The International Simutrans Forum

 

Author Topic: 1870-1890's carriages lack class definitions  (Read 5621 times)

0 Members and 1 Guest are viewing this topic.

Offline Vladki cz

  • Devotee
  • *
  • Posts: 2725
    • My addons, mostly roadsigns
  • Languages: EN, CS
Re: 1870-1890's carriages lack class definitions
« Reply #35 on: January 23, 2019, 07:57:16 AM »
Hi spenk,

I started to dig into this issue because of bad appearance. If you use some if the shorter carriages, you'll get irregular spaces between carriages.

However the capacity issue is also important. Especially the 3rd suburban vs. 3rd luggage is striking (shorter, less compartments but higher comfort at the same capacity?)

I wouldn't add mail nor tpo to luggage versions. There are already dedicated mail /tpo carriages and the mail brake van has also quite a lot mail capacity. Catering also does not make much sense when you cannot walk between cars during the ride.

Suburban vs. Luggage is already differentiated by comfort and loading time. That's fine. Luggage cars have higher comfort (no luggage on your knees) and higher loading time.

Offline Spenk009

  • *
  • Posts: 271
Re: 1870-1890's carriages lack class definitions
« Reply #36 on: January 23, 2019, 12:15:17 PM »
I can make the changes and if James sees fit they'll be incorporated. I'll try to post the improvements side by side.

Luggage cars have higher comfort (no luggage on your knees) and higher loading time.

I see your point, but I mean nothing about adding a  mail hold nor  accessible luggage catering during the trip (even if corridored sets allow this). I mean that a dedicated luggage carriage would allow passengers from others to decide that leaving their luggage in the separate car is sensible from a certain length of journey on (trips longer than 2/3 hours aren't usually routine commutes). So if the player adds a luggage carriage to a long distance train, it would be beneficial to longer trip taking pax (class dependent).

Offline Vladki cz

  • Devotee
  • *
  • Posts: 2725
    • My addons, mostly roadsigns
  • Languages: EN, CS
Re: 1870-1890's carriages lack class definitions
« Reply #37 on: January 23, 2019, 04:22:52 PM »
I think I understand what you mean by luggage catering, but that would have to be a bigger dedicated luggage van. The luggage compartments are imho small to fit only the luggage of passengers in the same car.

Also overcrowding makes sense to me only in the lowest class in train. Would you pay extra to travel in car with nice soft seats if you couldn't sit down? Also standing passengers usually stay in the corridor, not in compartments. So I would be more inclined to reduce the overcrowding even in 3rd class for these carriages

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Moderator
  • *
  • Posts: 18753
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: 1870-1890's carriages lack class definitions
« Reply #38 on: January 24, 2019, 12:02:48 AM »
There are some complex issues that have been raised. Firstly, length: it is generally preferable for all carriage graphics to have as accurate a length as possible. The lack of granularity in the length parameter is unfortunate, but all sorts of visual chaos will ensue if we start having carriages of deliberately the wrong length to work around this. Far better would be to work on allowing fractional lengths in the code (although this may be some way off unless someone else codes it).

Secondly, luggage: any representation of luggage in comfort must take account of the fact that the luggage carrying capacity of one vehicle in a train can affect the comfort of other vehicles. Doing this would require a change in the code (catering values are unsuitable, as they generate revenue as well as increasing comfort). For this reason, the simulation of luggage cars has generally been omitted, and, if included in some cases, would require revision of the entire pakset to add luggage vehicles to match all of the various railway carriages, which would be a very large task.

As for overcrowding, the policy so far has been to allow overcrowding in all but first class carriages/compartments. The general rule of thumb is, for non-corridor compartment stock, the overcrowded capacity is half as much again as the seated capacity, as generally as many people can stand in the middle of a compartment as can sit on either side of it. The same holds in corridor compartments, but people can also stand in the corridor, giving a higher ratio of overcrowded to seated capacity, and the same in saloon carriages.

Although people may prefer to stand in the corridor given the chance, people would undoubtedly have stood in compartments in corridor carriages if there were no other space available: it is not plausible that people would have refused to get on the train at all just because they have to stand in a compartment rather than in the corridor. People will stand anywhere with enough space if the train is crowded enough.

Edit: I have now had a chance to look over the code. I see that you have re-rendered quite a lot of vehicles: that is very helpful: thank you for that. I do take the point about the front/rear full brakes, as the graphics do show them having a particular end.
However, one thing that I do notice is that you have in some places (e.g. the LBSCR-4Wheel-First-fitted) added first class without modifying the capacity or comfort. The carriage, although modelled on a first class carriage, had the capacity and comfort that it would have had had it been built as a long-distance third class carriage: before classes were introduced, in the LBSCR carriages, class was used as a proxy for whether the carriages were long distance or suburban.

