[bug 8.x] Incomplete reversing of convois June 05, 2010, 05:26:07 pm With a train servicing sack stations, the convoi reverses placing the loco at the head of the train in order to depart from the station. However, wagons (such as brake wagons) that are required to run last in the convoi stay in their "original" position, which in this case becomes immediately after the loco; which of course is a bug in itself...Furthermore, at least with pak128.Britain-Ex the LNWR brake wagon (LNWR-1847-Brake) this causes the graphics to be messed up as well, as can be seen in the following two images, one with the brake wagon (w/o windows) in its proper place at the rear of the convoi, and the other with it immediately following the loco. Quote Selected
Re: [bug 8.x] Incomplete reversing of convois Reply #1 – June 05, 2010, 05:47:29 pm Sanna,thank you for the report. There are several things going on here. Firstly, the graphics: this is a difficult problem to fix completely because simply reversing some things in a train/convoy without reversing other things reveals weaknesses in alignment accuracy in paksets that is concealed when, as with Simutrans-Standard, reversing does not take place. If you test reversing with Pak64, for example, even after modifying the values of vehicles to make them reversible, then these issues do not occur (at least, from my recollection when I tested this some time ago).The normal result of this problem is a gap between the locomotive and carriages. However, in your second screenshot, there is a second problem: there is also a gap between the second and third carriages, and the first and second carriages are overlapping. That is caused by the LNWR-1847-Mail vehicle not having "bidirectional=1" set in its .dat file, which means that its graphic did not reverse when those of the other vehicles did, resulting in misalignment. I have fixed that issue on the latest push to the Github repository for Pak128.Britain-Ex.The other issue is one of design: I had set it deliberately to allow, in reversed formation, vehicles that normally have to be at the rear of a convoy to remain in their relative position in the case only of p****enger and mail vehicles, as I do not think that it was the practice to shunt a p****enger/mail brake vehicle around the end of the train when the locomotive changed ends. This is not the case for goods trains with brake vans (which, consequently, have a longer turnaround time).Thank you for the report, though! Quote Selected
Re: [bug 8.x] Incomplete reversing of convois Reply #2 – June 05, 2010, 08:28:53 pm Quote I have fixed that issue on the latest push to the Github repository for Pak128.Britain-Ex.i've seen a commit at jamespetts / simutrans-pak128.britain master,no commits for pak128.britain-experimental (since 1 March) and pakbritain-experimental (10 May). Is the pak britain experimental repository: jamespetts / simutrans-pak128.britain?i thought this was your development version of pak britain standard? Quote Selected
Re: [bug 8.x] Incomplete reversing of convois Reply #3 – June 05, 2010, 09:32:32 pm Sdog,I've had to change to using a new Github repository for Pak128.Britain-Ex: see here for details. Quote Selected
Re: [bug 8.x] Incomplete reversing of convois Reply #4 – June 05, 2010, 09:50:32 pm Quote from: jamespetts – on June 05, 2010, 05:47:29 pmThe other issue is one of design: I had set it deliberately to allow, in reversed formation, vehicles that normally have to be at the rear of a convoy to remain in their relative position in the case only of p****enger and mail vehicles, as I do not think that it was the practice to shunt a p****enger/mail brake vehicle around the end of the train when the locomotive changed ends. This is not the case for goods trains with brake vans (which, consequently, have a longer turnaround time).I am not altogether convinced that is a good idea to allow the constraints created by the pak manager to be broken when the convoi is reversed. It may very well be that it is ok in the specific case with brake wagons and p****enger trains, but if the reverse code does not honour the constraints, then I am afraid that more "strange" combinations will occur. I do believe it would be better if a convoi always had a structure that the constraints allow. Quote Selected
Re: [bug 8.x] Incomplete reversing of convois Reply #5 – June 05, 2010, 09:52:40 pm Quote from: sanna – on June 05, 2010, 09:50:32 pmI am not altogether convinced that is a good idea to allow the constraints created by the pak manager to be broken when the convoi is reversed. It may very well be that it is ok in the specific case with brake wagons and p****enger trains, but if the reverse code does not honour the constraints, then I am afraid that more "strange" combinations will occur. I do believe it would be better if a convoi always had a structure that the constraints allow.Hmm - how might this be accommodated without breaking all the p****enger/mail convoys and making them reverse in unrealistic ways? Quote Selected
Re: [bug 8.x] Incomplete reversing of convois Reply #6 – June 05, 2010, 10:07:42 pm IMHO, if the brake van in question does not need to be at the rear, then it should not be so constrained in the pakset. If this calls for another brake van (used in freight trains that do need to be at the end), then so be it...I still think that constraints in the pakset should be followed even when driving in reverse.EDIT: This of course then causes the p****enger wagons to be allowed to be at the rear of the convoi.... Quote Selected
Re: [bug 8.x] Incomplete reversing of convois Reply #7 – June 05, 2010, 10:14:30 pm Ahh, the difficulty with that is that the practice in the UK, as I understand it, was for a long time that there would be a brake vehicle at one end of the train, but that it would not be shunted around to the other end when the train reversed; simply enforcing the existing constraints will not achieve this result. Quote Selected
Re: [bug 8.x] Incomplete reversing of convois Reply #8 – June 05, 2010, 10:22:05 pm In that case, breaking the constraints by the pakset manager might make Exp work better with pak128.Britain, but then again work worse with other paksets? I know you have a dual role here both as dev. and pakset maintainer (for the -Ex pakset at least), but I really do not think that pakset considerations should influence the coding of Exp... or at least not in this manner. If the pakset definition needs to be expanded to make this work as desired in pak128.Britain, then that is of course fine; but making "hardcoded" choices on how specific vehicle should behave based on the intended behaviour in a specific pakset is imho not a good idea... Quote Selected
Re: [bug 8.x] Incomplete reversing of convois Reply #9 – June 05, 2010, 10:34:36 pm What is the practice with p****enger or mail carrying brake vehicles in other countries, do you know...? Quote Selected
Topic: [bug 8.x] Incomplete reversing of convois (Read 3935 times)
previous topic - next topic
