Page 1 of 3

Feb 01 - Initial testing / installation

Posted: Mon Feb 02, 2015 4:40 am
by mharmless
Started off by trying installation on multiple devices. Here is a short profile of the devices and installs.

1) HP Laptop
1.OS) Windows 7 SP1, 64 bit
1.Spec) AMD A6-3400M processor, 6gb of RAM, AMD A6-3400M APU graphics, 750gb conventional drive
1.Res) 1600x900, also tested with 1600x1200 external display.
1.Inst) Installed to the conventional drive, at the location "c:\users\mharmless\downloads\stars in shadow"
1.Notes) Installation went fine. Update screen was visible and went through all its steps. Upon launching the game, the centered Stars in Shadow logo is seen, and then the display fills with a solid white background and a blue arrowhead cursor. Moving the cursor around shows a change into the standard windows cursor at three locations. Middle one goes to a load game screen, and appears to be fine. Bottom goes to an option screen, which also appears to be fine. Exiting either of these appears to return the invisible main menu. Topmost option leads to another white screen. You can tell it is different because the cursor hotzones do not appear at all. Covered the display in roughly 50 pixel increments without finding a cursor change spot. Behavior is identical on both displays.


2) Primary PC
2.OS) Windows 8.1, 64 bit
2.Spec) Intel Core i7-4770, 16gb of RAM, AMD HD7850 graphics, 240gb SSD
2.Res) Two displays, 1600x1200 primary, 1200x1600 secondary.
2.Inst) Installed to the SSD drive, at the location "c:\games\stars in shadow"
2.Notes) Installation went fine. Update screen was visible and went through all its steps. Upon launching the game, received dialog:
Title:

Code: Select all

Error during program startup

Icon:

Code: Select all

Yellow Exclamation

Body:

Code: Select all

error in image callback.
Lua state\Drawers\Planets\Clouds\PlanetClouds.lua:43:glActiveTextureARB(GL_TEXTURE0 + 8) failed: GL_INVALID_VALUE
 Lua state\Drawers\Planets\Clouds\PlanetClouds.lua:43:
   Lua state\Drawers\Planets\Clouds\PlanetClouds.lua:40:
     Lua state\Drawers\Planets\Clouds\PlanetClouds.lua:37:
       Lua state\Drawers\Planets\Clouds\PlanetClouds.lua:142:
         Lua state\Drawers\Planets\planet_drawer_env.lua:471:fun:Arid


Attempted a relaunch via the updater, same behavior. Skipping launcher and going from sis.exe, same behavior. Started asset check, completed without incident, same error.

After success on PC 4 below, Media PC, I forced the primary display to 1280x800 and attempted to launch again. Same behavior.


3) Tablet PC
3.OS) Windows 8.1 with Bing, 32 bit
3.Spec) Intel Atom Z3735F, 1gb of RAM, Intel HD Graphics (Ivy Bridge), 16gb MMC, 16gb class 10 SD card
3.Res) 1280x800. HDMI out, did not test on a secondary display. Did not try portrait orientation.
3.Inst) Installed to the SD card, root of the drive "d:\stars in shadow"
3.Notes) Installation went fine. Update screen was visible and went through all its steps. Upon launching the game, received dialog:
Title:

Code: Select all

Error during program startup

Icon:

Code: Select all

Yellow Exclamation

Body:

Code: Select all

attempt to index a nil value
Lua state\GUI\~GalaxyMap\@GalaxySkin1\PlanetBar1.lua:28:PlanetBar:
 Lua state\GUI\~GalaxyMap\GalaxyMap.lua:30


Skipped launcher, same behavior. Launched updater again, now I'm getting the updater with background graphics and the drop window where text goes, no text. No buttons along the bottom either. Closing the window showed a brief flash of text, so relaunched updater and clicked approximately where the Update button should be. This caused everything to refresh and display correctly. Indicates up to date. Hitting asset check causes the same error outlined above. Will come back to this later, can't try it on the primary drive without clearing some space.


