VMWare Fusion: Mounting a network drive for reliable cross-platform development with LiveCode

Introduction

When I’m developing desktop applications in LiveCode I need to test on both macOS and Windows. My primary development platform is macOS and I run Windows using VMWare Fusion. Since I use the Levure framework I can switch over to VMWare at any time, launch a pre-compiled executable, and quickly see how my app behaves. The precompiled executable loads the current development files so I’m testing what I’ve been working on in the LiveCode IDE on macOS.

Up until yesterday there was a problem, however. If I find a bug while testing in Windows I like to switch back to macOS, make the necessary code changes, and then launch the pre-compiled executable to test again in Windows. But sometimes when I launched the executable I would get an error because the source file I just edited (e.g. a .livecodescript file) would be corrupted. For some reason, Windows running under VMWare Fusion was only seeing a truncated version of the file I had just edited. If I opened the file in NotePad I would see that some of the file contents at the end of the file would be missing. The only way to fix the issue was to restart Windows which really slows things down when you are trying to troubleshoot a bug. (Watch a screencast demonstrating my workflow.)

The interesting thing is that this had not always been a problem. In the past I had used the same workflow with VMWare Fusion and had no problems.

Yesterday I upgraded to VMWare Fusion 10 hoping that the problem might be fixed. When I saw that it wasn’t I decided to email VMWare support to see if they knew what was going on. I was pleasantly surprised to get a call back from their technical support team in no time at all. We got on a screen sharing session, I showed them what was going on, a knowledgable person then got on the phone with me and talked me through the solution which I will describe below.

Don’t use Shared Folders for development work

At some point I had set up VMWare Fusion to use Shared Folders. This is an easy way of making folders on macOS available to Windows. While it is convenient, the feature wasn’t designed to be used the way I’m using it for development work and the issue I described above can occur.

Mount a network drive

The solution is to mount your macOS folder as a network drive. This involves a little more work but is simple enough if you know what to do. I’m going to show you how to create a private network between VMWare and your Mac so that you don’t have to worry about which network you are on.

Add Device…

  1. To begin, turn off your VM and open the settings.
  2. Click on the Add Device… button.

Add a Network Adaptor

Select Private to my Mac

  1. Click on the Private to my Mac option
  2. Click Show All

You should now see two network adaptors in the Settings window.

Turn on File Sharing on your Mac

  1. Navigate to the System Preferences > Sharing panel and turn on File Sharing.
  2. Make note of the smb address.
  3. Click on the Options… button.

Turn on SMB file sharing for your account

In order to mount your folder as a network drive in VMWare you will need to check the box next to the account you want to share.

Map network drive

You are now ready to launch Windows in VMWare Fusion and map a network drive.

  1. Select This PC in the left column of a Windows Explorer window.
  2. Click on Map network drive and select Map network drive.

Enter the path to your macOS user folder

Using the smb address you noted earlier, enter the path to your user folder on macOS. In this example I used the following:

\\trevors-imac\trevordevore

Enter your macOS login and password

Voila!

Your macOS folder will now appear as a networked drive under Windows. You can drag any folders from the networked drive to the Quick Access area so that you can quickly navigate to the applications you want to test. Using this approach you can edit files in macOS and know that the file contents will be updated properly within Windows running in VMWare Fusion. Happy testing!

Share

Exporting SVG files from Adobe Illustrator for use with drawingSvgCompileFile()

In LiveCode 9 the image object has been updated with support for resolution independent vector data. In addition, a new Drawing Library is included which has two functions that will convert an SVG file or SVG data to a format that can be assigned to the text of an image. The end result is that developers can now import SVG files into the LiveCode image object. This is a great way to include resolution independent graphics in LiveCode applications.

Given that LiveCode 9.0 does not support the entire SVG specification it is possible to create SVG files that will not import into LiveCode properly. This article will show you how to export an SVG file from Adobe Illustrator that will display properly in LiveCode.

Continue reading

Share

Using the livecode standalone engine to run script only stacks on a server

Years ago I set up a license key generator on our server to generate license keys for our products. I used a revolution standalone engine and a plain text LiveCode script that was processed using the engine. Recently, a server upgrade removed a file that was necessary for the old revolution engine to function so I was forced to update my setup. Below is a description of the solution I came up with after a few hours of trial and error.

LiveCode Community Standalone Engine

To process my plain text LiveCode script files I installed the 64-bit version of the community version of the Linux standalone engine on my server. I grabbed the file named Standalone from the Tools/Runtime/Linux/x86-64 folder of a LiveCode Community installation. I renamed the file to <code> on my server and made it executable (e.g. chmod 755).

Note that you need to use the community version as it doesn’t need to be licensed. If you grab the Standalone engine from an Indy or Business installation the engine won’t run because it is unlicensed. Licensing occurs when you build a standalone in LiveCode.

The Script

For the script I created a Script Only Stack which looks like this:

script "clarifykeygenerator"

on startup
   put generateKey(the commandArguments)
   quit
end startup

function generateKey pParams
   ...
   return tKey
end generateKey

The script generates the key, outputs it to STDOUT, and then quits. That’s it.

