Messages

Still seems to persist. No rush to fix this though, I am just trying to create a Android packaging tool for AGS and am trying to see if I could compile the latest version. The error seems similar though:

jni/../jni/../../../Engine/platform/android/acpland.cpp: In function 'void selec
tLatestSavegame()':
jni/../jni/../../../Engine/platform/android/acpland.cpp:542:37: warning: compari
son between signed and unsigned integer expressions [-Wsign-compare]
jni/../jni/../../../Engine/platform/android/acpland.cpp: In function '_jstring*
Java_com_bigbluecup_android_PreferencesActivity_readStringConfigValue(JNIEnv*, j
object, jint, jstring)':
jni/../jni/../../../Engine/platform/android/acpland.cpp:302:1: warning: control
reaches end of non-void function [-Wreturn-type]
"Compile arm  : agsengine <= libc.c
jni/../jni/../../../Engine/platform/util/libc.c:22:8: error: conflicting types f
or 'malloc_usable_size'
compilation terminated due to -Wfatal-errors.
make: *** [obj/local/armeabi/objs/agsengine/platform/util/libc.o] Error 1

What version of the NDK are you using? The declaration of malloc_usable_size changed with NDK r9 and I adapted the source here accordingly: https://github.com/adventuregamestudio/ags/commit/fe3d87c86b830c5e6d7b7e990de5ec5ef68483e7#diff-84cd350904395d756aa62ee013e17353L22

We have an issue with voice clips being cut off. The vast majority of voice clips work fine, but we've encountered an issue where a voice clip is cut off near the end; the text and portrait animation is played properly for the correct amount of time, but the voice halts and cuts off the last few words. Here's a sample project that shows it; run the executable and press space bar to have EGO say something.

http://crystalshard.net/test/Voicetest.zip

I pushed a fix: https://github.com/adventuregamestudio/ags/commit/7fda041a715ff2792de7cc6d789929b4682716d7

I could also reproduce this error with 3.21 and that particular ogg file. Interestingly the engine seems to fail at keeping the buffer filled. Eventually playback ends here: https://github.com/adventuregamestudio/ags/blob/master/Engine/libsrc/alogg/alogg.c#L1022

For me the speech never ends though. Audio cuts out and the speech line stays on screen forever.

Edit: Hmm, removing this "buffer underrun" branch causes the full clip to play. May be a bug in the underrun detection.

The Draconian edition has this "pixel-perfect" scaling, it boils down to the changes in the d3d driver: https://github.com/AlanDrake/Adventure-Game-Studio/commit/5e0ae8b6c88663d5f2b4f2732dedac47cd8caa0c#diff-5d98445a638921481f8b22c5397f77b0R1392

[Makefile-defs.linux]

You can change the linker flags to produce verbose output. This will show what files the linker checks before the linking fails. In the file Makefile-defs.linux change the line:

Code: makefile

override LDFLAGS  += -Wl,--as-needed $(addprefix -L,$(LIBDIR))

to

Code: makefile

override LDFLAGS  += -Wl,-verbose,--as-needed $(addprefix -L,$(LIBDIR))

This will give a lot of additional output, for my system the relevant output lines are:

Code: makefile

attempt to open /usr/lib/x86_64-linux-gnu/libstdc++.so failed
attempt to open /usr/lib/x86_64-linux-gnu/libstdc++.a failed
attempt to open /usr/lib/gcc/x86_64-linux-gnu/4.7/libstdc++.so succeeded
-lstdc++ (/usr/lib/gcc/x86_64-linux-gnu/4.7/libstdc++.so)

In the options menu, did you try with different settings for "Drop frames if necessary"? If this is not checked, every frame will be displayed regardless of speed. If checked, frames will be skipped according to the video speed if the processor is too slow to display them all. I had added that option because slow devices might skip all frames of a video because they lag behind from the beginning.

There is also no hardware acceleration for the video decoding, so a large video will not play smoothly.

I will push these changes to the repository, so they will be in the normal builds. I guess mouse and joystick support could be always on.

From the looks of it, the OUYA controller does not feature a start or a select button. Strange design decision.

Thanks for testing the build! Sounds very promising. Implementing button combinations (Alt-X, etc.) is a bit difficult, so it won't happen too soon I think.