4) Media PC
4.OS) Windows 8.1 Pro, 64 bit
4.Spec) AMD A10-6800K, 8gb of RAM, AMD A10-6800K APU graphics, 120gb SSD
4.Res) 1280x800 on a projector.
4.Inst) Installed to the SSD, "C:\games\stars in shadow"
4.Notes) Installation went fine. Update screen was visible and went through all its steps. Upon launching the game, got to the main menu. I can see now that there was an Exit to Windows option that I never got, plus the three I found with cursor wandering.


Going to go ahead and play now on the Media PC and see what I can see. I'll try some more PCs either later tonight or after work tomorrow.

Re: Feb 01 - Initial testing / installation

Posted: Mon Feb 02, 2015 7:09 am
by mharmless
Typed notes to myself while playing, here's a cleaned up version of that. I've bolded items which could legitimately be considered bugs, and not simply interface preferences. This was done on PC4 from the above list.


Planet in the center left of the main menu looks like a disco ball, kaleidoscope of blue/white squares. Rotates approximately 10 degrees and then resets itself. New game. Race selection, they all have a disco ball. Humanity is at the bottom, and it is the only one that isn't a disco ball. I can see now that this is supposed to be the star.

Some racial text does not fit in the highlight region of the race selection area.
Ashdar Colonials cuts off "empire."
Gremak Empire cuts off "dinner."
Humanity cuts off "Galaxy."
Human text ends up cutting into the bottom UI element, with the "y" overlapping. No other racial text collisions.


I select Humanity. I like having a star. Also, HFY.

System Fargone. Arid planet. 4 people. Ships seem intuitive. 2 colony ships, 4 transports, one scout. Right click to deselect units for movement, also seems obvious. I have a second planet, gas giant by look. Can't click it for any info.

Sending scout to nearby star. Blue giant

+ sign by stardate to end turn. Took me a couple of seconds of looking. Only figured out due to the clean/sparse UI. Lack of elements to examine. + a stardate, got it. Could be more obvious, tiny target.

Blue giant has only poor worlds. Sensible. Moving to an orange star. Slaver encounter. I'm tempted to pay tribute, but that isn't an option. DIE SLAVER SCUM. System view. Attack planet. We have orbital control, no fight.

Decided to disembark two population units onto the Arid world in Fargone. Max is 4. 1 there now, room for growth. Each disembark was clumsy. It doesn't go back to my fleet view, leaves me at system view. Have to exit and go get to the next transport. Perhaps back to fleet view after a disembarkation?

Sending a now-empty transport to a red dwarf in range. Scout is locking down the pirate system.

Found another pirate group over the red dwarf. Castor II system. This turn I also lost a colony ship due to being broke. Scrapped the empty one at Fargone, then scrapped the one over Castor II. This crashed the game. I sent bug report. Related to scraping a ship used in an attack? Will dig more later.

Restarted the game. Gave me a new race. Exiting to main menu, I'm HFYing some more.

Fargone again, same start. Makes sense.
Sent ship to nearby yellow star. Fight! Nice little tactical battle engine. Clean, good looking. Captured enemy destroyer with Marines. HFY! Star is a disco ball, just like at race selection. Sending colony ship & transport to the new system. Arid world here, might as well take it.

Inferno world. Scrolling red text, handwritten, "Lights". Visible just as the edge of the inferno reaches the right part of the disc.

War with an AI. Birdmen looks like. Second ship reached system Alioth. There's a fight to be had here, but we crashed again. Bug report sent in. Unsure what the trigger was.

Restart game. Keeping the race I was delt this time. Teros.
Text of the Restaban system has the "undefined character" box where there should be an apostrophe. "Rastaban's geology reveals a long history" is the surrounding text. Second instance, same problem, around "weather of Rastaban's skies,"

Found slavers. This time I had a bribe option. Nope, war. Brought in colony ship, colonized arid in same system, set transport to work. Researched lasers. Now working on small craft

Designing my own ship. Dragging lasers over, they appear to highlight the box above where the mouse is located. Dropping onto Turret places it in Medium, dropping on System gets it into turret.

