Free Tools for GIS Developers
Navigation:
  What's New
 
GIS Tools:
 AVPython
 NTXShape
 SHPTRANS
 
Other Tools:
 SciTE
 Tomcat
 
About:
 The Author
 These Pages
 
 Best viewed with Mozilla / Firefox; or with IE and the Google Toolbar.
Powered by Blogger

Monday, January 12, 2004

  7:34 PM
SciTE 1.58-wbd-1 was originally compiled with Digital Mars C++, but I noticed some minor display glitches in the non-client area, so I recompiled it with MinGW and uploaded again.
0 comments [post]  

Tuesday, January 06, 2004

  7:52 PM
SciTE Lua Extension is modified to load the scripting engine just-in-time when it is first needed, rather than loading it when SciTE starts. This adds a small amount of complexity, but means that you don't pay for the overhead of Lua (which doesn't amount to much, honestly), and more importantly you don't risk the instability of new, experimental code unless you want to. In real terms what this means is a smaller footprint for my 1.58 installer (when that comes out), and easier switching between a stable work environment (with Lua turned off) and playing with the cool, new, but immature Lua scripting hooks.
0 comments [post]  

  1:23 PM
Digital Mars C++ - I used to be a user of Symantec C++; it was the first 32 bit compiler that our office purchased, back in 1995 just before Windows 95 came out. It was years ahead of the competetion, and was a breath of fresh air compared to MSVC. In terms of exception support, template support, and just in the quality of its code, it was great. However, with MSVC 5 Microsoft began to circle the wagons and reclaim the developer community. Its tools weren't better than Symantec's or Borland's, but somehow it became harder for a relatively new player to compete in Microsoft's turf. When Java came around, Symantec jumped on that bandwagon, did well there, and invested in that rather than making the major R&D investment required to get their C++ compiler ready for ANSI/ISO 1999 compliance and bring their IDDE up to the state of the art exemplified by Borland C++ Builder. But eventually the original author, Walter Bright (who also wrote Zortech C++, which was the basis for Symantic C++), got the rights back and resumed work on it. It's still lacking in a few ANSI/ISO features and its Win32 headers are rather crufty, but it is once again becoming an excellent compiler.

I've added support for Digital Mars C++ to my local copy of Scintilla and SciTE. Some of this will go into CVS; some will not. Also, the makefiles still require some Borland tools (make, brc32) at present. As long as it stands up to testing, the Lua 5 enabled custom build that I will include in my SciTE 1.58 installer will be compiled with Digital Mars.
0 comments [post]  

Monday, January 05, 2004

  7:14 AM
Tis the season to delete the holiday junk mail.

If I have filters in place that say I don't want to get any email about viagra, do the spammers really think it's all that useful to send me emails about alternatives like vagira and viarga? They've turned misspelling into an art form!

I use a Bayesian filter to cut down on my spam, and I just got my first email that probably not only gets past my filter, but messes up the statistics. It consists of hundreds of real words that are very uncommon in spam. The advertisement itself was only one line, and that line was also broken up by hidden tags:

Ban</chapman>ned C</flint>D Gov</crest>ernment d</retrofit>on't wan</depreciable>t m</windbreak>e t</religion>o s</valois>ell i</ordinal>t. Se</adroit>e N</chancellor>ow *
0 comments [post]  

Friday, January 02, 2004

  5:23 PM
My Python and Win32 Extensions setup program is updated to include Python 2.3.3 and Win32all 163. This setup program is an alternative to the non-redistributable ActivePython or the two-part install for the official ython and Win32all; it is intended to simplify redistribution for AVPython users, but it could also be of use for other Python developers. The source code for the setup wizard (an Inno Setup script, of course) is also there.
0 comments [post]  

  5:01 PM
SciTE Lua Extension now seems to be as robust on Lua 5 as it was on Lua 3.2, maybe moreso even. Also faster and more functionally complete. It works on all of the supported compilers, including Borland which was giving me the most problems before. Therefore I'm not going to be doing anything else with Lua 3.2. I'll just be working with other Lua-interested people like Jeremey Cowgar and Philippe Lhoste to prep the Lua 5 extension for inclusion in the main SciTE distribution.
0 comments [post]  


|