You will thus need to recalibrate the comfort and capacity of these carriages to match with them being first class (i.e. three a side in each compartment and with a comfort matching contemporaneous first class vehicles). For reference, this is what an LBSCR 4 wheel first class compartment looks like:
LBSCR 4 wheel first (interior) by James Petts, on Flickr
LBSCR 4 wheel first (interior) by James Petts, on Flickr

Offline Spenk009

  • *
  • Posts: 271
Re: 1870-1890's carriages lack class definitions
« Reply #39 on: February 04, 2019, 06:08:49 PM »
James, thank you for noticing this. I've gone ahead and changed the capacity, updated the comfort and removed/reduced overcrowding (1st=0,2nd=12 of 50). I'll be happy to use the gorgeous pictures in the rest of the album for future reference. I've also moved a blend file or two to the Blends repo and removed some of the older non-transparent images.

I was wondering whether we should constrain the LBSCR suburbans as the source provided mentions: "These suburban carriages were close-coupled in semi-permanent sets by a central buffing fixture with side chains, and standard buffers were only fitted to the brake end of brake-thirds."

Vladki:
- luggage 3rd should have only 4-compartments capacity (and full length) <- I've amended this for the luggage carriage
- luggage tricomposite - should have 4-compartments 6+8+2x10 and full length <- Your source is correct, but our carriage has only three visible compartments as seen here.
- luggage composite (1/2) - should have 3-compartments 6+2x8, and may be shorter, but should be rescaled. Current 3-compartment image is too long for length=4. Or it could be a 1st + 3rd composite with 4 compartments? <- Is this not a four compartment? Image In any case this can be rendered.
- suburban composite (1/2) should have capacity 12+16 instead of 16+16 <- It's a suburban vehicle, where we won't have too much reason to splurge on comfort where not needed. I think we should keep this.


Offline Vladki cz

  • Devotee
  • *
  • Posts: 2725
    • My addons, mostly roadsigns
  • Languages: EN, CS
Re: 1870-1890's carriages lack class definitions
« Reply #40 on: February 05, 2019, 04:48:02 PM »
- luggage 3rd should have only 4-compartments capacity (and full length) <- I've amended this for the luggage carriage
OK, good
Quote
- luggage tricomposite - should have 4-compartments 6+8+2x10 and full length <- Your source is correct, but our carriage has only three visible compartments as seen here.
- luggage composite (1/2) - should have 3-compartments 6+2x8, and may be shorter, but should be rescaled. Current 3-compartment image is too long for length=4. Or it could be a 1st + 3rd composite with 4 compartments? <- Is this not a four compartment? Image In any case this can be rendered.
I thought about swapping the images for these two. So the tricomposite would be 4-compartments with luggage space in the middle, and 1st/2nd composite would use the 3-compartment image.
Quote
- suburban composite (1/2) should have capacity 12+16 instead of 16+16 <- It's a suburban vehicle, where we won't have too much reason to splurge on comfort where not needed. I think we should keep this.
If there is no reason fir extra comfort, then there is no reason for the 1st classs. I take it that 1st class is 6 seats, 2nd class 8 seats. And the difference betweeen suburban and long distance is only the luggage compartment, that takes approximately the same space as one 3rd class compartment. And 1st/2nd class compartmens have some extra legroom, so only 4 fit the same lenght where 5 3rd class compartments would fit.

So for 3rd class it would be like: 5-compartments (suburban), 4-compartmens (luggage), 4-compartments (brake), 3-compartments (luggage+brake).
Composite would be: 2+2 compartments (suburban), 1+1+2 compartmetns (luggage),  3-compartment composite with luggage could be anything 1+2 or 2+1 or 1st/2nd class only

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Moderator
  • *
  • Posts: 18753
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: 1870-1890's carriages lack class definitions
« Reply #41 on: February 05, 2019, 11:34:34 PM »
Thank you very much for that: I have now incorporated that (having added the English translation texts for the class names for the 4 wheel carriages).

For the 4 wheel suburban carriages, I agree that these should have an exclusive constraint set. Also, you might reduce their length to simulate the shorter buffering, which will allow more of them to fit in short platforms.

Your work on adding classes to the pakset is much appreciated.

Offline Spenk009

  • *
  • Posts: 271
Re: 1870-1890's carriages lack class definitions
« Reply #42 on: February 15, 2019, 06:32:32 PM »
I have been scratching my head at some of the configuration of the lbscr carriages. I'm listing my findings with proposed solutions.

Craven Carriages (built 1866 - 1872) - suburban carriages due to their comfort values (lo-38, med-42).
The Cravens don't feature a first class. If this is intended due to the suburban nature, that's understandable.
The Cravens have mail carriages, which seems unusual for a suburban train. If you're supposed to use them to help connect up the mail networks using select stations with mail capability, we could mention it to players in the vehicle comment/copyright section.