Built a transport, loaded a person, and sent it to my eternal blockade of Sargas. Nope, that's not how you ground attack. Ground combat unimplemented possibly? Been researching looking specifically for some sort of invasion tech.

First real fight. Got blown up. Reactor hit. Destroyer down! Found bugmen. They bombed a colony I made. Need a fleet. Genocide on all the marauaders I'm camping over, to free up fleets. Congregate fleets, move system to system in a pack of 5 till I found their homeworld.

Used a scout ship self destruction to destroy their entire fleet in one go. Used my remaining combat ships to clear the station rapidly. Easy win. Scoutship self destruction FAR nastier than any weapons I have access to.

Bomb them into oblivion. Start taking worlds for myself. Found a Gaia world. These people integrated after colonization. 0 money, 0 industry, 2 science. Nice. Might have to see about moving them around to science worlds.

Colonized Bacabs IV, Iceball world. It came with 3 factories and 2 labs, the remains of the bugmen.

I scrapped a lab, and did not get a new spot to put a factory in. That makes sense sort of, iceballs are bad for me. Scrapped facilites till I saw a build spot. 2 spots, same as I'd expect from other unsuited worlds. Perhaps a world with too many facilities should scrap some when colonized by a race with less spots for that type?

Planet Grumium had a debug tag of special Hammers = 50, did not appear to do anything special. And just like that the night is gone.


Tl;DR Game feels great, even now. Can't wait till work is over tomorrow.

Re: Feb 01 - Initial testing / installation

Posted: Mon Feb 02, 2015 7:52 am
by Arioch
Thanks for the detailed reports. It's great that you were able to find so many issues on such a wide variety of systems.

Tactical combat is badly unbalanced, as I'm sure you've noticed. We didn't get a chance to make a much-needed pass at it before the beta deadline.

For information on how to invade planets (which isn't obvious), see my post in the Getting Started thread.

Planetary infrastructure is currently immune to destruction by orbital bombardment and invasion; this will almost certainly change. However, the situation in which a population cap change resulted in fewer infrastructure slots should be accounted for by the "underpopulation" mechanic; a planet can only work a factory or lab that they have enough population for, so the extra labs shouldn't be doing anything for you. I think it's better that the buildings stay in place, so that you can sell them, or they can be reused if you move more suitable population in and the population cap goes back up. There should probably be some UI hint that lets you know when a building is unusable due to underpopulation.

Re: Feb 01 - Initial testing / installation

Posted: Mon Feb 02, 2015 3:30 pm
by mharmless
Came in early to test the installation on my work PC.

5) HP Elitedesk 800 GI
5.OS) Windows 8.1 Enterprise, 64 bit
5.Spec) Intel i7-4770 processor, 16gb of RAM, Intel HD Graphics 4600 and ATI Radeon HD 5400 graphics, 120gb SSD, 1tb conventional drive
5.Res) 3 displays at 1200x1600, arranged side to side, one display at 1680x1050 above these. Primary is center 1200x1600 portrait display.
5.Inst) Installed to the conventional drive, at the location "d:\test\Stars in Shadow".
5.Notes) Installation went fine, no problems caused by our firewall, no blockages. Graphical distortion at the top of the main menu, and it looks like an odd bit of graphical tearing underneath the menu as well, near the copyright notice. Programmatically resized/inserted into background?. The primary distortion at the top looks like a continuous repeat of the topmost properly drawn pixels from the image.

Can't attach the image, file too large. Even the jpg varient was too large at 566kb. Uploaded to imgur instead.
Image

Re: Feb 01 - Initial testing / installation

Posted: Mon Feb 02, 2015 3:41 pm
by mharmless
And now in an Orcale VirtualBox

6) Oracle VirtualBox on PC 5
6.OS) Windows 7 Professional SP 1, 32 bit
6.Spec) Intel i7-4770 processor, 2gb of RAM, 32mb VirtualBox Graphics without Acceleration, 40gb VDI
6.Res) 1 adjustable display, tested at 1032x777
6.Inst) Installed it onto a mapped network drive. "E:\Stars in Shadow". Location maps to the host machine.
6.Notes) connecting to update.stars-in-shadow.com:80
updater\connect_to_server.lua:165: wrong version number.