You can now test this setup on the command line. Assuming the lc-standalone-v8 engine and a keygenerator.livecode script are in the same directory the following call would work:

./lc-standalone-v8 -ui keygenerator.livecode theLicenseCount theLicenseName

Note that when processing parameters in your script, -ui and keygenerator.livecode would be the first two parameters. Actual parameters that your script is interested in processing would start with parameter number 3 in the commandArguments.

Calling The License Generator From PHP

The final thing I had to do was call the license generator from PHP. You can do that using the exec() function. Below is some sample code. Note that I am base64 encoding the name. If the name for the license has accented characters in it then PHP will throw an error if you try to include it in the exec() function. Just make sure you call base64decode in your LiveCode script that generates the license key.

$license_generator = '/usr/home/name/public_html/cgi-bin/';

$execStr = $license_generator . 'lc-standalone-v8 -ui ' . $license_generator .         'keygenerator.livecode ' . 
        escapeshellarg($license_count) . ' ' . 
        escapeshellarg(base64_encode($name));

$key = exec($execStr, $output, $return_var);
Share

Accounting for new widget properties in an OnLoad handler

When writing a LIveCode widget the OnLoad handler is used to populate variables in the current instance of the widget with values that were stored with the widget. Version 1 of an OnLoad handler looks something like this:

public handler OnLoad(in pProperties as Array)
   put pProperties["text"] into mText
   put stringToColor(pProperties["text color"]) into mTextColor
   put stringToColor(pProperties["button color"]) into mButtonColor
end handler

Now assume that you released this widget and there are stacks using it. You then decide that you want to add some more properties to the widget. You make the necessary changes to your OnLoad handler so that version 1.0.1 looks like this:

public handler OnLoad(in pProperties as Array)
   put pProperties["text"] into mText
   put stringToColor(pProperties["text color"]) into mTextColor
   put stringToColor(pProperties["button color"]) into mButtonColor

   // New properties added in version 1.0.1
   put stringToColor(pProperties["border color"]) into mBorderColor
   put pProperties["stroke width"] into mStrokeWidth
end handler

You compile the widget and then launch your project. When your existing widgets are displayed on the screen they aren’t rendered. All you see is an empty area. You try to check the properties in the property inspector but no properties are set. What is going on?

Since your widgets created with version 1 of your widget they didn’t store any properties named “border color” or “stroke width”. When you try to access a key of the pProperties array that doesn’t exist LCB is unhappy and refuses to go on.

There is an easy fix, however. Whenever you release a new version of your widget with new properties you need to add a check to make sure the new properties are keys in the pProperties array.

public handler OnLoad(in pProperties as Array)
   put pProperties["text"] into mText
   put stringToColor(pProperties["text color"]) into mTextColor
   put stringToColor(pProperties["button color"]) into mButtonColor

   // New properties added in version 1.0.1
   if "border color" is among the keys of pProperties then
      put stringToColor(pProperties["border color"]) into mBorderColor
      put pProperties["stroke width"] into mStrokeWidth
   end if
end handler

Since I added “border color” and “stroke width” in version 1.0.1 I just check for the presence of one of the keys. If one of the keys is present then the other will be present as well.

With this change to the code, your existing widgets will still load when running with version 1.0.1 of your widget.

Share

Creating a Busy Indicator in LiveCode Builder

One of my favorite aspects of creating Widgets with LiveCode Builder (LCB) is that I don’t have to create multiple image assets in order to support various display resolutions. Since it is possible to create very complex shapes using the new drawing routines I no longer have to resort to using PNG images exported from Photoshop for more complicated aspects of my UI.

Busy indicators are one control in particular that are quite a pain to manage. You needed to generate a PNG image for each frame in the animation and you had to create the animation for at least two sizes (1x and 2x). With widgets creating a busy indicator is much simpler and you get better results. Here are a couple of notes from a busy indicator that I just finished working on. You can get the source from my livecode-extensions github repository.

Continue reading

Share

Dawn of the Planet of the Widgets

On March 12th Kevin Miller announced the first LiveCode 8 developer preview at the LiveCode Unconference in Brooklyn. At the heart of the announcement was the introduction of widgets, a new control available to LiveCode developers. Thanks to the LiveCode Github respository, I had been using LiveCode 8 in the weeks leading up to the announcement. As LiveCode engineers made improvements I would update my local repository, rebuild using Xcode, and test. I was so impressed with what I saw that I decided to move a project I’ve been working on for the last few months to version 8 right away. Why would I trust my project to a LiveCode engine that hadn’t even reached a public beta yet? Let’s take a look.

Continue reading

Share

Using SourceTree to monitor progress on LiveCode

When LiveCode became open source after the Kickstarter campaign all of the source code was moved to Github. Because the livecode repository on Github is public, you can monitor all changes which are submitted to it.

I’m going to show you how to easily monitor progress on the LiveCode engine using SourceTree, a free Git client for Mac and Windows made by Atlassian.

Continue reading

Share

A Plugin For Monitoring the focusedObject

Knowing which object currently has focus can help you troubleshoot situations when messages aren’t being sent to the object you think should be getting them. I have a plugin I use called ReportTheFocusedObject which continually monitors the focusedObject and selectedField so that you can quickly determine which object is the target of messages.

You can download the plugin here.

For instructions on how to install plugins for Revolution see this lesson.

Share