Stroudley Carriages (built 1872, improved 1875 - 1901) - main line carriages due to comfort (low-62 but brakes have 52, med-68, hi-77)
As previously mentioned, I've created a third class vehicle and also a composite (source). It is in the newest pull request.
These Carriages retire Jan 1872, 6 wheel improvements don't arrive until December 1872. I suggest moving retire_month to November.
These carriages are main line / long distance. I suggest I lower the capacities for second, if not third class, to reflect the improved comfort. The brake carriage will be fixed to 62 comfort.
I create a set of suburban carriages with higher 1st class capacity, lower comfort across the board, constrained to each other and slightly cheaper purchase price.

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Moderator
  • *
  • Posts: 18753
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: 1870-1890's carriages lack class definitions
« Reply #43 on: February 16, 2019, 12:33:06 AM »
Thank you for your thoughts in relation to this. A little historical information may be of assistance in working out how these are supposed to work.

Before the 1870s, there were no separate long distance and suburban carriages. It was the LBSCR who first introduced this idea, I think, in the 1870s (or, at least, at around the same time as other railways). Early railway carriages were really extremely uncomfortable: they were small and cramped, had a hard ride, were draughty, had no heating (other than the ability to hire hot water bottles at stations), no lavatories and crude lighting consisting of a rapeseed oil pot lamp (one per two compartments in third and sometimes second class quite often) which produced very little light but a fair bit of smoke.

Second and third class carriages would have had the same capacity: the difference was simply that the second class carriages had padding on the seats: the third class carriages merely had wooden benches. Going back further beyond the 1860s to the 1850s and 1840s, the third class carriages would not even have had proper windows (save for one in the door).

In relation to the retirement date of the Craven carriages, the retirement date is April 1872, the same month as the unfitted Stroudly carriages are introduced. They are available until January 1877. However, fitted Stroudley carriages are available from September 1875 - I am not sure, therefore, what the issue with the dates is: can you elaborate?

In relation to your pull request, I imagine that you will need to modify it to take into account what I wrote above regarding 1860s carriages and the differnce between second/third class (unless you did not incorporate these features into the pull request, in which case, please let me know, and I will review it).

Thank you very much for your work on this.

Offline Spenk009

  • *
  • Posts: 271
Re: 1870-1890's carriages lack class definitions
« Reply #44 on: February 16, 2019, 08:08:04 AM »
To the retirement dates: My apologies. I lost overview of the what was happening when. Of course braked four-wheel carriages are available before the introduction of six-wheel carriages.

I have not incorporated a capacity difference or suburban stock suggestion in the pull request. The intention was to reduce mainline compartment capacity from 10 to 8 (second and third following your description) and create an explicit suburban set (Cravens don't feature a first class which may be a result of being suburban). The main-line carriages shouldn't be constrained but the suburbans in this case could. I have thought about this further and unless players really want it, it was just a small idea.

I've amended files in the pull request to enable automatic merging. The changes are: addition of a composite vehicle, addition of a third class vehicle, fixing first class capacity, moved/removed a few .blend files.

I'd like to share that the 1850s to 1860s transition looks very good, the change in stock, locomotive power and track requirements is very nice. Thank you James for your patience with us. I would continue with lbscr-48/54ft-augmentation carriages next, or are there higher priority tasks in transparent workflow conversion?

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Moderator
  • *
  • Posts: 18753
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: 1870-1890's carriages lack class definitions
« Reply #45 on: February 16, 2019, 12:24:56 PM »
Thank you for this: that is helpful. It is probably better not to have suburban/mainline Craven carriages, as these did not exist in reality.

In relation to reviewing your pull request, a few things that will need looking at:

(1) you add some non-LBSCR fitted four wheelers (4-wheel-1870s-third-fitted), but the coupling constraints are not updated on all vehicles that can couple to the 1870s 4 wheelers to match;
(2) you change the name of LBSCR-4Wheel-Second to LBSCR-4Wheel-third without adding a compat.tab entry, meaning that any saved game in which players had a LBSCR-4Wheel-Second vehicle will be broken - the better solution here is probably to add a third class as well as the existing second class vehicle and modify the comfort of the second class vehicle;
(3) the full brakes need a "none" constraint as the last in the sequence of "next" constraints so that they can be at the back of trains; also, they should not have previous constraints, since they should be able to couple immediately behind the locomotive if the player desires - generally, it is preferable for full brakes (other than ones that can only go either at the front or back of the train, which are rare) to have no coupling constraints at all; and
(4) the comfort of the 1872 Stroudley carriages needs consideration: the pre-classes scheme had the 2nd class used as the long distance vehicle and 3rd class used as the suburban vehicle - in the comfort calibration overview.ods file, you will see that the suburban 1870s carriages are at a comfort of 52 and the low capacity carriages (i.e. long distance carriages) are at 62; I note that you change the comfort of the brake carriages from 52 to 62 "to match other vehicles in the set", but these vehicles are a different type and should have a different comfort: use the spreadsheet as the guide rather than the .dat files.

