Skip to: Site menu | Main content

Development Log October 2007

Newest Entries are on the Recent Entries page.

Wednesday, 31st October 2007

Trying to determine why the swf output isn't displaying properly. It's very close to correct from what I can tell by using Gnash (-vp) and flasm (-d) But I'm having trouble finding something that goes into more detail. Have resorted to creating the same output with Ming (nearly finished), to compare the end result. Only a few bytes different. Might have to take some time on the weekend to resolve those differences and find out why.


Tuesday, 30th October 2007

Having problems getting the swf matrix structure writing out correctly. Looking at how Ming does it, they have structures and functions in place which take care of this automatically. I think I'll copy their approach.


Sunday, 28th October 2007

Took quite some time today to look into various AJAX technologies, and especially the Google Web Toolkit, for awareness of how to build the online component of Flame. Looks promising. :)


Saturday, 27th October 2007

Worked on the swf output for background images. Fixed a bug where the fill type was being incorrectly encoded, and working on getting the output swf for the matrix component correct. Progress being made, but still needs work.


Monday, 22nd October 2007

Didn't have much spare time today, so only got around to updating the Makefile to compile the recently added functions in CVS. Still a lot of swf work to be done.


Sunday, 21st October 2007

Added initial code to generate shape records for background images for the swf output. It's in cvs, but it's untested and will likely need work.


Thursday, 18th October 2007

Figuring out how fill style records, and shape records, are structured in swf. Think I've got the fill style record bit worked out, but still need to better understand how shape records work. It's pretty fiddly. ;-> As an aside, noticed I started working on the swf 11 days ago. So, it's taking time. At this rate, it'll probably be another 11 days or more before things are all working together nicely.


Wednesday, 17th October 2007

Thinking the reason the images aren't being displayed in the swf output, is because they need to have a shape defined for them to be "filled" into. Started adding code for defining a background shape for them in the output swf. It's in CVS, but not usable yet, as I'm still getting my head around what bits go where.


Monday, 15th October 2007

Cleaned up the byte ordering in the swf output functions. Added a function to set the background color of the swf movie. The swf movies being generated aren't yet working, but finding it difficult to tell what's wrong. Flasm has been helpful up until now, but it's not giving me a clear picture any more. Hopefully the only problem is that the bitmap images I'm embedding need to be resized first.


Thursday, 11th October 2007

Figured out how to select individual bytes inside words and double words, and write them to the output steam in the order I want. So, now the swf header has the length of the file, number of frames, and the frame rate all written out correctly. Still need to get the dimensions of the swf movie written out correctly though.


Wednesday, 10th October 2007

Started debugging through the swf output code. Fixed a crashing bug due to the use of an incorrect variable, came up to another bug at the point of finally outputting the file from the buffer. Guess that's what I'll look at tomorrow? :)


Sunday, 7th October 2007

Spent many hours today adding functions and code for generating swf. Doesn't yet work, and pretty sure the byte ordering has serious issues, but as a generalisation it's getting reasonably close to having a structure I can debug through. Which in turn should let things be changed and adjusted until it's outputting valid swf bytecode.


Saturday, 6th October 2007

Finished updating the includes in each function to only include the others that it depends upon. This makes recompiling (i.e. for developers) a lot faster, as a single change doesn't mean everything needs to be recompiled.


Thursday, 4th October 2007

Added initial (non working) stub functions for the creation of flash byte code for highlight, mouse cursor, and text layers. Also began adjusting the includes in each function to only call the ones it needs, in order to dramatically speed up recompilation time for developers.


Wednesday, 3rd October 2007

Noticing that the Ohloh site has some useful features, such as having a sorted (most recent first) list of CVS commits:

 http://www.ohloh.net/projects/8895/commits 

Also added initial untested code to convert Flame bitmap layers to Flash dictionary shapes. Lots more work to be done.


Tuesday, 2nd October 2007

Came across an interesting OSS project evaluation site today called Ohloh, so added an entry for Flame to it:

 Flame Project page on Ohlog 

Also filled in a bunch of the missing info for PostgreSQL, and sent invites to the relevant people I know who it saw the CVS commits of but that it didn't have further details for.


Monday, 1st October 2007

Started into the coding of Flash (swf) output. Still very tentative steps, with only one partial stub function written thus far. It's going to take some time until its working.