The build also has some sort of gamepad support in that two buttons are mapped to left click and right click, MENU is mapped to the "select" button and ESC to the "start" button. Wiggling the analog stick around should also move the mouse cursor, but that doesn't work right yet.

It side loads just fine on OUYA. The only thing I wish I could do (and honestly, this would be good on ALL Android version), is use a real mouse. I have a USB Keyboard and Mouse hooked up to my OUYA, and it works great in DosBox and ScummVM. In this AGS port however, it moves to the point I click and then I have to click again if I want to single click (double Click to single click... kinda). having a way to enable real mouse tracking would be great, and would also work with OTG cables on phones.

Thanks,

Can you try this build with experimental mouse/keyboard support? I can only test it in a virtual machine, so I cannot tell how it behaves on a real device. Moving and left/right-click should work as well as most keyboard keys (including ESC and F1-F12). Is the AGS mouse cursor for you offset from the Android cursor? http://jjs.at/temp/AGS_mouse.apk

Actually I had some heavy performance problems (a maximum of 13 FPS) when testing a 1024x768 game (32 bit).

Have you tried using one of the hardware renderers from the preferences menu? It should boost the performance like using the D3D driver on Windows does. It is a bit incomplete though because there is no character tinting yet.

I know it's been mentioned/requested before, but MAN would I love to see an OUYA port.

It should be possible to sideload AGS onto the OUYA. There would be the problem of missing controller support though. I don't have an OUYA (no intention on buying one either) and the SDK probably still doesn't work right on the emulator (I will test if that holds still true). If someone else wants to implement OUYA support be my guest.

I guess that is my fault. For the PSP I decreased the default sprite cache size from 20 MB to 5 MB. This could cause a reloading of sprites for high-res games. https://github.com/adventuregamestudio/ags/blob/master/Common/ac/spritecache.cpp#L52

It looks like you can download VS 2008 Express with SP1 too. The english Microsoft site doesn't mention it anymore but the links to the ISOs still seem to work: http://download.microsoft.com/download/e/8/e/e8eeb394-7f42-4963-a2d8-29559b738298/VS2008ExpressWithSP1ENUX1504728.iso You can eventually get there when googling for visual studio 2008 express with sp1.

E: The log looks alright except for the linker problem.

Are you sure that you cannot install the service pack in the express edition? I see no indication for that.

The jailbreak requirement is only for the version in this thread as it is the only way to install non-store apps. Other changes for submitting to the store would be mostly adding meta files like all the required images for the different devices. You can see these changes in JanetG's repository linked above. It also has modified controls that are more suited for touch control.

This version seems to crash when using the Lua plugin and the project has a GUI.

The issue is probably fixed with this commit.

I don't know anything about the ARCH package system, so the makefile is probably not correct when setting include paths. Dependencies should be libz and liblua-5.1, but I only tried compiling on Debian-style systems.

The scripts could probably be compiled at runtime, but from what I've read the library for serialization of the data (for savegames) just dumps out the raw state of the lua interpreter. So I suppose that savegames will not be cross-platform compatible. But I don't really have a big insight into it.

Well, there are several problems.

1. The error messages indicate that the engine cannot find the plugin at all. The libagslua.so library must be either in the games directory or in the same directory as the ags binary.

2. There was a problem with the library exports not being externally visible, it should be fixed with the latest commit that I just pushed.

3. Now the big one: The plugin might work on 64 bit systems in theory. But because the scripts are compiled to lua bytecode, which is platform dependant, the game will not load. You will get some error message from the plugin that indicates a corrupted script file. The byte code is compatible across all 32 bit little endian systems though (e.g. it works on and ARM Android device).

As for the build failing on the lua plugin: You have to initialize and update the lua submodule. See here: http://www.adventuregamestudio.co.uk/forums/index.php?topic=44768.msg636451342#msg636451342

I switched the SDK target to 10 now, so the backward compatibility stuff should be in full effect now. The new daily build has that change.

I tried Space Quest II on my phone and on the emulator and I could enter characters via the onscreen keyboard, so I cannot reproduce the problem.

Antialiased sprite drawing is also available for the software renderer. I made a quick test and it works for me on Linux (tried with Pleurhburg and Ben Jordan 8) when I force the option via the acsetup.cfg file.

Uru?