Thank you again for your work on this: it is much appreciated.

Offline Spenk009

  • *
  • Posts: 271
Re: 1870-1890's carriages lack class definitions
« Reply #46 on: February 16, 2019, 01:47:44 PM »
Thank you for this: that is helpful. It is probably better not to have suburban/mainline Craven carriages, as these did not exist in reality.
I meant this in regards to Stroudley introduced stock, which were converted main-line carriages.

(1) I must have made an error in the fundamental workflow here. Could you point out the missing additions?
(2) The Second class vehicle is not intended to be replaced by either the third or composite carriages. I supplemented as the source mentions were built. Just in case, I started a savegame from an older nightly source, running a train with the assembly as I expect intended, saved and loaded in a freshly compiled pak. I encountered no issues.
(3) I tried to find the source of this. I think it may be due to the fact that I've added a full brake front and amended the rear entry in en.tab, but this has not propagated to your branch.
I do take the point about the front/rear full brakes, as the graphics do show them having a particular end.
entries between upstream/master and spenk009/master.
(4) That is an application I had not thought of. It's just that the brake carriages have comfort=52, there are no third class carriages to make use of as suburban and the comfort value defined across all comforts in the carriages introduced at the exact same time, scheme, type, looks have: comfort[0]=0, comfort[1]=62, comfort[2]=68, comfort[3]=77.

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Moderator
  • *
  • Posts: 18753
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: 1870-1890's carriages lack class definitions
« Reply #47 on: February 16, 2019, 11:58:59 PM »
(1) My apologies: I seem to have missed in my previous post that you had added these in the commit on at 1736h on the 15th of February.

(2) Looking at this again, I think that I might have misinterpreted the way in which Git for Windows presented the commit: it appeared to involve a deletion of LBSCR-4Wheel-Second and its replacement with LBSCR-4Wheel-third, but in fact it shows this replacement on a copy of the file which you had renamed to provide a new third carriage. My apologies.

(3) This relates to the 6 wheel carriages, in particular the carriage with the name LBSCR-6Wheel-guard: the current version in the pakset does not have any constraints defined, but your commit adds the same constraints as for a carriages such as the LBSCR-6Wheel-4-com-first, which would not allow it to be at either the beginning nor end of a train. The 6 wheel version does not have separate front/rear brake graphics.

(4) The 1870s carriages, unlike the 1860s carriages, need two schemes: one for the suburban carriages and one for the long distance carriages.

Offline Vladki cz

  • Devotee
  • *
  • Posts: 2725
    • My addons, mostly roadsigns
  • Languages: EN, CS
Re: 1870-1890's carriages lack class definitions
« Reply #48 on: February 17, 2019, 10:34:35 PM »
Hello, I have noticed that some of the LBSCR 6-wheel carriages have retire year 1892 (passenger non-break), while others (passenger break and mail) have retire year 1907. Is that intended?

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Moderator
  • *
  • Posts: 18753
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: 1870-1890's carriages lack class definitions
« Reply #49 on: February 17, 2019, 11:06:22 PM »
I think so - the passenger carriages were replaced with bogie carriages, but, initially, bogie carriages were not built for mail/brake applications.

Offline Vladki cz

  • Devotee
  • *
  • Posts: 2725
    • My addons, mostly roadsigns
  • Languages: EN, CS
Re: 1870-1890's carriages lack class definitions
« Reply #50 on: February 19, 2019, 08:34:56 PM »
But the bogie variants  have same intro (LBSCR-non-cor-lav  in 1889 and LBSCR-sub-bogie-42ft in 1894) for both brake and non-brake (although no mail).
So either the passenger 6 wheelers (both brake and non-brake) should end at same time, probably 1894 or both should be extended to 1907 to match the mail 6-wheelers.

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Moderator
  • *
  • Posts: 18753
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: 1870-1890's carriages lack class definitions
« Reply #51 on: February 23, 2019, 12:51:46 AM »
Those are the passenger brakes, not the full brakes: the only full brakes available are the 6 wheelers.

Offline Vladki cz

  • Devotee
  • *
  • Posts: 2725
    • My addons, mostly roadsigns
  • Languages: EN, CS
Re: 1870-1890's carriages lack class definitions
« Reply #52 on: February 23, 2019, 01:55:13 AM »
Yes I just thought that the passenger brakes should have the same intro/retire as passenger non-brakes. So that only the mail 6-wheelers will be available longer.

