wildweasel wrote:My trick to prevent this is to lower the "height" setting on the status bar to one pixel less than the actual pixel height of the bar image. So in Doom's case, "height" would be set to 31 instead of 32. Barely perceptible difference to most players, and the bar gets drawn over that one line of imprecision that occurs at certain screen resolutions.
True, but the side effect is that the map name displayed on the automap is also lowered by 1 line. Looks not great either.
MG_Man wrote:Perhaps, but it seems like a fairly inexpensive operation. I notice no performance differences. It would be true, though, that a 1px border within the status bar would be much less taxing.
From a programmer's point of view I'd also say that simply including the "completeborder true;" statement is the more robust way of fixing the problem, rather than programming something to somehow "work around" the problem where it is not even clear if it would not introduce problems with custom status bars. I also doubt it makes much difference in regards to CPU cycles.
Plus, the fix does not require any work on the programmers' sides. It is simply including one line in the sbarinfo definitions.
So my suggestion is also that "completeborder true;" should be used as a fix for this problem.
[quote="wildweasel"]My trick to prevent this is to lower the "height" setting on the status bar to one pixel less than the actual pixel height of the bar image. So in Doom's case, "height" would be set to 31 instead of 32. Barely perceptible difference to most players, and the bar gets drawn over that one line of imprecision that occurs at certain screen resolutions.[/quote]
True, but the side effect is that the map name displayed on the automap is also lowered by 1 line. Looks not great either.
[quote="MG_Man"]Perhaps, but it seems like a fairly inexpensive operation. I notice no performance differences. It would be true, though, that a 1px border within the status bar would be much less taxing.[/quote]
From a programmer's point of view I'd also say that simply including the "completeborder true;" statement is the more robust way of fixing the problem, rather than programming something to somehow "work around" the problem where it is not even clear if it would not introduce problems with custom status bars. I also doubt it makes much difference in regards to CPU cycles.
Plus, the fix does not require any work on the programmers' sides. It is simply including one line in the sbarinfo definitions.
So my suggestion is also that "completeborder true;" should be used as a fix for this problem.