hyperion

Mh, probably Ubuntu 22.04. python-is-python3 sounds like a package installing a symlink python pointing to python3 and the the package python-is-python3 just so depends on a package python3 which in turn depends on a package python3.10 for example. My guess the package python3 if it exists would be what you want. Not sure what the canonical way to install python 3 on Ubuntu is but looking at the results of "apt search python" should probably make it obvious.

Is correct but might be incomplete. Edit: tested to be complete, builds with 2.11.5, 2.12.1, 2.12.2 and git HEAD. Don't think this improves readability over your first suggestion. It's not complexity that is the issue just that people aren't used to it in this context so it takes an extra one or two seconds to mentally parse, at least for me. The greater than char in the greater than equal token doesn't help Whether or not the tradeoff vs. not using the preprocessor is worth it I don't know.

As long as it's not the cast variant of solving this I'd be fine with either. Looking at the linked ticket a future 2.12.3 should no longer have any issue, still technically it's a downstream bug, just that the fallout was unexpectedly large and the missing headers where injected back upstream. So I'd add the parser header regardless as this also allows the use of all 2.12.x releases.

Well, some of the changes you posted look valid at a glance, not so the use of an internal type of libxml2 tho.

A more recent version than mine, looking at my distros bugracker there are lots of similar bugs to this one for various packages with 2.12 and so it's still hard masked.

Obviously this is just an opinion due to lack of a definitive metric

What version of libxml2 are you using and is it an install from your package manager and with no local build in path?

The gameplay is sufficiently unique. A bug in the placement function I'd say, same as chickens or berries in mines or trees in iber walls. <- @maroder

https://github.com/wraitii/0ad_grand_strat

What needs fixing the most is not being able to register on phab.

I suggest you play Polar Sea map once.

Well, wasn't there an olympic discipline of running with hoplite gear?

Trees, units, buildings from an art perspective are actors, you find these xml files under art/actors. There is an introduction to actors on the wiki. The schema is named actor.rng and can be found in the source tree.

https://www.khronos.org/files/collada_spec_1_5.pdf But I suspect Blender has the required code, it's just not wired up for the colada exporter. Did you try fbx for example, did it work?

Well, me thinks the biggest issue with walls is placing and repairing

Some blender exporters have a scale factor, might try exporting and re-importing using one of those formats as a quick hack. Never tried myself so might not work.

Alpha channel of spec map can be used for emission with corresponding material definition, eg see basic_glow_norm_spec.xml Another option would be to grab a Kronos ref implementation of PBR shader and modify it to work with pyrogenesis, Vladislav has plans to move to PBR at some point, so you'd be somewhat future proof and have a probably better time using modern tools.

Finite resources are somewhat important for rts, so me think it's up to mapmakers to add more fish (more or bigger patches) if that improves a certain map. One problem is the role of fishing isn't that obvious in 0ad as let's say aoe2, where its used to grow eco faster early on and still not lose all purpose later with fish traps by converting wood to food over time in balance with fields.

Well, for Linux only sources are provided and distributions have to build 0ad themself. Also they wouldn't accept bundled deps and so the situation is different than for Mac where the goal is an all-in-one dmg. If you build for yourself fixing the build script to use homebrew packages is obviously possible. At least you get a working build script you can use as base unlike on Windows where the deps are prebuilt.

All data files are OS-independent so you can just copy them over as long as you preserve the structure, for where to find them, there is the helpful page https://trac.wildfiregames.com/wiki/GameDataPaths

Well, guess you figured out at least 90%, so if you add that and then post in the forum asking for review and pose a question or two, you likely get proper answers for the reminder if you really can't figure it out. I don't intend to target you but there is a rather common pattern here, complain and expect someone else to sit down and do research or do work. So it comes down to if you can't be bothered why should I be. While not absolute, I'm far more willing to help people who actually contribute. As for brilliant ideas, I'm sure all contributors have lots of them, so why would they need others to provide more.

Server side, maybe to much load, try again after a while.

To my knowledge those tarballs are exports of gecko-dev.git and are in essence stripped tarballs of firefox source releases for size reason, the following should work as drop-in. https://ftp.mozilla.org/pub/firefox/releases/102.12.0esr/source/

@Ben Phegan SM-102 just became EOL, meaning it won't change anymore and can be used without it changing right under our nose. So there is no fundamental blocker anymore, ofc needs more testing and maybe fixing. So your findings might be of some help pushing it forward. As for CI, the built binaries they should work on later versions anyway. The issue with newer SDK is it causes issues with opengl, so the new vulkan support is probably the way forward. As for the other build issues, well, not that nice is all I can say, let's hope those patches can get merged sooner rather than later.

The CI uses 10.12 IIRC, that's also the one used for building the releases. Anything newer should work or it's a serious bug. As for SM you might want to try 102, maybe you'll have more success https://code.wildfiregames.com/D5002