "MyFile.dat"; DestDir: "{code:GetMatlabToolboxLocalPath}"
[Code]
var
MatlabToolboxLocalPath: string;
function GetMatlabToolboxLocalPath(Param: string): string;
begin
Result := MatlabToolboxLocalPath;
end;
function InitializeSetup(): Boolean;
var
MatlabExePath: string;
begin
MatlabExePath := FileSearch('matlab.exe', GetEnv('PATH'));
if MatlabExePath = '' then
begin
MsgBox('Cannot find MATLAB', mbError, MB_OK);
Result := False;
Exit;
end;
MatlabToolboxLocalPath := ExtractFilePath(MatlabExePath) + '..\toolbox\local';
Result := True;
end;

Related

How to reload a postgres C extension?

I've created a C extension defining the shell of a function that I can call from SQL code in Postgresql 12.3. I'm using PGXS to build and install the extension. I can build, install and call the function, but if I make changes and reinstall, the changes don't show.
The make file is:
MODULES = bar
EXTENSION = bar
DATA = bar--0.0.1.sql
PG_CONFIG = pg_config
PGXS := $(shell $(PG_CONFIG) --pgxs)
include $(PGXS)
bar.control is
comment = 'Simple bar function'
default_version = '0.0.1'
relocatable = true
module_pathname = '$libdir/bar'
bar--0.0.1.sql is:
CREATE OR REPLACE FUNCTION
bar(jsonb) RETURNS int AS 'MODULE_PATHNAME','bar'
LANGUAGE C STRICT;
bar.c is:
#include "postgres.h"
#include "fmgr.h"
PG_MODULE_MAGIC;
PG_FUNCTION_INFO_V1(bar);
Datum bar(PG_FUNCTION_ARGS)
{
PG_RETURN_INT32(101);
}
I run make install to do the install. And then
drop extension bar; create extension bar; but the changes won't show when I call the function. The only way I can install a new version is to restart postgres. What step am I missing?
Well, I discovered that if I just create a new connection to the DB the changes will show. That isn't so bad, I suppose. Does that have to do with how postgres uses fork() to create a new process for each connection?
When you execute your function, the shared library gets loaded into your backend process and stays there. Replacing the shared library on disk does not modify the in-memory copy loaded into your process. You have to start a new database connection and execute the function again to get the new version.

how to open new instance of acrobat

I'm using windows 10 task view and need to open different instances of excel word and acrobat.
I can't seem to get acrobat working. I get error: 0x800401F3 - Invalid class string on the AcroApp := ComObjCreate("AcroExch.App") line.
Any suggestions?
^+n::
oWord := ComObjCreate("Word.Application")
oWord.Documents.Add
oWord.Visible := 1
oWord.Activate
xlApp := ComObjCreate("Excel.Application")
xlApp.Visible := true
xlApp.Workbooks.Add()
xlApp := ""
run notepad
run chrome
AcroApp := ComObjCreate("AcroExch.App"); Error when running
AcroApp.Visible := true
AcroApp.Open
return
While the COM approach seems cool, it seems really unnecessary.
I don't have Acrobat, but a quick Google search tells me they have a command line option to open a new instance, as I suspected (documented here).
So, a quick little run command should do the trick:
Run, % """C:\Path\To\Acrobat.exe"" /n"
Also, maybe this is why your COM approach didn't work?