An update just pushed perhaps? Regardless, time to get to work.

Re: Feb 01 - Initial testing / installation

Posted: Mon Feb 02, 2015 3:52 pm
by sven
mharmless wrote:Came in early to test the installation on my work PC.

Graphical distortion at the top of the main menu, and it looks like an odd bit of graphical tearing underneath the menu as well, near the copyright notice. Programmatically resized/inserted into background?. The primary distortion at the top looks like a continuous repeat of the topmost properly drawn pixels from the image.

Can't attach the image, file too large. Even the jpg varient was too large at 566kb. Uploaded to imgur instead.



Ok -- these, at least, are easy ones. The start-screen graphic has known issues at many display resolutions -- it's all perfectly fixable, just haven't gotten around to it yet. And I've attempted to boost the allowances for forum image uploads, so posting error screens should become easier in future.

Re: Feb 01 - Initial testing / installation

Posted: Mon Feb 02, 2015 3:58 pm
by sven
mharmless wrote:updater\connect_to_server.lua:165: wrong version number.[/b]

An update just pushed perhaps? Regardless, time to get to work.


This looks like a firewall / security software issue. (More exactly, this may be a case of a security program trying to insert itself into the updater's TLS connection, and getting the protocol wrong.)

Try hitting the "alt-port" button on the updater, and see if that helps. (These kinds of "protocol monitoring" tricks are often only used over port-80.)

Also -- can you tell me what security software is active on this machine?

Re: Feb 01 - Initial testing / installation

Posted: Mon Feb 02, 2015 4:09 pm
by mharmless
ESET Endpoint AV on all machines. We use Watchguard as our firewall/packet filter, so primary machine working and VM not working would make some sense. Primary machine is exempt from the monitoring, and the VM is not. All traffic from it would be ripped apart and examined. I'll try to make time at lunch to exempt the VM and give it another go.

Re: Feb 01 - Initial testing / installation

Posted: Mon Feb 02, 2015 4:11 pm
by sven
mharmless wrote:Planet in the center left of the main menu looks like a disco ball, kaleidoscope of blue/white squares. Rotates approximately 10 degrees and then resets itself. New game. Race selection, they all have a disco ball. Humanity is at the bottom, and it is the only one that isn't a disco ball. I can see now that this is supposed to be the star.


Ok -- this is exactly the category of behaviors I was worried about. Nearly all the alpha testing has been done on either NVIDIA cards or intel integrated chipsets.

Basically, it appears that *all* of your AMD cards are breaking, but, amusingly, they're all breaking in different ways. What your seeing here are shader compilation bugs, which I can probably work around by inserting special-case code. But, you've tested on so many different machines now that my logging hacks are being a bit overwhelmed. When you get home, would you upload the entire contents of the Stars in Shadow\logs directory from this media-center machine? That way, I can easily write a fallback case that will trigger for your hardware, without interfering with the builds for the testers working with NVIDIA/intel cards.

Re: Feb 01 - Initial testing / installation

Posted: Mon Feb 02, 2015 4:20 pm
by sven
1.Spec) AMD A6-3400M processor, 6gb of RAM, AMD A6-3400M APU graphics, 750gb conventional drive

...

and then the display fills with a solid white background and a blue arrowhead cursor



This is a different category of graphics bugs. Probably a quirky OpenGL implementation in the AMD integrated graphics chip. And, again, I'll need to the contents of your entire Stars in Shadow\logs directory to look into it. One easy test here would be to try running at a lower screen resolution -- it may just be a case of conservative texture allocation rules.

