[squeak-dev] Scraping Squeak Wiki into 6000+ subclasses of Doclet overloads the Browser

gettimothy gettimothy at zoho.com
Sun Dec 16 10:49:02 UTC 2018


Hi folks, just an FYI and not a problem per-se. I don't know if keeping thousands of classes hanging around is a valid use-case, but I figured I would give it a try to see what happened as a first-stab at in-image persistence for the live SWiki. For SeasideDoc ( http://menmachinesmaterials.com/SeasideDoc ) under SWIKI in the left menu, I have scraped the live wiki and stored the pages as subclasses of DocletSwiki (?) The package requires Seaside and Zurb and is available at the menmachinesmaterials.com via anonymous ftp. The application works just fine. I have installed on the latest Squeak5.2 and there is no stray junk in the image per: 'http://wiki.squeak.org/squeak/2176' SystemNavigation obsoleteClasses SystemNavigation obsoleteBehaviors SystemNavigation obsoleteMethodReferences Smalltalk garbageCollect Everything works fine until I attempt to browse the superclass of those 6000+ subclasses. From a Browser, Find Class -> Doclet Then, the image, while not hanging, spikes to 100 percent on the attempt to browse. After about 5 minutes, the browser opens on the thing, but it is clearly churning on the scenario. When I close/minimize the Browser or change to a different project, performance returns to normal. I don't know if keeping thousands of classes hanging around is a valid use-case, but I figured the team would like to know about it purely as an intellectual curiosity. Cordially, t
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20181216/3f480872/attachment.html>


More information about the Squeak-dev mailing list