How to import other source code files in dm script

Is there a way to use multiple code files in dm-script to structure the code? Something like:
import "path/to/utility_functions.s";
utility_functions.do_something_general();
Note that I do not want to have the code as a menu item if possible. The code contains only functions that I use in the main script.
I tried the following:
File 1: test.s
void test(){
result("test\n");
}
File 2: require-test.s
AddScriptFileToPackage("path/to/test.s", "test", 3, "test-function", "", "", 1);
ExecuteScriptString("test()"); // works immediately but feels wrong
test(); // works after restart
Now I have the following problems:
I have to restart DigitalMicrograph after executing this script, otherwise test() does not work (ExecuteScriptString("test()"); works but it feels wrong to use strings for invoking code, if possible I'd like to avoid that)
When I restart DigitalMicrograph another time AddScriptFileToPackage() sais 'The script cannot be added because the package exists and is read-only. [...]'. Is there a way around it or do I have to use try blocks?
I feel like I am not doing this wrong at some place.
DM script does not support on-demand-loading of packages, but there are two different ways to achieve what you want:
Using library packages
When you "install" a script, you can choose to either install it as menu-command or as a library. This is true for both installing scripts via the menu command (which get stored in the global preferences file) or via the scripting-command (which can be used to
create .gtk files which one can then add/remove from the plugins
folder as needed).
The "menu" option will install a script such that it is invoked once via the menu-item but does not stay in memory otherwise.
The "library" option will execute a script once on startup and keep the script itself in scope. So you can define methods (or classes) in a library file and have it generally available. And you can put some executable code in a library if you need some startup-action.
Using script libraries as .gtk plugins is possibly the recommended way to achieve what you want. They are, however, always loaded.
Piece of advise: If you make libraries ensure you use very unique class and method names to avoid any conflict. I would recommend pre-fixing all class/method names with some library-name, i.e. use MyLib_MyClass instead of MyClass and the like.
Clarification: Scripts added as library packages are permanently added to the software, i.e. those packages get created once and are then placed in the plugins-folder. They will always load on startup of DM and be available. The Library package method is not suitable for temporarily 'loading' external scripts. Such 'on demand import' is not supported by DM-scripting.
The script commands to create packages are utility commands to help one create packages in an easy and manageable way. Typically, one would create a "Create package XY" script with several such commands adding all scripts from a location into a package. The script would be called once to create the package-file (It is already in the plugins folder afterwards.)
Only when the included scripts change and the package therefore requires to be updated, is the create-package script called again. Note, that in this case it is first required to remove the package-file from the plugins folder and start DigitalMicrograph without loading it, so that a new package is created. Otherwise the script would append to the package, which would not be possible if methods of the same name already exist in the package.
The F1 help documentation has an example script:
A typical examples, using GMS 3.4.0:
Script stored at: C:\Tmp\testLib.s
void TestCall()
{
Result("\nTest")
}
Script stored at: C:\Tmp\menuAction.s
Result("\nPerforming an action here.")
One-time run script to install a package:
// General package parameters
// *********************************************
string pkNa = "myPkg" // Filename of plugin
number pkLe = 3 // level 3 (.gtk) only needed for load order
string pkLo = "user_plugin" // plugin location
string scriptRoot = "C:\\Temp\\"
// List of Scripts to be installed as menu items
// *********************************************
// Each entry needs a (unique) command-name, a menu-name and an optional sub-menu name.
// The "isLibary" flag is set to 0
// It is possible to add the same script multiple times. The script will be executed when the menu item
// is chosen. Methods and Classes of the script are not available otherwise
// A separator can be added by installing and empty script with a (unique) command name starting with "-"
AddScriptFileToPackage( scriptRoot + "menuAction.s", pkNa, pkLe, pkLo, "Call 1", "MyMenu", "MySubMenu", 0 )
AddScriptFileToPackage( scriptRoot + "menuAction.s", pkNa, pkLe, pkLo, "Call 2", "MyMenu", "", 0 )
AddScriptToPackage( "", pkNa, pkLe, pkLo, "-sep1", "MyMenu", "", 0 )
AddScriptFileToPackage( scriptRoot + "menuAction.s", pkNa, pkLe, pkLo, "Call 3", "MyMenu", "", 0 )
// List of Scripts to be installed as library
// *********************************************
// Each entry needs a (unique) command-name. Menu-name and sub-menu name are "".
// The "isLibary" flag is set to 1
// The script will be executed once on startup (if there is executable code). It is also executed once
// here during the install.
// Methods and Classes of the script are permanently available and need unique names.
// Adding a script to the package as libary can be used to create on-load-version info output.
AddScriptFileToPackage( scriptRoot + "testLib.s", pkNa, pkLe, pkLo, "library-1", "", "", 1 )
AddScriptToPackage( "Result(\"Script packages myPkg loaded.\\n\")", pkNa, pkLe, pkLo, "myPkg-versionInfo", "", "", 1 )
After running the install-script there will be:
A menu like this:
Output in the results window like this:
A package file in the folder C:\Users\USERNAME\AppData\Local\Gatan\Plugins\myPkg.gtk
The script command TestCall() generally available in all scripts.
The package will load each time when DM starts as long as the .gtk file remains in the plugins folder.
Calling script code from within scripts
The scripting language supports two commands to call a script from within a script:
Number ExecuteScriptString( String text )
Number ExecuteScriptFile( String file_path )
Using the command to execute scripts form disc can do what you want, but maintaining a useful 'library' that way could be tedious. It also does not allow you to install classes.
Example of calling a script from within a script:
// Direct example
void Demo()
{
ClearResults()
Result( "I am a test call.\n")
number n = 5
Result( "I am working on the number: " + n )
}
Demo()
//Having the script as a string
number otherNumber = 11 // To show how you can modify a script call as an example
string scriptStr
scriptStr += "void Demo()\n{" + "\n"
scriptStr += "ClearResults()" + "\n"