Re: Feb 01 - Initial testing / installation

Posted: Mon Feb 02, 2015 4:25 pm
by sven

Code: Select all

error in image callback.
Lua state\Drawers\Planets\Clouds\PlanetClouds.lua:43:glActiveTextureARB(GL_TEXTURE0 + 8) failed: GL_INVALID_VALUE
 Lua state\Drawers\Planets\Clouds\PlanetClouds.lua:43:
   Lua state\Drawers\Planets\Clouds\PlanetClouds.lua:40:
     Lua state\Drawers\Planets\Clouds\PlanetClouds.lua:37:
       Lua state\Drawers\Planets\Clouds\PlanetClouds.lua:142:
         Lua state\Drawers\Planets\planet_drawer_env.lua:471:fun:Arid


Opps. Yet a third category of graphics errors, but, this one, in theory, ought to easy be to fix. And as this is your primary gaming PC, with a card that the game really *needs* to support, of all the graphics bugs, I think this one's our top priority. Again, I could use the contents of your entire logs directory, but, there's a chance I'll have an experimental patch up by the time you get home.

Re: Feb 01 - Initial testing / installation

Posted: Mon Feb 02, 2015 4:32 pm
by sven
CODE: SELECT ALL
attempt to index a nil value
Lua state\GUI\~GalaxyMap\@GalaxySkin1\PlanetBar1.lua:28:PlanetBar:
Lua state\GUI\~GalaxyMap\GalaxyMap.lua:30


Skipped launcher, same behavior. Launched updater again, now I'm getting the updater with background graphics and the drop window where text goes, no text. No buttons along the bottom either. Closing the window showed a brief flash of text, so relaunched updater and clicked approximately where the Update button should be. This caused everything to refresh and display correctly. Indicates up to date. Hitting asset check causes the same error outlined above. Will come back to this later, can't try it on the primary drive without clearing some space.


My working hypothesis with this one is that we're looking at some kind of disk-write problem. The behaviors could be explained by missing / corrupted files. As always, uploading the whole logs directory would give me more data to work with. (I've run the alpha on Win8 tablets in the past without trouble, so I'm hopeful that this is one that may have a simple fix.)

Re: Feb 01 - Initial testing / installation

Posted: Mon Feb 02, 2015 4:37 pm
by mharmless
I'll pack up all four log folders regardless of a test patch, so you can confirm suspicions about the behavior. Then I'll retry them all if there is a new update by then. Assuming the board's upload limits allow, I'll start defaulting to attaching a compressed copy of the entire logs folder with future posts.

I've also got a few other PCs at home that I hadn't tried the install on yet, I'll get to that tonight as well.

Regarding cards needing support, the AMD APUs are actually remarkably beefy and I wouldn't discount them. My media PC is able to play the new tomb raider and look very sexy doing it on nothing but that APU, and the laptop specs listed for PC1 has been my primary bit of mobile gaming for several years. Routinely play Civ5 and Distant Worlds on it without issues. They really are remarkable chips for the budget gaming options, completely disproportionate power relative to cost.

I've heard of people having unique issues with AMD's dual-graphics mode where you work the APU alongside an equivalent dedicated card. I'll see about rustling that up once these earlier issues start getting delt with.

Re: Feb 01 - Initial testing / installation

Posted: Mon Feb 02, 2015 4:45 pm
by sven
Regarding cards needing support, the AMD APUs are actually remarkably beefy and I wouldn't discount them.


Yeah -- the problem isn't the capacity of the chip -- the problem is that I'm using OpenGL for graphics, rather than DirectX, as almost all windows games do. There are pros and cons to that choice -- but, the huge CON is what you're seeing right now: OpenGL support varies from card to card, and that means the game needs special case code that can toggle potentially buggy graphics features on or off in response to vendor and version strings. We should be able to get the game running on all your AMD chips with a bit of fiddling, but, it's going to take some experimentation.

Re: Feb 01 - Initial testing / installation

Posted: Mon Feb 02, 2015 4:57 pm
by sven
Decided to disembark two population units onto the Arid world in Fargone. Max is 4. 1 there now, room for growth. Each disembark was clumsy. It doesn't go back to my fleet view, leaves me at system view. Have to exit and go get to the next transport. Perhaps back to fleet view after a disembarkation?


Yeah, the transport UI is still very raw -- coded it up under a deadline, and it needs polish. Resetting to fleet view is an easy improvement to make though. Posted a little patch to the dev distro.

(Also, I should write a little post explaining how to toggle between the dev/stable distro options..)