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Moderator
  • *
  • Posts: 18753
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: 1870-1890's carriages lack class definitions
« Reply #53 on: February 23, 2019, 10:36:12 AM »
That would leave a period without any full brakes available, which would be problematic as it would prevent the formation of an all mail train.

Offline Vladki cz

  • Devotee
  • *
  • Posts: 2725
    • My addons, mostly roadsigns
  • Languages: EN, CS
Re: 1870-1890's carriages lack class definitions
« Reply #54 on: February 23, 2019, 11:29:04 AM »
I'm not sure if I am clear enough. I propose to change the retire date of:

name=LBSCR-6Wheel-brake-front
intro_year=1877
retire_year=1907
freight=Passagiere

name=LBSCR-6Wheel-brake-rear
intro_year=1877
retire_year=1907
freight=Passagiere

to retire_year=1892 to match all other passenger (non-brake) 6-wheelers.

And to keep the full brake (and all mail/fruit/milk 6-wheelers) as is:

name=LBSCR-6Wheel-guard
intro_year=1877
retire_year=1907
freight=Post


Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Moderator
  • *
  • Posts: 18753
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: 1870-1890's carriages lack class definitions
« Reply #55 on: February 23, 2019, 12:29:44 PM »
My apologies for misunderstanding: I see what you mean now. I have made the adjustment as requested.

Offline Vladki cz

  • Devotee
  • *
  • Posts: 2725
    • My addons, mostly roadsigns
  • Languages: EN, CS
Re: 1870-1890's carriages lack class definitions
« Reply #56 on: March 04, 2019, 08:48:49 PM »
Speaking about lbscr-6wheelers, I have noticed that the lbscr-6wheel-mail*.png is much bigger than other lbscr-6wheel-*.png, so it need re-rendering.
Also lbscr-6wheel-third-umber*.png is a little bit bigger, although the difference is not so remarkable as the mail car.
And the lbscr-6wheel-brake-front_SE/S.png and lbscr-6wheel-brake-rear_N/NW.png are darker than other 6-wheelers in orange livery
« Last Edit: March 04, 2019, 09:47:38 PM by Vladki »

Offline Vladki cz

  • Devotee
  • *
  • Posts: 2725
    • My addons, mostly roadsigns
  • Languages: EN, CS
Re: 1870-1890's carriages lack class definitions
« Reply #57 on: March 16, 2019, 10:34:21 PM »
I think my previous post should be split as separate bug report. Also the lbscr-6-wheel 3rd class in 1904's livery needs rescaling

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Moderator
  • *
  • Posts: 18753
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: 1870-1890's carriages lack class definitions
« Reply #58 on: March 17, 2019, 01:49:41 AM »
I think that these are more in the category of vehicles that could do with being re-rendered in due course than bug reports - unless there are vehicles in the same set that have not been re-rendered?

Offline Vladki cz

  • Devotee
  • *
  • Posts: 2725
    • My addons, mostly roadsigns
  • Languages: EN, CS
Re: 1870-1890's carriages lack class definitions
« Reply #59 on: March 17, 2019, 10:12:59 AM »
Most of the lbscr 6-wheelers are already re-rendered. These are a few I have noticed to look different from the rest.

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Moderator
  • *
  • Posts: 18753
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: 1870-1890's carriages lack class definitions
« Reply #60 on: March 17, 2019, 11:19:30 AM »
I wonder whether Spenk009 might like to re-render the remaining LBSCR 6 wheelers, in that case?

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Moderator
  • *
  • Posts: 18753
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: 1870-1890's carriages lack class definitions
« Reply #61 on: April 22, 2019, 11:30:14 AM »
I have now remedied the issue with the LBSCR 6 wheel mail carriage: the problem was in fact caused by using an old .blend file that had not been rescaled. The mail carriages were actually intended to be created from the full brake carriages with the layer containing the guard's lookout duckets hidden, which is what I have now done.

Incidentally, I also note that some of the .blend files for the re-rendered carriages had been saved in their rotated state (i.e., after the automatic creation of the images). It is important that this never be done, as it is very difficult to revert to the original rotation again for the next export: always save then export, then reload the save from before the rendering/export before doing anything else. I had to recover old files from the Github archive to fix this issue.



I have also made some progress in the last few days adding class definitions to some more carriages, including some late 1870s Midland Railway bogie carriages, and the complete set of outstanding Metropolitan and District Railway carriages (including EMUs).

Offline Spenk009

  • *
  • Posts: 271
Re: 1870-1890's carriages lack class definitions
« Reply #62 on: April 27, 2019, 11:41:09 AM »
My apologies for embarking on some of the maintenance of the pakset. I believe I rebuilt the erroneous mail carriages and pushed them to github. Accordingly, the pak repository would have also been updated on my end.

