I have recently encountered problems with distilling some GSAK output HTML into a Plucker PDB file. The distiller (HTML Parser) hangs, and it is impossible to get any indication of which page is causing the problem.

I switched to Sunrise, an alternative to the Plucker Desktop. This software seems to work pretty well, although:

a) It cannot distill as large a set of cache pages as Plucker. If there are too many pages or images it runs out of memory.

b) It is written in Java and requires download of the Java runtime from java.sun.com.

I looked at sunrise a while back but the author stated

quote:

In future, Sunrise will come with its own offline web viewer, which will be a commercial product. Once the viewer reaches public beta state, Plucker support will be removed entirely.

. But now Sunrise website is down. Also the google's cache copy, says

quote:

this site will close down in two weeks. (October 16, to be precise. With the closedown of this site, Sunrise maintenance will come to an end. It's unlikely there will be many releases after this.

.

Note I just had GSAK output HTML for all the geocache in north GA (above N32 45.667) and then plucked them it processed all 1,727 geocaches without a problem. It was about 11 megs, but it worked.

It does appear that Sunrise is a "dead end", as the author is moving to develop a C++ version for PocketPC and vade mecum viewer, abandoing the Palm/Plucker space.

As an experiment, I created a GSAK database with 1329 caches and exported it to HTML. Plucker/distiller hangs on this output. I then split it into two equal subsets, and Plucker handled both halves OK. It thus appears that the issue is not an individual page's HTML, and is likely a memory usage problem in the python distiller. Sunrise was able to distill the 1329 caches.

Since I do not included images from cache pages, my PDB file is 3.8Meg.

In any case, creating multiple PDB files for different areas seems to be a reasonable workaround.

Yesterday I noticed that Sunrise is now open-source on SourceForge.net. I downloaded the source code and saw that the memory usage for the JVM is controlled by a registry value (if not set defaults to 64Meg). I created the registry value and set it to 256Meg. Afterwards I was able to distill 2600 cache pages into a single PDB.

Really... Well, in that case I may see about getting involved and see if something can be done about the performance issues.