Outside The Box

I wrote before about corrupt text causing bogus “Missing Language Pack” warnings in AutoCAD. That case prompted me to write a file repair utility. However, since the case appeared to be isolated, I never released the utility to the public.

The problem has been rearing its ugly head again, so I decided to dust off the original code and release it. The repair utility is now available on the freebies page (look for CleanLanguage.zip) for AutoCAD 2007 and later versions of AutoCAD. I have not tested any recent cases, so I can’t be sure it is the same problem. If you try the utility, please report your results by leaving a comment here.

It’s not too often that lawyers give away their work, but Gene Quinn of IPWatchdog.com is doing just that by providing some free sample confidentiality agreements. If you can afford to pay an attorney for simple agreements like this, more power to you; otherwise you may want to grab these free agreements while you can. There is of course a disclaimer, but thankfully it’s short and simple just like the agreements.

I had the good fortune of hearing Deelip Menezes deliver the keynote address at the Bricsys 2010 Conference in person. If you missed it, check out the video now at the Bricsys web site. The question and answer session after the speech is an excellent harbinger of the discussions to come if Deelip is correct in his prediction about CAD on the cloud.

Developers, if you use VC Build Hook and/or need the same capability in Visual Studio 2010, please let Microsoft know by voting for a fix. I’ve written before (and here) about the problem. I think it is possible to hack a third party solution, but by far the best solution is for Microsoft to add back the ability to produce ANSI response files in service pack 1.

I use GlowCode for profiling ObjectARX applications. As far as I know, this is the only profiler that works with AutoCAD. What makes GlowCode special is that it can instrument code on the fly, including a monster like AutoCAD. It has a clunky interface that takes some patience to figure out, but it’s a lifesaver when you need it.

Recently a CADLock customer sent us a drawing file that was apparently not very complicated, yet it took over 20 minutes to create a deep vault. To determine the cause, I used GlowCode to profile the vault creation process. I could see almost immediately that most of the time was being spent in AcDbObjectId::operator ==, which was being called from AcDbObjectIdArray::find.

This turned out to be a classic case of using the wrong type of container. Searching an AcDbObjectIdArray isn’t bad when the array is small, but this drawing contained a special case where the array was very large, and this special case exposed the flaw. I changed the array to a std::set and tested again. The new time was 30 seconds to create the vault – a slight reduction of approximately 95%.

[Full Disclosure: Electric Software has for many years provided me with a complimentary license of GlowCode in exchange for my testing and feedback.]