scriptStr += "Result( \"I am a test call.\\n\")" + "\n"
scriptStr += "number n = " + otherNumber + "\n"
scriptStr += "Result( \"I am working on the number: \" + n )"+ "\n"
scriptStr += "}\n"
scriptStr += "Demo()\n"
If ( TwoButtonDialog("Script-call","Show it", "Run it") )
{
ClearResults()
Result( scriptStr )
}
else
ExecuteScriptString( scriptStr )
The following explicit example of build script usage may be closer to what you are looking for. It shows that in the course of a single DM session, one can edit the module source files and repeatedly rebuild the package without having to relaunch DM, contrary to the clarification about package creation provided in the answer from BmyGuest. This example also makes use of the very convenient GetCurrentScriptSourceFilePath function which greatly simplifies file path references when one can locate the build script and module source files in the same folder (this is the approach I take with my own development projects).
Here is the arrangement of my files for this example:
The two source modules are very simple function and class libraries.
Here is Module1:
void Module1SayHello()
{
OKDialog("Hello from module 1");
}
And here is Module2:
class Module2TestClass
{
void Module2SayHello(Object self)
{
OKDialog("Hello from module 2");
}
}
Here is the build script:
void main()
{
// Establish the source code directory relative to the current build script location
String buildScriptSourceFilePath;
GetCurrentScriptSourceFilePath(buildScriptSourceFilePath);
String sourceFileDir = buildScriptSourceFilePath.PathExtractDirectory(0);
// Add the modules
AddScriptFileToPackage(sourceFileDir.PathConcatenate("Module1.s"), "MultiModuleTest", 3, "Module1", "", "", 1);
AddScriptFileToPackage(sourceFileDir.PathConcatenate("Module2.s"), "MultiModuleTest", 3, "Module2", "", "", 1);
}
main();
Contrary to the above-mentioned clarification, this build script can be run multiple times during a DM session and the content of the package file gets replaced each time. So now one has a very nice development environment where one can open the source file for a module, edit it as desired, save it, and then rebuild the package file. One can use the following test script to see that the behavior changes as one edits, saves, and rebuilds the implementation of any function or method in the module source files:
void main()
{
Module1SayHello();
Alloc(Module2TestClass).Module2SayHello();
}
main();
Because of the way the DM script interpreter parses, tokenizes, and executes code, all functions and methods invoked anywhere in a script must have been previously defined before a script is executed. This is why the above test script, or any other script that uses the added modules, cannot simply be appended to the end of the build script (except if embedded in a string passed to the ExecuteScriptString function, as pointed out in the posed question). The concept of imported code modules (e.g. as in Python) is therefore not really possible in DM scripting (as pointed out in a comment to the answer by BmyGuest). In this sense, DM scripting shows its roots in 1990’s coding concepts, which commonly involved separate compilation, linking, and execution phases.
Nevertheless, the build script approach described here allows one to take advantage of the features of a true integrated development environment (IDE). For example, one can add the module source files (and build script) to a project in Visual Studio and get all the benefits of a modern multi-file code editor and revision control (e.g. via Git). This is what I do with the Enabler framework.
The one caveat is that once the DM session is closed, the plug-in (package) file does become finalized in some way so that it can no longer be replaced by the build script in a future DM session. In this case, one does have to remove the package file from the plug-ins folder before resuming another development session in DM (as covered in the clarification from BmyGuest).
For everybody else who needs this, I am using AddScriptFileToPackage() now, inspired by both, #BmyGuest and #MikeKundmann.
The following main.s is always open in my GMS. The real code I'm working on is in program.s. To test your code execute the main.s. This file can be executed multiple times in one session!
For opening GMS I use the (Windows) batch file below. This deleteds registered plugins automatically which makes the main.s usable again. For debugging I created a python script that combines all the files listed in the main.s. This way GMS jumps to the errors. This python program can be downloaded from my github page.
/**
* File: main.s
*/
String __file__;
GetCurrentScriptSourceFilePath(__file__);
String __base__ = __file__.PathExtractDirectory(0);
/**
* Load and add the file `filename`, the name will be the `filename` without
* the extension.
*
* This is dynamic only for the current session. If GMS is restarted, using
* this will create errors except if the plugins folder does not contain the
* required files (delete `%LOCALAPPDATA%\Gatan\Plugins\` before starting).
*
* #param filename The filename (or path) relative to the path of this file
* #param name The internal name to register the script with
*/
void require(String filename, String name){
// AddScriptFileToPackage(
// <file_path>,
// <packageName: filename of .gtk file in plugins>,
// <packageLevel: load order [0..3]>,
// <command_name: id/name of the libary/command>,
// <menu_name: name of the menu, ignored if isLibrary=1>
// <sub_menu_name: name of the submenu, ignored if isLibrary=1>,
// <isLibrary: wheter to add as library (1) or as menu item (0)>
// )
AddScriptFileToPackage(__base__.PathConcatenate(filename), "__require_main_" + name, 3, name, "", "", 1);
}
/**
* Require the file `filename` with the basename of the `filename` as the name.
*
* #see require(String filename, String name);
*
* #param filename The filename (or path) relative to the path of this file
*/
void require(String filename){
require(filename, PathExtractBaseName(filename, 0));
}
void main(){
// add libaries
require("string-lib.s");
// add main file
require("program.s");
}
main();
The (Windows) batch file to start GMS. This deletes the plugins folder automatically. Then the main.s does not cause any problems.
#echo off
rem
rem File: start-gatan.bat
rem ---------------------
echo Deleting GMS cached libaries...
SET plugins_path=%LOCALAPPDATA%\Gatan\Plugins\
SET gms_path=%PROGRAMFILES%\Gatan\DigitalMicrograph.exe
if exist %plugins_path% (
echo Deleting all .gtk files in %plugins_path%...
del %plugins_path%__require_main_*.gtk /F /Q
del %plugins_path%__require_main_*.gt1 /F /Q
del %plugins_path%__require_main_*.gt2 /F /Q
del %plugins_path%__require_main_*.gt3 /F /Q
if exist "%gms_path%" (
echo Starting GMS
start "" "%gms_path%"
) else (
echo GMS path %gms_path% does not exist.
pause
)
) else (
echo Plugins path %plugins_path% does not exist.
pause
)

