There is an odd bug in mirc 6.20/6.21 that I can't seem to be able to accurately reproduce, so instead a screenshot:
http://www.nihongoresources.com/downloads/mircbug.jpgThe odd part is that this bug pops up every now and then, and seems completely unreproducible. I've tried to test it in a testchannel with the same number of chracters in the name so that the line mirc gets fed is of equal length with just a different channel name, but this does not lead to the same kind of screw up (that is to say, consistently). I've also tried to reset the topic in the original channel a second time, which does not have the same effect, instead leading to the last bit of the topic being cut off a byte "too early" so the last bit of UTF8 encoded test becomes the classical utf garble (but then this is not mirc's fault but more the cutoff at the server being set in bytes).
The curious part is that I do believe this is somehow a deep mirc bug, as hitting the "up" button to go through the input history for the line I sent off to the server shows me a magically distorted line that matches what is shown in the channel; it seems mirc pre-processed the line, borked it, then stored it in the history and fired it off to the server.
If anyone has an idea of what to test to see if a consistent reproduction of the problem can be set up, I'm all eyes.
- Pomax
