Re: [Freecol-developers] Role vs. Equipment

2013-08-18 Thread Michael T. Pope
On Sun, 18 Aug 2013 09:20:04 +0200 Michael Vehrs wrote: > I'll have a look at that. And I think that all getName/getLabel code > should be moved to Messages if possible. Messages is a good place. The only real alternative I see is to try to have a common "public StringTemplate getLabel()" metho

Re: [Freecol-developers] Role vs. Equipment

2013-08-18 Thread Michael Vehrs
On 18.08.2013 03:37, Michael T. Pope wrote: > On Sat, 17 Aug 2013 14:42:28 +0200 > Michael Vehrs wrote: > >> It probably doesn't matter, as long as we ensure that setRole() or its >> successor is called during initialisation. >> > AFAICT role is either read directly or fixed with setRole

Re: [Freecol-developers] Role vs. Equipment

2013-08-17 Thread Michael T. Pope
On Sat, 17 Aug 2013 14:42:28 +0200 Michael Vehrs wrote: > It probably doesn't matter, as long as we ensure that setRole() or its > successor is called during initialisation. AFAICT role is either read directly or fixed with setRole() in all the Unit initialization paths, so I think you are right

Re: [Freecol-developers] Role vs. Equipment

2013-08-16 Thread Michael T. Pope
On Thu, 15 Aug 2013 17:39:37 +0200 Michael Vehrs wrote: > Alright. Let's get rid of them. I suggest we split the work along > packages. A quick grep shows that there are about 150 call sites, about > half of them in the model package. Shall I tackle the model? Go for it. I only had a brief loo

Re: [Freecol-developers] Role vs. Equipment

2013-08-15 Thread Michael Vehrs
On 13.08.2013 12:02, Michael T. Pope wrote: > How should we handle backward compatibility for the roles? > Specifications in saved games from<= 0.10.5 will not have a > section.>= 0.10.6 have a roles section with two bugs: omitted > model.role.default and model.role.scout does not include the > cap

Re: [Freecol-developers] Role vs. Equipment

2013-08-15 Thread Michael Vehrs
On 15.08.2013 10:40, Michael T. Pope wrote: > On Sun, 11 Aug 2013 15:54:29 +0200 > Michael Vehrs wrote: > >> Update: I have fixed all test failures except for six that occur only >> when running all tests, but not when running the test suite >> individually. I conclude that some tests must be

Re: [Freecol-developers] Role vs. Equipment

2013-08-15 Thread Michael T. Pope
On Sun, 11 Aug 2013 15:54:29 +0200 Michael Vehrs wrote: > Update: I have fixed all test failures except for six that occur only > when running all tests, but not when running the test suite > individually. I conclude that some tests must be messing around with the > specification, causing the sta

Re: [Freecol-developers] Role vs. Equipment

2013-08-13 Thread Michael T. Pope
How should we handle backward compatibility for the roles? Specifications in saved games from <= 0.10.5 will not have a section. >= 0.10.6 have a roles section with two bugs: omitted model.role.default and model.role.scout does not include the captureUnits ability. I can not think of anything be

Re: [Freecol-developers] Role vs. Equipment

2013-08-12 Thread Michael T. Pope
On Sun, 11 Aug 2013 15:54:29 +0200 Michael Vehrs wrote: > Done. I hope my lack of git-fu hasn't caused any regressions. I had time today for a quick look at the first test fail (the second probably has the same cause) and traced it as far as Unit.getLandMoveType() which does a lot of "if (getRole

Re: [Freecol-developers] Role vs. Equipment

2013-08-11 Thread Michael Vehrs
On 11.08.2013 12:59, Michael T. Pope wrote: > On Sun, 11 Aug 2013 12:30:20 +0200 > Michael Vehrs wrote: > >> Update: I have fixed all test failures except for six that occur only >> when running all tests, but not when running the test suite >> individually. I conclude that some tests must be

Re: [Freecol-developers] Role vs. Equipment

2013-08-11 Thread Michael T. Pope
On Sun, 11 Aug 2013 12:30:20 +0200 Michael Vehrs wrote: > Update: I have fixed all test failures except for six that occur only > when running all tests, but not when running the test suite > individually. I conclude that some tests must be messing around with the > specification, causing the

Re: [Freecol-developers] Role vs. Equipment

2013-08-11 Thread Michael Vehrs
On 11.08.2013 09:13, Michael Vehrs wrote: > On 10.08.2013 12:32, Michael T. Pope wrote: > >> On Sat, 10 Aug 2013 11:39:40 +0200 >> Michael Vehrs wrote: >> >> >>> Excellent. The change role message is, of course, an essential part of >>> switching to a role model. >>> >>> >> No w

Re: [Freecol-developers] Role vs. Equipment

2013-08-11 Thread Michael Vehrs
On 10.08.2013 12:32, Michael T. Pope wrote: > On Sat, 10 Aug 2013 11:39:40 +0200 > Michael Vehrs wrote: > >> Excellent. The change role message is, of course, an essential part of >> switching to a role model. >> > No worries. What is your preferred sequence here? I am busy chipping >

Re: [Freecol-developers] Role vs. Equipment

2013-08-10 Thread Michael T. Pope
On Sat, 10 Aug 2013 11:39:40 +0200 Michael Vehrs wrote: > Excellent. The change role message is, of course, an essential part of > switching to a role model. No worries. What is your preferred sequence here? I am busy chipping away at the PlayerExploredTile problems again ATM, but I can probab

Re: [Freecol-developers] Role vs. Equipment

2013-08-10 Thread Michael Vehrs
On 10.08.2013 11:23, Michael T. Pope wrote: > On Sat, 10 Aug 2013 09:35:32 +0200 > Michael Vehrs wrote: > >> Recently, I attempted replacing the current equipment-based model with a >> role-based model. The required changes are pervasive and are certain to >> introduce any number of bugs. Ther

Re: [Freecol-developers] Role vs. Equipment

2013-08-10 Thread Michael T. Pope
On Sat, 10 Aug 2013 09:35:32 +0200 Michael Vehrs wrote: > Recently, I attempted replacing the current equipment-based model with a > role-based model. The required changes are pervasive and are certain to > introduce any number of bugs. Therefore, I propose a gradual migration. > As a first ste