Can't test if directory already exists from within Fastlane with Ruby

I can’t test whether a directory exists from within a Fastlane action nor lane. None of the following work for save_path = "./dir_name" nor save_path = "dir_name" in the current directory (from which fastlane is being run):
!Dir.empty?(save_path)
!Dir[save_path].empty?
Dir.exist?(save_path)
File.exist?(save_path)
File.directory?(save_path)
I even tried to expand the relative path:
File.exists? File.expand_path(save_path)
I have referred to the following:
Check if directory is empty in Ruby
How to check if a given directory exists in Ruby
https://blog.bigbinary.com/2017/02/28/dir-emtpy-included-in-ruby-2-4.html
https://ruby-doc.org/core-2.2.0/Dir.html
How does one test to see if a directory exists from within Fastlane? Thank you for reading and for your help!
The correct format to check this:
Dir.exist? "#{save_path}"
Will return true if it exists or false if not.
LOL Well, I discovered my issue. I didn't realize that I had to include the entire path to each of these statements, since I'm using Dir.foreach to iterate over part of the filesystem; the updated syntax is:
Dir.foreach(save_path) do |dir|
next if dir == '.' or dir == '..'
if (File.directory?("#{save_path}#{dir}"))
begin
FileUtils.mkdir_p("./#{save_path}#{dir}/images")
rescue StandardError => e
UI.message("Failed to make directory ./#{save_path}#{dir}/images: #{e.to_s}")
return
end
end
end
This is versus File.directory?("#{dir}") in that if statement. I thought that the Ruby Dir and FileUtil libraries would look relative to the current directory for the remainder of the path.

How do you add installation related computed strings to InnoSetup scripts?

Currently when I upgrade a program I backup the existing files to a folder named backup in the program directory eg .
Source: "{app}\filename.exe"; DestDir: "{app}\backup"; Flags: external skipifsourcedoesntexist uninsneveruninstall
Is there a way to specify or compute a strings so the code is something like
Source: "{app}\filename.exe"; DestDir: "{app}\backup{date}"; Flags: external skipifsourcedoesntexist uninsneveruninstall
or combine a date with info on the previous version
Source: "{app}\filename.exe"; DestDir: "{app}\backup{previous version}{date}"; Flags: external skipifsourcedoesntexist uninsneveruninstall
Please read Pascal Scripting: Scripted Constants and Pascal Scripting: Support Functions Reference. Also read this and this.
Anyway, this is how I did it:
[Dirs]
; create an empty folder...
Name: "{app}\{code:MyDateTimeString}"
[Code]
function MyDateTimeString(Param: String): String;
begin
Result := GetDateTimeString('yyyy.mm.dd_hh.nn.ss', #0, #0);
end;
You can add runtime dynamic values using {code:...} constants.
An example of getting the current date and making a backup of the installation folder can be seen on the ISXKB wiki