taner wrote: ↑Fri Jun 30, 2023 11:41 am
towards the FFAStrans headquarter.
Wait, we have an headquarter?! I wanna see it!
I thought it was just Grandmaster's basement! hahahahaha
Ok, back to serious mode.
I understand why people may wanna see 1.4.0 being released, but believe me, we're really not there yet.
The first thing that should be noted is that each release "isn't just a number", but it has a particular meaning.
If we take a look at the naming convention x.y.z (like 1.4.0) stands for "major", "feature", "bugfix".
In other words, if 1.3.0 was a feature release of the "1" series, then 1.3.1 was a bugfix release of the feature release 3.
It should be noted that we only really bump the "y" number when we're confident enough that we have enough features and that everything is stable.
This means that the development process doesn't follow a linear pattern; for instance, did you know that the development of 1.4.0 actually began after the development of 1.3.0 but BEFORE the development of 1.3.1?
This means that while the "master" repository is always updated with the latest and greatest features (it is, indeed, master), there are branches dedicated to bug fixes which make up the individual subversions.
That being said, as pretty much everyone is aware, 1.3.1 was released in the first part of 2023, but look at the date of the commits/pull requests we worked on for 1.4.0 ehehehehehehe
- Screenshot from 2023-07-01 14-57-07.png (20.89 KiB) Viewed 6738 times
- Screenshot from 2023-07-01 14-58-13.png (8.9 KiB) Viewed 6738 times
- Screenshot from 2023-07-01 14-59-06.png (52.59 KiB) Viewed 6738 times
- Screenshot from 2023-07-01 14-59-47.png (49.23 KiB) Viewed 6738 times
As you can see, commits/pull requests to 1.4.0 have been undergoing since 2022 and the development is indeed a slow meticulous process.
The whole team, Grandmaster, emcodem, Benjamin, Thomas and I is working on this and we're trying our best, but we're not there yet.
We have a few builds of 1.4.0 which have been compiled from master and I'll show you some very early screenshots for fun in the next few days, but keep in mind that those are what I call "development snapshot", so they don't represent by any stretch of meaning what the final release will be like and things might drastically change in the meantime.
Currently, 1.4.0 involved a massive redesign of the filter_builder.au3 (courtesy of a massive work done by Grandmaster Steinar) among other things and we're currently on the way of testing it out, especially in how it behaves with the current Avisynth frame properties etc.
Other changes are like ProRes XQ support in the ProRes encoder along standard writing library metadata so that you're no longer limited to Lavc (as some official Apple decoders were picky about it, but obviously we kept Lavc too for backwards compatibility), RF64 support for PCM muxed in .wav larger than 4 GB so that they're now properly muxed, improved support for images decoding (sometimes the A/V decoder would otherwise cause a loop leading to memory leak when BlanckClip() was wrongly set to length=0), high bit depth support for the A/V decoder for video and high audio sampling support for audio, IMF packages decoding support by reading the CPL, DolbyE & DolbyED2 decoding support, new options to control range and GOP in the generic MPEG encoder, new things in the REST API and plenty plenty plenty more stuff in general.
Every time we add something new, though, we gotta make sure we don't break anything and although the task of testing new functionalities and regressions might appear easily doable, believe me, it's a mammoth task.
Luckily we have test books and routine tests we can use for regressions (emcodem did a great job automatizing those for a series of meaningful tests), however we still have to spend a lot of time checking that everything is fine and even when we do that, the combinations in which users can do things is almost infinite, so we might miss something (which is why there are always bug fix releases after a while, 'cause it's you guys, the community, who actually uses it in many different ways and provides valuable feedback, far more valuable than any automatic test we might be doing on our own).
The TL;DR at this point is: yes, we're working on 1.4.0, but don't hold your breath on it 'cause it's a mammoth task.