There vehicles in the works, mainly a few additions to classes. I'm inclined to use the same images for composite and other-class versions of carriages, but this is bad practice and doesn't concur with the level of detail invested in other vehicles. I don't have much documentation on how vehicles looked and decent images of rarer carriages are scarce.

I've found employment in a different company requiring very long hours (including weekends) and 3 hours of daily commute. I'm finding it very difficult to make time for social responsibilities and hobbies. I apologize again for lack of presence and missing the prior two posts.

I will check what progress I've made and report it.

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Moderator
  • *
  • Posts: 18753
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: 1870-1890's carriages lack class definitions
« Reply #63 on: April 27, 2019, 12:34:32 PM »
Don't apologise for working on the pakset - work is much appreciated. Just be careful not to save the .blend files in a rotated, post-export state, as this is quite time-consuming to undo.

Also don't apologise for not having much time: that is a problem common to many of us who are not retired: nobody is obliged to do this work, although it is always appreciated when people do.

If you are having problems finding images of what composite or first or second class carriages of any particular type look like, please let me know on this thread: I have quite a number of books about railway carriages, and I am also a member of a model railway club with an extensive library, so I can fairly easily find pictures or drawings of carriages of a very wide range of types.

It is better to ask than to use the same image for composites as third class vehicles, as having the two look identical might be confusing for players. However, some second class carriages (and some second/third compsoites) would have looked indistinguishable from their third class counterparts (railway practice was generally to give second class passengers the same amount of space as third class passengers, but give them padded seats (or later, slightly better padded seats)), such that the differences on the outside of the carriage would be confined to different lettering/numbering on the doors which would be invisible at our scale.

Thank you again for your work on the pakset: it is much appreciated.

Offline Vladki cz

  • Devotee
  • *
  • Posts: 2725
    • My addons, mostly roadsigns
  • Languages: EN, CS
Re: 1870-1890's carriages lack class definitions
« Reply #64 on: May 02, 2019, 08:31:20 PM »
I dare to continue here - a question about lbsrc-craven carriages. They have brake-front, brake-rear and full-brake (mail), just like many other carriage sets of late 1800's, but lack the constraints forcing particular order. Only constraints are next=any for non-braked carriages. Is it just forgotten, or intentional?

Offline Spenk009

  • *
  • Posts: 271
Re: 1870-1890's carriages lack class definitions
« Reply #65 on: May 04, 2019, 11:45:40 AM »
I have looked into the changes and updates made, tried to push a commit but failed somehow. I don't think I can recover the changes made. I have created four-wheel suburban carriages, which is in the pull request currently outstanding.

I dare to continue here - a question about lbsrc-craven carriages. They have brake-front, brake-rear and full-brake (mail), just like many other carriage sets of late 1800's, but lack the constraints forcing particular order. Only constraints are next=any for non-braked carriages. Is it just forgotten, or intentional?

The lbscr-cravens follow the same pattern as the normal Four-Wheel carriages, which can be arranged without a brake van adjacent to the locomotive (a single brake van at the rear). We spoke about it around reply 45, so I think that will be kept.

Most of the lbscr 6-wheelers are already re-rendered. These are a few I have noticed to look different from the rest.
Could you list the offending carriages? I will try to fix these then.
« Last Edit: May 04, 2019, 12:53:58 PM by Spenk009 »

Offline Vladki cz

  • Devotee
  • *
  • Posts: 2725
    • My addons, mostly roadsigns
  • Languages: EN, CS
Re: 1870-1890's carriages lack class definitions
« Reply #66 on: May 04, 2019, 04:59:03 PM »
See my post from March 4th

Offline Spenk009

  • *
  • Posts: 271
Re: 1870-1890's carriages lack class definitions
« Reply #67 on: May 05, 2019, 12:02:51 PM »
I've tried to find the issues in the post from March 4th, but have only been able to find and fix the darker tone in carriage. I've moved example carriages onto a siding here. Can you point out my error in looking into the issue?

The blend files feature a few vehicles that are designed wider than others. I believe I had fixed these, but the changes have not propagated to the upstream repository due to the bad saving practice. James, can you point out the commit that added the erroneous files?

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Moderator
  • *
  • Posts: 18753
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: 1870-1890's carriages lack class definitions
« Reply #68 on: May 09, 2019, 12:19:17 AM »
Thank you for your work on this.

First of all, to answer Vladki's question - the coupling constraint scheme for many unfitted vehicles (i.e., those without continuous brakes) is much simpler than for the fitted vehicles: they unfitted vehicles have the same types of coupling constraints as wagons (that is, all but brake carriages have constraint[next]=any).

Spenk - I am a little confused as to what exactly is on your Github repository. Looking at the merge data from your branch, I see a large number of changes going back to February (e.g. "Added LBSCR 4 wheel third class coach") - is this something that you have already done? There are changes for which I do not understand the rationale - for example, you have modified the LBSCR long distance unfitted carriages to be able to be upgradable to the suburban fitted carriages; may I ask what the reasoning was for this?

