On Tue, Jun 26, 2007 at 03:31:32PM +0200, Vincent Lefevre wrote: > On 2007-06-26 09:24:44 -0400, Thomas Dickey wrote: > > On Tue, Jun 26, 2007 at 02:58:31PM +0200, Vincent Lefevre wrote: > > > A naive replay won't work: you need (triple-)clicks while the output > > > occurs to be able to reproduce the bug (such clicks more or less > > > freeze xterm's output). > > > > I would take that into account. The main thing that a reply would > > do would be to set the screen up as it was when you clicked on it, > > e.g., erasures, wrapping-marks, etc. > > But clicks are not logged, so it would not be the same state (clicks > shouldn't affect the buffer state, but the present bug is that they > seem to affect this state).
hmm. Going back through the thread, it seems you're talking about selecting text while it's moving. All I'm able to see is a case that I can explain: if the text doesn't move, and I happen to be clicking near the place where two different lines are displayed, I can get a selection-extension at the same time that I'm (nominally) doing a triple click. Slowly cat'ing the text in a loop gives the same effect - occasionally the clicking results in a selection-extension. Outside of that - I'm not seeing any instance where it selects past the current line. From your report, I cannot tell if there's a particular place on the screen where problem occurs. For instance, selecting near the bottom, or close to one side might be relevant. -- Thomas E. Dickey http://invisible-island.net ftp://invisible-island.net
pgpa6G5fj15i2.pgp
Description: PGP signature