To Do List & Bugs

[Omega]

That reminds me... I noticed that you can no longer pick the font from the INI. The values are still there, but they do nothing, and the fonts are forced. It would be nice to be able to use ANY font. If the name of the font isn't there, it will look in the data folder (where you currently store the fonts).

[silnarm]

a tiny contribution, hopefully the first of many:

Thanks Will, a helpful addition, guess I should really make it so the icons can be sized too...

I admit I couldn't locate where RenderFontMenu and kin in the config are used; it looks like they are unused, right?

Yep, they were used by the old GUI, the new one currently uses hardcoded values, I was planning to make them changeable from widget.cfg, but putting them back in the ini might be wise.

They have been removed from Config in the 0.3.x branch, but that change obviously hasn't made its way back to trunk yet.

What is the appropriate route and place for these tweaks?

patches would be nice, create a ticket on the tracker and attach, or just post them here... If there are many more to come, you'll probably be given SVN access fairly quickly, less work for us applying patches .

That reminds me... I noticed that you can no longer pick the font from the INI. The values are still there, but they do nothing, and the fonts are forced. It would be nice to be able to use ANY font. If the name of the font isn't there, it will look in the data folder (where you currently store the fonts).

As hinted at above, they were removed with the old GUI (which used platform-specific font rendering code), and the intention is that the new ones will be configurable, they just aren't yet 

[the warlord of the reich]

i always dont see bugs in lots of my games. i think i should feel lucky (i hope there wasent bugs that i dident realize they were bugs just parts of the game lol)

 

[Omega]

Oh, and I was also wondering, when (if ever) can we expect to see Subfactions working? I know the AI topic has come up a lot in the past, but I'm not looking for some overhaul, simply to get the AI to use subfactions (which, are used exactly like upgrades. Though they'd have to randomly pick a subfaction). With transport units implemented, this is the LAST lil step separating us from being able to do FPM (although game play wise, this is THE game breaker). I think MRising also uses subfactions, and they should be a high priority, since it's used in mods, it isn't overly complex, and is broken existing feature (one that has, abet, been broken for as long as I remember). It would be nice to see FPM revived, after this last step...


In passing, who deleted the topic about that new google image format? It was requested to be locked due to offtopic, but when I tried to lock it, I found it deleted... Just as a note, unless it's spam or flat out trolling, please don't delete topics. Lock them, and let them remain for archival purposes. It's crucial to note that there is NO WAY TO RECOVER DELETED TOPICS. Unlike a file on your computer or a page on the wiki, there is no recycling bin on our board (and never will be, since 99.9% of deleted topics are porn, ads, or various other spam, and the only "recycling bin" like function there currently is is simply another board where deleted topics go (and if deleted from there, they are gone for good). It's not a problem this once, but please try and keep this in mind.

[Psychedelic_hands]

Oh, and I was also wondering, when (if ever) can we expect to see Subfactions working? I know the AI topic has come up a lot in the past, but I'm not looking for some overhaul, simply to get the AI to use subfactions (which, are used exactly like upgrades. Though they'd have to randomly pick a subfaction). With transport units implemented, this is the LAST lil step separating us from being able to do FPM (although game play wise, this is THE game breaker). I think MRising also uses subfactions, and they should be a high priority, since it's used in mods, it isn't overly complex, and is broken existing feature (one that has, abet, been broken for as long as I remember). It would be nice to see FPM revived, after this last step...

Yes, Constellus also plans to use Subfactions for a few things I think. When is the AI overhaul planed to happen?

[silnarm]

... When is the AI overhaul planed to happen?

Now, and over the next month or two probably.

[will]

Building 0.3.2 from source, I have some diverse questions:

0) the bz2 source contains a CMakeCache.txt file; you want to strip that, people will have to know to delete it before building otherwise since it seems to think the user name is James ;

1) unpacking megapack_v4/glest/* into addons/megapack_v4/, the scenarios didn't turn up.  Has the directory name for the scenarios changed?

2) trying to run the 'benchmark1' (MG?) scenario that I found in another thread, it couldn't find tech/megapack/resources, so I symlinked tech/megapack_v4 to tech/megapack and got it to go further, but then it crashes.

