TrueBrain wrote:

This bug is very real, but hard to fix. Even more because it rarely happens, I don't think it has any priority at all.

This comment was imported from FlySpray: https://bugs.openttd.org/task/73#comment439

GrimRC wrote:

A savegame just before one of the dates where this happens would be helpful.

This comment was imported from FlySpray: https://bugs.openttd.org/task/73#comment484

GrimRC wrote:

I've attached a savegame which is just before a new aircraft is introduced. Select an aircraft in the 'build aircraft' menu and press 'yes' in the aircraft preview mode box which shows around 1st Jan., to test the bug.

This comment was imported from FlySpray: https://bugs.openttd.org/task/73#comment499

GrimRC wrote:

I've attached my patch, which somewhat fixes this problem. It only fixes the 'build aircraft' window, and can be considered a proof-of-concept. Would someone please tell me if this is the right approach? The patch changes .sel_index to 0 normally, and only sets it temporarily if the user clicks on a vehicle. .sel_eng is used to track which vehicle is currently selected.

My patch breaks 'vehicle renew' for aircraft, although I'm not sure how as I don't know how to activate and test this feature in the game. This functionality needs reimplementing too, as it depends on .sel_index somehow. I have put some (lame) fixes to this area to prevent assertions, and allow my patch to be tested as it is (in an incomplete state).

Test this with my savegame in my previous comment.

Attachments

• build_aircraft.patch (5.17 KiB)

This comment was imported from FlySpray: https://bugs.openttd.org/task/73#comment500

GrimRC wrote:

I've given up on fixing this as my patch wasn't deemed an acceptable approach. My patch does fix the problem, but seems to break the autorenew code at the same time. I haven't looked in to the possibility of fixing the autorenew code to work with my patch, but my patch's breaking of the autorenew code wasn't the reason given for it being considered unacceptable. I'm not able to write a fix in a way which would be considered cleaner. The current code can be considered unclean (irrespective of its bugginess), and my patch is simply a tweaking of it (rather than a hefty rewrite).

My offer of extending this (crude) patch to fix all vehicle build types in a similar manner remains open for the foreseeable future, although someone else would have to fix the autorenew code to be compatible with this approach.

This comment was imported from FlySpray: https://bugs.openttd.org/task/73#comment509

Darkvater closed the ticket.

Reason for closing: Fixed

This problem has been fixed as a side-effect of r6722? (eg store a list of vehicles, and reference a selection by EngineID instead of offset)

This comment was imported from FlySpray: https://bugs.openttd.org/task/73