April 2015 update of “Ini++ v1.5”

The Multimedia Fusion v2.5 extension manager now contains the April 2015 version of Ini++ v1.5. This is a small update that fixes a problem with one of the expressions.

This version is only available from the extension manager. My website no longer contains the latest version.

The description text for the latest version ends with the words “April 2015 build”, as shown below:

Extention manager


String Replace object updated

I’ve updated the String Replace extension for Multimedia Fusion in order to support a new action that clears the list of rules.

I did look at updating the extension to support Unicode and properly support the Unicode version of Multimedia Fusion (which includes the new Fusion 2.5). However, the SDK I was using is too out of date and so it would’ve been quite a lot of effort. Therefore the object does not support Unicode and also the array and global string functionality does not work in Fusion 2.5 or the Unicode version of Multimedia Fusion 2.

When the Fusion 2.5 SDK is released I may look at rewriting the object (in a compatible manner) so that it does support Unicode, and also perhaps get it to work on other platforms. I’m not sure yet, but it seems like it’d be a good way to get back into extension development for a possible future Ini++ Lite object. I have mentioned this idea in a previous post on this blog.


New version of Ini++ fixes crash

I have updated the Multimedia Fusion 2 extension Ini++. The update fixes a crash. This crash only occurred when both encryption and disk access were enabled, as far as I know. This build hopefully fixes the problem. It is referred to as the October 2013 build (sorry for it taking until mid-November to release it!)

Thank you to Logan Apple for his extremely useful help in reproducing this problem.

Update: It was reported that Build 258.2 and higher of Multimedia Fusion breaks this extension. I could not reproduce this and it appears to work fine in Fusion 2.5.

Download Link

MMX Instructions

In 1997 Intel added a set of instructions to their Pentium processor that were expected to be useful for audio and image processing. This was called MMX. The major feature it allowed was the adding of four 16-bit variables in one instruction (or alternatively eight 8-bit values or two 32-bit values).

It also features a mode where the value of the addition  ‘saturates’ rather than wraps. For instance, 100+200 = 255 using unsigned 8-bit saturation arithmetic. This is particularly useful in image and audio processing as wrapping often does not make sense in these cases. For instance, suppose we have a bitmap image where, after a header, every value is stored as 3 bytes in a row. Suppose we wish to make the image a bit lighter. It is simple to achieve this, as shown below.

// Assume 'data' is a character array holding the bitmap data, excluding
// the header. len holds the length of this array.
for ( int pos = 0 ; pos != len ; ++pos )
 int newValue = data[pos] + 50;
 if ( newValue < 0 )
   newValue = 0;
 if ( newValue > 255 )
   newValue = 255;
 data[pos] = newValue;

If I run this on my test image (which contains 844,048 pixels) it takes about 3.1ms, excluding the time to read to and write from file.

I wondered how much faster it would be using the MMX byte-wise saturated adding function. I shall be using the MSVC compiler and so MASM style inline assembly will be used. The instruction required is paddusb. The first parameter is a MMX register and the second location is either a MMX register or a memory location. The result is written to the first parameter.

So the basic idea is that we loop ‘data’, send the current 8-bytes chunk into an MMX register, call the saturated add function and then put the resulting variable back into the memory location we found it in. This gives us something like this:

unsigned char* end = data + len;
unsigned char toAdd[] = {50,50,50,50,50,50,50,50};
while ( data < end )
   __asm mov eax,data
   __asm movq MM0,[eax]
   __asm paddusb MM0,toAdd
   __asm movq [eax],MM0
  data += 8;
__asm emms // Restore state
// If data != end the last few should be adjusted without
// the use of these special instructions. I won't do this
// for demo purposes

Note that the ’emms’ instructed is required after we have finished with MMX, as otherwise floating point calculations won’t work properly (MMX and the floating point instructions use the same registers). ‘movq’ accepts a memory address as the second parameter, so we use the location that ‘data’ is pointing to. (We have to dereference it and so it must be in a register. It is not allowed to have an instruction like ‘movq MM0,[data]‘)

The same example now runs in 0.4ms – nearly 8 times faster.

Finally, we can optimise this a little. We are using ‘data’ as a counter and then keep transferring it to ‘eax’. Why not just use ‘eax’ directly? Finally, since we are not short on registers, we should make sure that ‘toAdd’ is always in a register. Let’s leave it in MM1. This gives us the following final code:

 // eax will contain the current offset
 mov eax, data
 // ebx will contain the final offset
 mov ebx, data
 add ebx, len
 // MM1 contains the constant to add
 movq MM1,toAdd
    // load in the value to MM0
    movq MM0,[eax]
    paddusb MM0,toAdd
    movq [eax],MM0
    add eax,8
    cmp eax,ebx
    jl TOP
  // Again, deal with remaining bits here
  // for real production code

It now runs in 0.3ms – over 10 times faster than the original version.