My question is twofold then:

2a) how to benchmark graphics performance?  (I want to track it before and after any changes I introduce)

2b) how to get debug symbols in the gae-crash.txt?  My gae-crash.txt is:

Code: [Select]

Crash
Version: Advanced Engine 0.3.2_dev_edition
Time: Mon Dec 13 09:39:26 2010
Description: SIGSEGV: address not mapped to object
Address: 0x8
Backtrace:
glestadv(_ZN6Shared8Platform24PlatformExceptionHandler7handlerEiP7siginfoPv+0x33a) [0x881096]
/lib/libpthread.so.0(+0xfb40) [0x7f68406deb40]
glestadv(_ZN5Glest10ProtoTypes11FactionType4loadEiRKSsPKNS0_8TechTreeE+0xb3) [0x77c55f]
glestadv(_ZN5Glest10ProtoTypes8TechTree4loadERKSsRKSt3setISsSt4lessISsESaISsEE+0x59b) [0x737699]
glestadv(_ZN5Glest3Sim5World8loadTechEv+0xdc) [0x8196a6]
glestadv(_ZN5Glest3Sim19SimulationInterface9loadWorldEv+0x129) [0x838d13]
glestadv(_ZN5Glest3Gui9GameState4loadEv+0x3e5) [0x65956d]
glestadv(_ZN5Glest4Main7Program8setStateEPNS0_12ProgramStateE+0x72) [0x6decee]
glestadv(_ZN5Glest4Menu17MenuStateScenario10launchGameEv+0x175) [0x70ef11]
glestadv(_ZN5Glest4Menu17MenuStateScenario6updateEv+0xe0) [0x70ed90]
glestadv(_ZN5Glest4Menu8MainMenu6updateEv+0x8c) [0x71141e]
glestadv(_ZN5Glest4Main7Program4loopEv+0x1d1) [0x6de5b9]
glestadv(_ZN5Glest4Main9glestMainEiPPc+0x377) [0x6dabd0]
glestadv(main+0x72) [0x6dae06]
/lib/libc.so.6(__libc_start_main+0xfe) [0x7f683ec56d8e]
glestadv() [0x5f00a9]

I edited CMakeList.txt and

Code: [Select]

set option(GAE_DEBUG_EDITION "Enable debug edition extensions." ON) and I then rebuilt clean and I can see my CMakeCache.txt includes

Code: [Select]

CMAKE_CXX_FLAGS_RELWITHDEBINFO:STRING=-O2 -g but no debug symbols in the stack trace anyway

[silnarm]

0) the bz2 source contains a CMakeCache.txt file; you want to strip that, people will have to know to delete it before building otherwise since it seems to think the user name is James

My fault, I must have accidentally generated with build dir == source dir and failed to clean up properly, fixed and new package up.

1) unpacking megapack_v4/glest/* into addons/megapack_v4/, the scenarios didn't turn up.  Has the directory name for the scenarios changed?

Yes, scenarios a structured differently, and located in gae/scenarios/category/scenario/,
Where 'category' and 'scenario' are the category ('tutorials', a campaign, or whatever) and specific scenario.

2) trying to run the 'benchmark1' (MG?) scenario that I found in another thread, it couldn't find tech/megapack/resources, so I symlinked tech/megapack_v4 to tech/megapack and got it to go further, but then it crashes.

My question is twofold then:

2a) how to benchmark graphics performance?  (I want to track it before and after any changes I introduce)

MG and GAE scenarios are not compatible.
Best bet would be to work in trunk, where I have recently added configurable (only 'in code' atm) performance stats to the debug info.

2b) how to get debug symbols in the gae-crash.txt?  My gae-crash.txt is:
...
I edited CMakeList.txt and

Code: [Select]

set option(GAE_DEBUG_EDITION "Enable debug edition extensions." ON) and I then rebuilt clean and I can see my CMakeCache.txt includes

Code: [Select]

CMAKE_CXX_FLAGS_RELWITHDEBINFO:STRING=-O2 -g but no debug symbols in the stack trace anyway

You're complaining about that stack trace ?!? It's full of symbols, they are c++ mangled, but it's all there!  Crash apparently occurred in FactionType::load()

Translation:

Code: [Select]

glestadv(_ZN5Glest10ProtoTypes11FactionType4loadEiRKSsPKNS0_8TechTreeE+0xb3) [0x77c55f] == Glest::ProtoTypes::FactionType::load()

GAE_DEBUG_EDITION isn't for that kind of debugging, it adds a bunch code to help visualise and debug the game/world/mechanics, stuff that is hard to debug with a traditional debugger. If you want to measure performance, best turn this off

[will]

Oh I'm not complaining about the stacktrace, was just thinking I was missing something obvious

My profiling sadly says my cpu spends all its time in an opaque blob called i915drm which isn't so descriptive

[hailstone]

I think i915drm is a kernel module.