Also, I believe that I have now fixed the issues with the LBSCR 6 wheel carriages. May I ask what you did by way of re-rendering in the commit "brighter image to match the rest of the set"? Vladki - may I ask which LBSCR 6 wheel carriges, if any, you believe are still incorrectly rendered with the latest nightly build of the set?

You write that you failed somehow to push a commit - can you elaborate? I am not entirely sure that I understand what you are describing here. May I ask to what changes that you are referring when you write of being unable to recover the changes made?

I note that the commit still contains additional suburban versions of the Craven carriages; I believe that I had explained in an earlier post that these carriages did not exist, as there was no distinction between suburban and long distance carriages in the 1860s. You will need to revert these changes before I can merge from your branch.  I believe that there are some other issues discussed above that will likewise need to be dealt with before I can merge this.

I also note that there is a commit with a commit message, "Update on 1870s LBSCR carriages, with two additional carriages made available", but the only thing in that commit is a removal of all the constraints on one of the LBSCR 6 wheel carriages. I presume that this was not intended? May I ask what was intended here?

Offline ACarlotti

  • *
  • Posts: 483
Re: 1870-1890's carriages lack class definitions
« Reply #69 on: May 09, 2019, 03:50:39 AM »
I tried investigating what was in the relevant commits, and found it initially quite confusing. I've listed some the things that contributed to my confusion, along with some suggestions for making things clearer in the future.
I expect some of these points could be relevant more generally; they're probably also picking up on the sort of things that people might not notice until they've tried investigating the full history of lots of bits of code (which I have been doing a lot of both while working out how to merge stuff from Standard accurately, and more generally while trying to learn how bits of code were intended to work while debugging or investigating inconsistenties in the code).

Spenk009:
1. In the merge commit 35a2de1 you made some additional changes that weren't present in either parent. This effectively makes them much harder to identify - using 'git log -p' (similar to my normal way of viewing history) doesn't show changes made during merge commits. I had to discover and add the '--cc' flag to view the changes made in that merge. (At least I can now know about that flag for future use). This is an extract of the output:
Code: [Select]
  LBSCR-4Wheel-Second-fitted
- LBSCR four wheel five compartment carriage (continuously braked) (second class)
+ LBSCR four wheel five compartment carriage (second class) (continuously braked)
 +LBSCR-4Wheel-third-fitted
- LBSCR four wheel five compartment carriage (continuously braked) (third class)
++LBSCR four wheel five compartment carriage (third class) (continuously braked)
 +LBSCR-4Wheel-composite-fitted
- LBSCR four wheel five compartment carriage (continuously braked) (first/second class)
++LBSCR four wheel five compartment carriage (first/second class) (continuously braked)
  LBSCR-4Wheel-Brake-rear-fitted
- LBSCR four wheel three compartment brake carriage (rear) (continuously braked) (third class)
+ LBSCR four wheel three compartment brake carriage (third class) (rear) (continuously braked)
  LBSCR-4Wheel-Brake-front-fitted
- LBSCR four wheel three compartment brake carriage (continuously braked) (third class)
- LBSCR-4Wheel-full-brake-fitted
- LBSCR four wheel mail/brake carriage (continuously braked) (rear)
+ LBSCR four wheel three compartment brake carriage (third class) (continuously braked)
 -LBSCR-4Wheel-full-brake-fitted
 -LBSCR four wheel mail/brake carriage (continuously braked) (rear)
++LSCR-4Wheel-full-brake-fitted
++LBSCR four wheel mail/brake carriage (third class) (continuously braked) (rear)
  LBSCR-4Wheel-full-brake-fitted-front
--LBSCR four wheel mail/brake carriage (continuously braked)
++LBSCR four wheel mail/brake carriage (third class) (continuously braked)
  LBSCR-6Wheel-5-com
The lines beginning '++' are ones which were present in neither parent. Obviously it is sometimes necessary to introduce new lines (or new versions of existing lines) to deal with merge conflicts, but in this case I think it would have been better to make those changes in a separate new commit, as this makes it easier to trace their history.

