On Mon, 18 May 2026 06:57:44 GMT, Lukasz Kostyra <[email protected]> wrote:
> The changes compile correctly and everything seems to work fine. I do have a > question though. > > I have a general concern relating future gstreamer updates. Inevitably, we'll > have to update it on our side anyway, and this seems to add more complexity > to the upgrades. Outside of removing code we don't need, what do we gain from > doing this? Main benefits: - Binary size reduction, since UNICODE tables takes space. Example: Windows: glib-lite.dll 674.50 KiB -> 359.50 KiB. MacOS: 1,234,000 bytes -> 1,040,896 bytes. - GSTREAMER_LITE define does not add complexity too much when merging code with upstream. Removed files will actually reduce effort to merge with upstream, since less code to merge. - Eventually I would prefer to remove as much unused code as possible. I think we still have plenty of unused code in GStreamer/GLib. ------------- PR Comment: https://git.openjdk.org/jfx/pull/2168#issuecomment-4483015328