[silnarm]

I think i915drm is a kernel module.

OpenGL driver.

If you're using the 0.3.2 source, you can turn on the code profiler in shared/util/profiler.h,

then you just need to put the macro _PROFILE_FUNCTION() at the start of Renderer::renderUnits().

Most of the other instances of that macro should be commented out by now, but there may be some strays... output is written to $(ConfigDir)/profiler.log

But I still recommend svn co trunk, where you can see this on-screen in-game.

[will]

In the svn trunk 1049, if you triple or quadruple click in quick succession on a unit command icon you crash; e.g. click on a worker, click on the build command, and then tripple-click on the farm icon quickly.

it seems that:

Code: [Select]

glestadv(Glest::Entities::Faction::reqsOk(Glest::ProtoTypes::CommandType const*) const+0x14) [0x61c3f6]
glestadv(Glest::Gui::UserInterface::mouseDownDisplayUnitSkills(int)+0x350) [0x6ab154]
glestadv(Glest::Gui::UserInterface::commandButtonPressed(int)+0x84) [0x6a889e]
glestadv(Glest::Gui::Display::mouseUp(Shared::Platform::MouseButton, Shared::Math::Vec2<int>)+0x28f) [0x6bf1c7]
glestadv(Glest::Widgets::WidgetWindow::eventMouseUp(int, int, Shared::Platform::MouseButton)+0x9a) [0x8021e4]
glestadv(Shared::Platform::Window::handleEvent()+0x153) [0x882989]
glestadv(Glest::Main::Program::loop()+0x2b7) [0x6d06a5]
glestadv(Glest::Main::glestMain(int, char**)+0x35c) [0x6cca71]
user_interface.cpp:872 in mouseDownDisplayUnitSkills(), m_display->getCommandType(posDisplay) can return NULL.

I put a printf in there and confirmed that, but simply avoiding that dereference only makes more pop up instead and it still crashes only later:

Code: [Select]

glestadv(Glest::Gui::UserInterface::giveOneClickOrders()+0x84) [0x6a9f6c]
glestadv(Glest::Gui::UserInterface::mouseDownDisplayUnitSkills(int)+0x5e9) [0x6ab3ed]
glestadv(Glest::Gui::UserInterface::commandButtonPressed(int)+0x84) [0x6a889e]
Therefore, I guess there is a deeper bug in how the command type can be NULL in the first place?

A simple return stops the crash, but is this just hiding a deeper problem?

Code: [Select]

} else {
const Unit *unit= selection.getFrontUnit();

if (selection.isUniform()) { // uniform selection
if (!m_display->getCommandType(posDisplay)) {
std::cerr << "WTF" << std::endl;
return;
}
if (unit->getFaction()->reqsOk(m_display->getCommandType(posDisplay))) {
activeCommandType = m_display->getCommandType(posDisplay);
activeCommandClass = activeCommandType->getClass();
} else {
posDisplay = invalidPos;

[hailstone]

commandTypes[ i ]= NULL; in Display::clear() is the reason why it is returning NULL. clear is called from UserInterface::ComputeDisplay. The only place commandTypes are being set is in ComputeDisplay if not selecting second menu. ComputeDisplay is being called a second time before it gets a chance to go back to first menu to refill commandTypes.

Either putting m_display->setCommandType(activePos, activeCommandType); in two-tier select or commenting out commandTypes[ i ]= NULL; seemed to work but I suspect they aren't ideal solutions.

PS: UserInterface needs to go on an exercise plan.

[silnarm]

Nice catch Will, I've just doctored it up by checking the active command type in the if condition, which is much the same as as just returning, except it resets UserInterface::activeCommandType & UserInterface::activeCommandClass, which would have still been set from the first click (?!?).

I wouldn't dwell on this one too much...

