Oleg Oshmyan

Could this be caused by the fact an actual Fourier transform with floating-point maths is used? As far as I understand (which may be wrong), it can introduce errors by...

I’ve already pasted the full number (in the linked gist). It’s only in base 36 to make it shorter than it would be in base 10. (I couldn’t even get...

To be honest I still don’t understand what this gives you that I haven’t already provided, but here: ```javascript xhr = new XMLHttpRequest() xhr.open('GET', 'https://gist.githubusercontent.com/astiob/091b8dbadf0698affb7a469e93450a0b/raw/2f13cc26872fd0d60787f23a2614bdd21156b730/base.base36') xhr.onload = function () {...

Bisecting shows 925c04a37f194ca992f5e5a9ee98ffc1b597dbaa as the first bad commit.

> Is the image immediatelly on the regressed commit the same as in HEAD of master (bitstream and/or decoded png)? Yes. Some intermediate commits during bisection showed other glitches, but...

Just tried encoding to AVIF with libheif and decoding with libavif to exclude libheif’s decoder from the equation: makes no difference; the round trip is still lossy. Curiously, `avifdec` reports:...

If you’re fixing the defaults—if the defaults in that code are even relevant in the first place—then shouldn’t `ScaledBorderAndShadow` also be changed? The context lines in your patch suggest it...

Yep. Assuming I understand correctly that this object mimics the standard Python dictionary, the second parameter provides a default value and itself defaults to `None` (Python’s name for `null`). The...

I have little knowledge of Kodi code, but a search for `PrimaryColour` combined with the report that this is a regression in 20a3 points to https://github.com/xbmc/xbmc/commit/ca1f335e1d382ca4f20621a1c5047b9588226672 as a primary suspect,...

> If I'm not mixing things up, writing such a script also isn't straightforward and no complete script (yet) exists. Indeed. No such tool exists yet, and when I worked...