[Ipe-discuss] Antialiasing, delete

[Dmitriy Morozov]

> Dmitriy Morozov wrote:
>> While we are asking questions "Why..." before filing bug reports, I was
>> wondering why anti-aliasing is not turned on? I'm talking about
>> anti-aliasing for lines of course, not fonts. Qt4 supports it; it takes
>> exactly 3 lines (2 in ipecanvas.cpp, 1 in ipecanvaspainter.cpp) to add
>> it, and the result is so much easier on the eyes.
>
> I've just tried it out, it does look very nice.  Thanks for pointing me
> to this feature (it's not very visible in the Qt documentation).

You are welcome. For me personally it's the main reason I've been
anticipating the Qt4 port, so I was very surprised to see it disabled.
There are some tiny bugs when it's on (maybe actually when it's off too, I
haven't really used Ipe Pre 25 without antialiasing), e.g., the red cross
cursor is not erased properly in some cases, so there are some trails of
dots left around occasionally (although rarely). I'll try to file it as a
proper bug report once I figure out the exact use case when it happens.

>> On a related note, why is
>> double-buffering explicitly turned off (with
>> setAttribute(Qt::WA_PaintOnScreen);). It's not really noticeable
>> without anti-aliasing, but I still don't see a reason to turn it off.
>
> Have you tried removing the line?  It worked fine for me under Windows,
> but under Linux become unusably slow.  I didn't have time to figure out
> whether (a) I'm doing something wrong or (b) Qt is still buggy or (c) my
>   X server is not configured right...  I'd be nice if someone could
> figure this out for me :-)
>
> Note that I don't expect it to make a big change: Ipe already draws all
> objects onto a bitmap, and only draws the overlays directly onto the
> screen.  I couldn't see any difference with or without double buffering
> under Windows.

I have removed this line because the lines were flickering noticeably with
it (not dramatically, but noticeably) when antialiasing was on. In light
of what you say about Ipe's implicit off-screen drawing, I'm not sure why
this happens. I use only Linux, and it has not become slow (i.e., I saw no
change in speed). However, it would be interesting to have other users
check its performance too.

>> Also, can somebody confirm that the following is a bug, and not just a
>> problem with my installation before I file it in Bugzilla?
>
> Yes, it's a bug (and has already been filed).

Well, by me, so it wasn't already when I asked. ;-)