PS: UserInterface needs to go on an exercise plan.

It also has an unhealthy relationship with Display.  The Display class needs re-implementing, while I changed it to be a Widget, it's still very similar to how it was, it is not a Container widget like it should be.

After I invert the Y axis on the Widget system I'll spend some time on it.

[will]

I'm trying to play the new Imperial Beta

But I can't load it:

Code: [Select]

glestadv: /home/will/Games/glestae-src-0.3.2/source/game/graphics/renderer.cpp:2497: void Glest::Graphics::Renderer::renderProgressBar(int, int, int, int, int, const Shared::Graphics::Font*): Assertion `progress <= maxProgressBar' failed.

will@ubuntu:~/.glestadv$ c++filt < gae-crash.txt
Crash
Version: Advanced Engine 0.3.2
Time: Sat Mar 26 00:01:06 2011
Description: SIGABRT:
Address: 0x3e8000020ad
Backtrace:
glestadv(Shared::Platform::PlatformExceptionHandler::handler(int, siginfo*, void*)+0x33a) [0x85cffe]
/lib/libpthread.so.0(+0xfb40) [0x7f12c9a36b40]
/lib/libc.so.6(gsignal+0x35) [0x7f12c7fc3ba5]
/lib/libc.so.6(abort+0x180) [0x7f12c7fc76b0]
/lib/libc.so.6(__assert_fail+0xf1) [0x7f12c7fbca71]
glestadv(Glest::Graphics::Renderer::renderProgressBar(int, int, int, int, int, Shared::Graphics::Font const*)+0x3f) [0x676809]
glestadv(Glest::Util::Logger::renderLoadingScreen()+0x3c2) [0x6377c0]
glestadv(Glest::Util::Logger::add(std::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, bool)+0x339) [0x636fb9]
glestadv(Glest::ProtoTypes::UnitType::load(std::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, Glest::ProtoTypes::TechTree const*, Glest::ProtoTypes::FactionType const*, bool)+0x6e) [0x73e52c]
glestadv(Glest::ProtoTypes::FactionType::load(int, std::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, Glest::ProtoTypes::TechTree const*)+0x71e) [0x758586]
glestadv(Glest::ProtoTypes::TechTree::load(std::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::set<std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::less<std::basic_string<char, std::char_traits<char>, std::allocator<char> > >, std::allocator<std::basic_string<char, std::char_traits<char>, std::allocator<char> > > > const&)+0x59b) [0x710b25]
glestadv(Glest::Sim::World::loadTech()+0xdc) [0x7f5986]
glestadv(Glest::Sim::SimulationInterface::loadWorld()+0x129) [0x814ff3]
glestadv(Glest::Gui::GameState::load()+0x3e5) [0x6486c5]
glestadv(Glest::Main::Program::setState(Glest::Main::ProgramState*)+0x72) [0x6b847e]
glestadv(Glest::Menu::MenuStateNewGame::update()+0x8c5) [0x6d836d]
glestadv(Glest::Menu::MainMenu::update()+0x8c) [0x6eab9e]
glestadv(Glest::Main::Program::loop()+0x1d1) [0x6b7d49]
glestadv(Glest::Main::glestMain(int, char**)+0x37a) [0x6b4297]
glestadv(main+0x72) [0x6b4587]
/lib/libc.so.6(__libc_start_main+0xfe) [0x7f12c7faed8e]
glestadv() [0x5e2a89]

will@ubuntu:~/.glestadv$ cat glestadv-error.log
Glest Advanced Engine: Error log file.

No files found: techs/great_empires/factions/imperial/upgrades/*.
XML Error in techs/great_empires/factions/imperial/units/citizen:
Node "repaired-units" doesn't have 8 children named  "unit"

Tree: repaired-units (unit unit unit unit unit unit unit building )
XML Error in techs/great_empires/factions/imperial/units/knight:
Upgrade not found: training_field

will@ubuntu:~/.glestadv$ cat glestadv.log
Glest Advanced Engine: Program log file.

