From: Alan G. <ag...@sp...> - 2009-04-23 14:04:56
|
Thanks to the brilliant refactoring of ecnode, I finally understand why ktechlab confuses wires with busses. It starts out with a network of connectors, and then goes to a network of connectors that carry some electrical signals, to a network of connectors that carries one electrical signal. When in truth, the case where there is only one electrical connector is NOT a special case of the case where there are many. The design choice to organize the code that way anyway is causing many headaches where container classes are used in place of simple pointers and the business about ktechlab allowing you to connect a bus to a single wire with unpredictable results. -- New president: Here we go again... Chemistry.com: A total rip-off. Powers are not rights. |
From: P Z. <zol...@gm...> - 2009-04-23 15:22:11
|
On Thu, 23 Apr 2009 14:59:02 +0200, Alan Grimes <ag...@sp...> wrote: > Thanks to the brilliant refactoring of ecnode, I finally understand why > ktechlab confuses wires with busses. > > It starts out with a network of connectors, and then goes to a network > of connectors that carry some electrical signals, to a network of > connectors that carries one electrical signal. When in truth, the case > where there is only one electrical connector is NOT a special case of > the case where there are many. The design choice to organize the code > that way anyway is causing many headaches where container classes are > used in place of simple pointers and the business about ktechlab > allowing you to connect a bus to a single wire with unpredictable > results. I noticed that bug some time ago, and listed it here: http://ktechlab.org/wiki/index.php?title=User:Zolee#Known_bugs_.28known_by_me_at_least.29 The result is not unpredictable: the first wire of the bus is connected and all the others are floating. We should centralize all the known bugs in the bugtracker... |