2. The history graph is quite complicated in a couple of places ('--oneline' used here for concision):
Code: [Select]
$ git log --graph --boundary --oneline  upstream/master..Spenk009/master
* c76627b25 (Spenk009/master) LBSCR 6 wheel carriages
*   60b0015d1 Merge remote-tracking branch 'origin/master'
|\ 
| *   9849fbc64 Merge branch 'master' into master
| |\ 
* | | 50b471bbf UPDATE: LBSCR 4 Wheel Carriages
* | | 578cfe9bc Addded Suburban versions of 4 wheel LBSCR Craven Carriages.
|/ / 
* | 6ef518663 Update on 1870s LBSCR carriages, with two additional carriages made available
* | 6f6ef36f1 Fixing full-brake entries and addition of rear on a brake vehicle
* |   fce4ab076 Merge branch 'master' of github.com:Spenk009/simutrans-pak128.britain
|\ \ 
| * \   35a2de11a Merge branch 'master' into master
| |\ \ 
* | | | c920d6371 Modified comfort value to match other vehicles in the set.
|/ / / 
* | | e07950f34 Added LBSCR 4 wheel composite carriage
* | | 71692e69c Added LBSCR 4 wheel third class coach
o | | 141431bdb Catching up to James, 04.02.19
 / / 
o | 50c3eae6c FIX: Incorrect constraint for the LNWR 42ft brake carriages CHANGE: Add street/stop name
 / 
o ece17613c FIX: The previously modified TPOs could not be coupled to the Mark 1 BG
I don't know why this is happening, but it seems that when you are merging from jamespetts' master, you are subsequently making some commits onto the pre-merge version, and then having to create a second merge to include the first one. It would be much cleaner if you only had to merge once. The simplest way to do this in future would be to remember to pull from Github before making further changes (or remembering to commit before carrying out the merge if that was the issue); however, if you forget to do this you might also have the option of rebasing you latest work onto the merge from Github so that only one merge commit is present.

3. I notice that commit 141431b (commit message "Catching up to James, 04.02.19") was in fact a change to the .gitignore file which doesn't seem to fit the commit message, and was subsequently included and then reverted my James. I presume there was just some mistake there.

4. Commits 71692e6 and e07950f made corrections to en.tab that had already been present upstream for a week or two. I guess it's easy to have that happen accidentally - the only thing I might have done differently is edited my commit history to remove my own changes once I realised they were upstream. But that's probably a matter of taste.

James:
I'm confused about this recent merge commit:
Code: [Select]
$ git log --boundary --graph upstream/master~~..upstream/master~
*   commit 80c7c5f191b529b64ad36c2a46f9d42a53b187bf
|\  Merge: ece17613c d25a25818
| | Author: James E. Petts <jamespetts@yahoo.com>
| | Date:   Thu May 9 01:03:23 2019 +0100
| |
| |     Merge branch 'master' of https://github.com/MatthewForrester/simutrans-pak128.britain.git
| |   
| *   commit d25a25818d10f0af1f4c00d479f0ccfc48b349f9
| |\  Merge: e60dd93c8 6e7e1c2ad
| | | Author: MatthewForrester <38819845+MatthewForrester@users.noreply.github.com>
| | | Date:   Thu Jun 14 05:12:58 2018 +0100
| | |
| | |     Merge pull request #1 from MatthewForrester/hold-fix
| | |     
| | |     Add the same way constraint as this ship's hold
| | |
| | o commit 6e7e1c2ade8fa391be7e250dfac20fd241224e24
| |/  Author: MatthewForrester <38819845+MatthewForrester@users.noreply.github.com>
| |   Date:   Thu Jun 14 05:08:54 2018 +0100
| |   
| |       Add the same way constraint as this ship's hold
| |
| o commit e60dd93c8e3ec2f22b8fc7277911a7b16609b871
|   Author: James E. Petts <jamespetts@yahoo.com>
|   Date:   Mon Jun 11 08:47:44 2018 +0100
|   
|       FIX: Incorrect capacities for the fitted 1870s composite carriage (Rollmaterial)
|
o commit ece17613cc4a58b08bfaee26120ee789e291b028
  Author: Rollmaterial <r.sivonen@tu-bs.de>
  Date:   Mon Apr 29 22:38:16 2019 +0200
 
      FIX: The previously modified TPOs could not be coupled to the Mark 1 BG
It made no change to the source code, and adds only one other new commit - this was a merge between commit 6e7e1c2 and its parent (both already included in your master) which was almost a year old.



I may be assuming too much Git competency here, but my suggestion for how tidy up this branch history to make it clearer (and to avoid having to include seperate commits including and removing suburban versions of the Craven carriages) would be to create a new branch on top of James' master, and cherry pick the commits that you still want to include onto the new branch (I think there're at most five).
As a general point, I'm very much in favour of editing local commit history when doing so can make it less confusing - as long as those commits haven't been shared and (possibly) built upon by someone else, then this shouldn't cause any of the problems caused by more invasive history-editing (such as duplicate commits arising, or forcing other people to rebase onto your amended commits). There're many examples of this sort of thing in commits I've uploaded so far - typical examples are when I add some stuff to an existing commit, or when I realise that I made a mistake porting changes from Standard five commits ago and go back to correct the original commit (instead of making a new one and leaving several broken versions in between).