80: Core data
3490: TechTree: techs/great_empires
3500: Resource type: techs/great_empires/resources/energy
3500: Resource type: techs/great_empires/resources/food
3510: Resource type: techs/great_empires/resources/gold
3520: Resource type: techs/great_empires/resources/stone
3520: Resource type: techs/great_empires/resources/wood
3520: Faction type: techs/great_empires/factions/imperial
3540: Unit type: techs/great_empires/factions/imperial/units/aeronautical_facility
3560: Unit type: techs/great_empires/factions/imperial/units/citizen
3580: Unit type: techs/great_empires/factions/imperial/units/garrison
3600: Unit type: techs/great_empires/factions/imperial/units/generator
3610: Unit type: techs/great_empires/factions/imperial/units/imperial_cruiser
3630: Unit type: techs/great_empires/factions/imperial/units/ironclad
3640: Unit type: techs/great_empires/factions/imperial/units/keep
3680: Unit type: techs/great_empires/factions/imperial/units/kite_fighter
3700: Unit type: techs/great_empires/factions/imperial/units/knight
3720: Unit type: techs/great_empires/factions/imperial/units/man_at_arms
3750: Unit type: techs/great_empires/factions/imperial/units/packed_trebuchet
3780: Unit type: techs/great_empires/factions/imperial/units/ship_yard
3800: Unit type: techs/great_empires/factions/imperial/units/tavern
3810: Unit type: techs/great_empires/factions/imperial/units/tower
3850: Unit type: techs/great_empires/factions/imperial/units/trebuchet
3880: Unit type: techs/great_empires/factions/imperial/units/wall
3900: Unit type: techs/great_empires/factions/imperial/units/war_factory
3940: Unit type: techs/great_empires/factions/imperial/units/war_machine
3980: Unit type: techs/great_empires/factions/imperial/units/worker
3990: Faction type: techs/great_empires/factions/tech
4000: Unit type: techs/great_empires/factions/tech/units/aerodrome
4020: Unit type: techs/great_empires/factions/tech/units/air_ballista
4040: Unit type: techs/great_empires/factions/tech/units/airship
4070: Unit type: techs/great_empires/factions/tech/units/archer
4090: Unit type: techs/great_empires/factions/tech/units/barracks
4100: Unit type: techs/great_empires/factions/tech/units/battle_machine
4140: Unit type: techs/great_empires/factions/tech/units/blacksmith
4160: Unit type: techs/great_empires/factions/tech/units/castle
4170: Unit type: techs/great_empires/factions/tech/units/catapult
4200: Unit type: techs/great_empires/factions/tech/units/cow
4220: Unit type: techs/great_empires/factions/tech/units/defense_tower
4240: Unit type: techs/great_empires/factions/tech/units/dreadnought
4260: Unit type: techs/great_empires/factions/tech/units/farm
4280: Unit type: techs/great_empires/factions/tech/units/guard
4300: Unit type: techs/great_empires/factions/tech/units/horseman
4550: ~GameState


And glestadv-crash.tga looks like its stopped with loading the horseman (or maybe that succeeded, just was the last thing drawn)

[John.d.h]

Judging from this:

Code: [Select]

No files found: techs/great_empires/factions/imperial/upgrades/*.
XML Error in techs/great_empires/factions/imperial/units/citizen:
Node "repaired-units" doesn't have 8 children named  "unit"

Tree: repaired-units (unit unit unit unit unit unit unit building )
XML Error in techs/great_empires/factions/imperial/units/knight:
Upgrade not found: training_field
That sounds like something wrong with the mod, not the engine.

[John.d.h]

I can get it to load in 0.3.2, but not in git-master, so maybe I'm wrong about it being only the mod's fault.

My error log: http://pastebin.com/JgFy4EhM

[hailstone]

so maybe I'm wrong about it being only the mod's fault.

It's a bit of both. Somehow the errors in the mod caused more units to be loaded than it thought existed triggering the assert in renderProgressBar (ie at the time of the crash 33 unitsLoaded but only 32 totalUnits resulting in 103%) so I think there could be improvement in error handling.

To get it working:
- In citizen.xml change <building name="aeronautical_facility"/> to <unit name="aeronautical_facility"/>
- Copy advanced_architecture and training_field from tech
- In training_field.xml remove barracks so it is just <effects/>

Alternatively remove training_field as a requirement for knight, although there needs to be at least one upgrade currently (added ticket #325).

[Mr War]

hopefully all fixed now on the Mod side. Re-uploaded  http://www.mediafire.com/download.php?myu13a1kkt8hqnv