Thanks for your update, xlucas. It's good to know this is a formally studied type of sequence, and there's at least one method of reliably generating them. I made a version based on the sample Python code on the Wikipedia page. The alphabet and the subsequence length are adjustable as constants: con...

Thanks, this is looking better. It occurs to me that full Circles are drawn differently from partial ones - the former uses the Bresenham algorithm, while the latter use an unsophisticated Sin/Cos based approach. (It would be great to fix this some time..) But as far as I know, full circles shouldn'...

Thanks fxm. Now I look back at my words, probably it's enough to say "not recommended", rather than "strongly advised against". I think the recommendation should go on Paint's page only - my personal feeling is that it's not needed to talk specifically about Paint on the Line/Cir...

The documentation notes should be specific that the problem occurs in conjunction with GFX_ALPHA_PRIMITIVES. It would be better to phrase the problem as overdraw, chaing the resulting colour of affected pixels. It should try not to make any implication that it will behave in these ways, only that th...

Thanks. I just generated a quick greyscale palette with for i as integer = 0 to 255: palette i, i,i,i: next i It looks like your Screen 18 is set to go fullscreen, that might make a difference. It would probably be a good idea to go fullscreen on both or neither, to confirm whether that's the proble...

I don't know.. On the one hand, the cause for this problem was not obvious, and baffled multiple people. On the other hand, it feels like a lot of text just to say, in effect, "warning: blended borders will be blended". Really, the problem here is that boxes are drawn twice at the corners,...

using opaqueness: It only takes one pixel overlap in a perimeter to leak, because this pixel will be a different colour Using line()-(),b overlaps at each corner. Nice discovery, dodicat! I've put a note about this in the bug report: 896 Unfortunately fixing the box drawing to prevent this double-b...

When drawing a line or box in QB, an optional 'style' parameter can be used. This contains a 16-bit repeating pattern, that controls which pixels of the line are set. A box is made up of up to four lines. With no style, it can only be drawn one way (just a solid border), but when there is a style, i...

Isn't a bug something, which should draw "topmost attention"? (aka: has top priority) I'm aware that this is, to a certain extent, depending on who reports it, because many don't check their code, for syntactical correctness, etc. causing ' false positives '. (not enough research, before ...

I'd agree with you that capitalisation is sensible and correct to use if you're using an acronym, keyword, constant, etc. that is supposed to be capitalised. But unnecessary emphasis can alter the tone of a post, particularly when its purpose is to draw attention to a fault. Regarding the example co...

Thanks for the report. I'd prefer though if you didn't capitalise the word "bug". It just comes across as shouting. Your example code has a subtle problem: the red border changes colour if it's partially transparent, so isn't actually equal to the border colour. But if I fix that - using t...

Yeah, in most cases, I expect people will not want to mix formatted and unformatted output in a single Print expression, but the possibility is there if anyone wants to do that. In practice, maybe it's not a feature worth having, since makes the documentation a little more complex, and possibly mean...

You have dkl from the FB development team to thank for fbfrog. *str++ = *dest++ better like that ( str=dest ; dest+=1; str+=1;) Don't forget to dereference: '*str = *dst' A couple more thoughts occur to me: - (a,b) can be translated to FB (if you really have to!) as 'iif(a,b,b)', as long as 'a' retu...

I had another go at translating. Here is the code with indentation: if (! ( ( bits= bits&0x7fffffff? ( resbits=bits, bits+bits ) : ( src+=4, resbits=*((Ulong *)(src-4)), (resbits<<1)+1 ) ), resbits>>31 ) ) { goto uselastofs; } That means it can be translated to something like: if (bits & 0x7...

What horrible code! Unfortunately it won't translate directly to a single FB expression. Maybe the first thing to do would be to translate it several separate lines of C, which can then be ported.. - (b=c) is an expression that assigns c to b, then returns b. - (a,b) evaluates a, then evaluates b, t...