Since it appeared to be the most acutely developed version, I compiled it and tested it. But it crashes on start. Is this because it is not in a working condition?
The branch is in working order, but there is an incompatibility with the settings .xml file. If you delete settings-experimental.xml (or settings-experimental-debug.xml if you are compiling a debug version), it should work.
Oh, gosh, that was incredibly daft of me. I'm so stupid. That has happened before so I should have remembered to test that. It now works (generally), though it crashes on generation unless industries are disabled, but this I think is due to something with the pak somehow. How would I go compile a double-height enabled pak.britain ex, by the way? I already compile from source, but is the latest experimental makeobj compatible or would I have to compile that too from source and use that? (assuming way improvements branch includes double-height in the first place?)
Don't worry - that is an easy mistake to make. I have not seen a crash on generating a map - can you give me more details on the steps needed to reproduce? I have been able to generate maps with industries without difficulty and have done so many times in testing.
The latest code on the half-heights branch of the Pak128.Britain-Ex Github repository is half-heights enabled. You will need to compile the Makeobj from the way-improvements branch for it to compile properly and be a usable half-heights pakset, however.
Hmm, I shall try to see after I recompile the industries anew from pak source... Meanwhile, I forget what includes to use for makeobj compile; I added libpng thingy, and now it still says "debuglevel: undeclared identifier:" in imagelist_writer and skin_writer.cc, so I am missing one thing or another.
Junna,
to save you time, I have uploaded the latest Windows binary of Makeobj here (http://www.bridgewater-brunel.me.uk/downloads/Makeobj.exe). Please note that this is a pre-production build, and there is no guarantee that anything built with this version will work with any release version, now or in the future; but it should be usable for testing.
I hope that this helps.
Oh, thank you, that did indeed work, and so far I have not been able to recreate the crash, so it seems (for now) fixed.
Another concern: for non-experimental paksets, cities universally end up as fields of the lowest level buildings (I think that this is because other paksets often have building levels for larger buildings set very high, as compared to pak.Britain). Trying to change this by increasing the renewal rate of buildings and the density does not tend to produce desirable results; low density tends to remain, despite a high %.
(Also: town growth seems still to be excessively high, or is this some attempt to achieve work equilibrium by the game after generating a map?)
Town generation and growth is set for major reworking fairly soon, in which, hopefully, these issues will be addressed.