Server Development
Posts: 236
Joined: Fri Apr 10, 2015 3:08 pm
Tue Sep 01, 2015 2:02 pm
[0.7.81] Industry tracker
Creating a new topic to keep track of the bugs, related to industry (manufacturing, research, copying, invention).
Did a thorough testing of current industry functions.
Here are the current results:
Manufacturing - operational. Tested functions was Job start, Delivering, Aborting.
- Job starts as per normal - sets the correct time for both job and assembly line it uses.
- Delivering puts the specified amount of items in the hangar.
- Aborting the task removes it from the jobs list, returns a blueprint, does not return the materials.
So, as a conclusion - manufacturing's fine. There are just a couple notes that will go to the other parts of report.
Copying - functional, a bit bugged.
As blueprints unpacking is not yet functional, there are a few glitches that does not really break the functionality, but look weird.
- 1/1 copy run returns...an actual copy! I mean - you can't see the stats, but icon in item info window gets loaded, and it indeed shown as a copy.
- multiple runs/1 copy returns 1 copy with specified amount of runs.
- 1 run/multiple copies - returns the specified amount of 1-run copies. Problem is - they get stacked, so that leads us to a little sub-test:
- Manufacturing multiple stacked BPC's - If stack was used for manufacturing start, system will subtract the run from EVERY blueprint in that stack. It means, that if you have dozen of 1-run BPC's, you start a 1-run job using them, ALL 10 BPC's will get nullified in process.
- Manufacturing the depleted copy - if copy have no runs left, client will give a "The job cannot be created as the blueprint does not have enough production runs left." pop-up.
- Manufacturing, selecting more runs that copy have - same alert pop-up
- Manufacturing using single-run copy - seems to work. It only glitched once for me - after i installed the job with that BPC, "Get Jobs" button was returning me the empty window. Server had the following client stack trace:
When delivered, system will return the goods AND the copy. Empty copy, with no runs left, so i just had to trash it.
- Manufacturing, using multi-run copy - in this case, system will subtract the specified runs amount from the overall number. If i have a 3-run copy and i start a 1-run job, system will return 1-run amount of goods and 2-run BPC.
Research - functional, a bit bugged.
Again, due to missing un-packing feature of BP's, it's not crashing the logic, but bugged.
- Research start/delivery/abortion works as per normal.
- Blueprint, indeed, gets it's attributes changed. The problem is - if you stack a researched BP with un-researched one, it seems like BOTH will bestow the same attributes (in my case it was 2 researched BPO's).
Invention - Not coded.
Dead-block is in place - any attempt to start the invention job will result in "The activity requested is not supported." alert pop-up.
So, as a conclusion:
The biggest issue in industry now is a stackable blueprints. In 2 cases of 4, it makes a dupe possible (ME/PE levels for the whole stack and turning a copy into original). In the rest it just causes a bit of a stir, without breaking much of mechanics.
Not the highest-priority issue to work on, but we'll need to put it in the list.
That's pretty much it.
Comments and observations are welcome.
Cheers!
Did a thorough testing of current industry functions.
Here are the current results:
Manufacturing - operational. Tested functions was Job start, Delivering, Aborting.
- Job starts as per normal - sets the correct time for both job and assembly line it uses.
- Delivering puts the specified amount of items in the hangar.
- Aborting the task removes it from the jobs list, returns a blueprint, does not return the materials.
So, as a conclusion - manufacturing's fine. There are just a couple notes that will go to the other parts of report.
Copying - functional, a bit bugged.
As blueprints unpacking is not yet functional, there are a few glitches that does not really break the functionality, but look weird.
- 1/1 copy run returns...an actual copy! I mean - you can't see the stats, but icon in item info window gets loaded, and it indeed shown as a copy.
- multiple runs/1 copy returns 1 copy with specified amount of runs.
- 1 run/multiple copies - returns the specified amount of 1-run copies. Problem is - they get stacked, so that leads us to a little sub-test:
- Manufacturing multiple stacked BPC's - If stack was used for manufacturing start, system will subtract the run from EVERY blueprint in that stack. It means, that if you have dozen of 1-run BPC's, you start a 1-run job using them, ALL 10 BPC's will get nullified in process.
- Manufacturing the depleted copy - if copy have no runs left, client will give a "The job cannot be created as the blueprint does not have enough production runs left." pop-up.
- Manufacturing, selecting more runs that copy have - same alert pop-up
- Manufacturing using single-run copy - seems to work. It only glitched once for me - after i installed the job with that BPC, "Get Jobs" button was returning me the empty window. Server had the following client stack trace:
- Code:
15:23:47 W Client::BeanCount: (BeanCount/alert) BeanCount error reporting and handling is not implemented yet.
EXCEPTION #21 logged at 09/01/2015 15:23:46 Unhandled exception in <TaskletExt object at 32235920, abps=1001, ctxt='<NO CONTEXT>^<bound method Button.OnClick of uicls.Button object at (snip)ff5e370, name=button, destroyed=False>>'>
Caught at:
/common/lib/bluepy.py(98) CallWrapper
Thrown at:
/common/lib/bluepy.py(86) CallWrapper
/../carbon/client/script/ui/control/buttons.py(247) OnClick
/client/script/ui/shared/manufacturing.py(1908) LoadJobs
/../carbon/common/script/sys/cfg.py(1215) Get
self = <Instance of Recordset.InvType>
Key column: typeID, Cache entries: 18710
Field names: typeID, groupID, typeName, description, graphicID, radius, mass, volume, capacity, portionSize, raceID, basePrice, published, marketGroupID, chanceOfDuplicating, soundID, iconID, dataID, typeNameID, descriptionID, copyTypeID
key = None
flush = 0
TypeError: int() argument must be a string or a number, not 'NoneType'
Thread Locals: session was <Session: (sid:76554782698568293, clientID:0, mutating:0, locationid:60010927, corprole:0x0, userid:2, languageID:EN, role:0x62f8000280c41000L, charid:140000000, address:10.0.2.2:64574, userType:1, sessionType:5, regionid:10000032, constellationid:20000400, corpid:1000115, stationid:60010927, stationid2:60010927, worldspaceid:60010927, solarsystemid2:30002736, hqID:60011566, rolesAtAll:0x0, rolesAtHQ:0x0, rolesAtBase:0x0, rolesAtOther:0x0, genderID:1, corpAccountKey:1000)>
EXCEPTION END
When delivered, system will return the goods AND the copy. Empty copy, with no runs left, so i just had to trash it.
- Manufacturing, using multi-run copy - in this case, system will subtract the specified runs amount from the overall number. If i have a 3-run copy and i start a 1-run job, system will return 1-run amount of goods and 2-run BPC.
Research - functional, a bit bugged.
Again, due to missing un-packing feature of BP's, it's not crashing the logic, but bugged.
- Research start/delivery/abortion works as per normal.
- Blueprint, indeed, gets it's attributes changed. The problem is - if you stack a researched BP with un-researched one, it seems like BOTH will bestow the same attributes (in my case it was 2 researched BPO's).
Invention - Not coded.
Dead-block is in place - any attempt to start the invention job will result in "The activity requested is not supported." alert pop-up.
So, as a conclusion:
The biggest issue in industry now is a stackable blueprints. In 2 cases of 4, it makes a dupe possible (ME/PE levels for the whole stack and turning a copy into original). In the rest it just causes a bit of a stir, without breaking much of mechanics.
Not the highest-priority issue to work on, but we'll need to put it in the list.
That's pretty much it.
Comments and observations are welcome.
Cheers!