Nothing clever was done here – just some (now) standard instructions were used in the most basic and standard way. However, the increased speed was incredible and so I thought I would share it.

Update to Ini++ v1.5 and thoughts about a new extension

I have issued an update to Ini++ v1.5. It fixes a couple of memory leaks do with the undo stack. The memory leak applied even if the undo stack was disabled. For this reason it is important that every user update to the latest version.

In addition, a minor bug is fixed where the undo and redo stack size could be set to 128 layers. This would actually register as infinite. Now the properties panel will not allow any values greater than 127. If you previously choose 128 it will stay selected, so you ought to be aware of this.

There are still known bugs. I have been able to reproduce a crash with the encryption. Hopefully I’ll  be able to fix this in future, but I do not have the tools available to do so quickly. I apologise for the instability. Thank you to HitmanN and XStar for sending me examples of these problems.

The new version can be downloaded here.    It has since been updated again, see here.

In addition, I wanted to float a new idea. Many people have asked for Ini++ to be ported to other platforms. This will not happen. Ini++ is a monster of an object with many, many features. One only has to look at the file size and options to realise this. Therefore it would be a lot of effort to port to any other platform.

There is a demand for a data storage object on other platforms though, and it seems the format of Ini++ is one that appeals to people. Therefore I propose a cut-down version of Ini++ be created and ported to every platform. This is obviously a fairly big project for me to undertake, and I’m not sure yet that I want to do this. But I wanted to ask people what features are important to them.

Some of the features that Ini++ support are:

  • Compression
  • Encryption
  • Built in features to save the file in specific directories
  • Setting initial data
  • Debugging tools, including a dialog that can be opened at runtime
  • Case sensitive/insensitive options (including the ability to change it at runtime)
  • Repeated groups and items via renaming
  • The ordering of the file is preserved (but comments, etc. are not)
  • Options to auto-save or not
  • Searching features
  • Sorting and reordering features
  • Merging functions
  • Half-baked ‘subgroups’
  • Global data, including across sub-apps. (Is the latter even possible in other runtimes?)
  • Hashed values
  • Get/Set as text, CVS
  • Listing groups and items
  • ‘Perform calculation’ expressions that allows summary information to be easily extracted. Has anybody ever used that?
  • Undo and redo stacks
  • Escape characters
  • Alphanum sorting that puts, say, “Hello 12” before “Hello 3”
  • Interaction with the chart object, array object
  • Functions to get the ‘part’ of a string directly
  • Lots of actions had custom dialog boxes in MMF
  • Saving objects and global values directly
  • Ability to allow/disallow empty groups, etc.
  • ‘Current group’ set of features

Some extra features I was going to add were: Filtering of the object (so that it would changes what the object lists and so forth without actually changing the object), proper subgroups, saving to formats other than Ini files (in particular, making the whole thing work the same but write to the registry or via a server) and full preservation of white space, etc. Unicode support everywhere would be good, too. Encryption is memory was an idea too, but it is questionable how useful that is.

Which features are the most important? If people could tell me (on the Clickteam Forums) their top 5 or so features in order I would find it very useful. Remember, it would take too long to add every feature to every platform, so be conservative.

Preprocessor Programming (All 9 parts as a PDF)

The recent series of articles about Preprocessor (Meta-)Programming is now available as a PDF file. Hopefully it is easier to read like this. It is about writing code which is evaluated with the C preprocessor.


Click to view/download

The original posts have now been removed from the blog as they are all included in this one.

Chart Object now can be added to MMF HWA applications

The Chart Object I wrote for Multimedia Fusion 2 would not allow itself to be added to hardware accelerated (HWA) applications. At the time the object was released, this was a separate runtime. It was disallowed as it didn’t appear to work properly with it – nothing would be displayed (except for the pie chart mode, which writes to an internal buffer first).

Today I started work to make the Chart Object HWA compatible (although not actually take advantage of the HWA). The first thing I did was remove the tests which stopped it being added to HWA applications. I tried it, and it seemed to ‘just work’ so long as the ‘Display Mode’ setting in the application properties is not ‘Direct3D 8’ or ‘Direct3D 9’ (both of which may cause crashes).

Well, I thought this would probably do for now. I’d be interested to know if there are any people who are strongly bothered by the inability to use it with Direct3D. Perhaps if it is an actual problems I will convert it to use off-screen buffers, but I’d prefer not to.

Download the version of Chart Object without the HWA checks

Edit: A quote I found from my notes from Fimbul which – probably fairly – describes the situation:

[T]he graph object. It’s great, but it’ll probably never see HWA, java, etc. It has an excellent function that allows you to get the coordinates for the points, so if you want, you can plot a graph yourself, but you can’t use that in the other runtimes (and I think it glitches even in HWA if you try to use it there). This is another example of something that is already using a form of coding, but that instead of being “click-friendly”, it’s mostly “click-impossible”.

Alas, it was created before alternative run-times, but released at about the same time as them.

%d bloggers like this: