The SDL forums have moved to discourse.libsdl.org.
This is just a read-only archive of the previous forums, to keep old links working.

SDL 2.0 is almost ready for release!

It is currently being used by a number of commercial games on all supported platforms, and is being seen by millions of people in Valve's recent game updates.


Please bang on the current snapshot and see if there are any issues for you:

http://www.libsdl.org/tmp/SDL-2.0.zip


Please add fixes for any issues you have to Bugzilla:

http://bugzilla.libsdl.org/


Here are the big pieces still pending that I am aware of:

* Final decision on sensor API

* Adding synchronous callback support for iOS application state changes
* ???


Cheers!

--Sam

Just went try it... mouse wheel stopped working for me o_o Ubuntu
12.04 64-bit. I'll check my code to make sure I didn't do something
stupid but that seems like a big problem to me.

I'll go try testing fullscreen later because last time I tried it
still broke up miserably.

2013/4/2, Sam Lantinga:
_______________________________________________
SDL mailing list

http://lists.libsdl.org/listinfo.cgi/sdl-libsdl.org

Looked up, this is the code (everything else is pretty much not
specific to SDL and would have already broken up otherwise):

// Mouse wheel?
case SDL_MOUSEWHEEL:
if (!event.wheel.windowID)
break;
if (event.wheel.y > 0)
input.cursor.wheel_up = 1;
if (event.wheel.y < 0)
input.cursor.wheel_down = 1;
break;

The only thing that seems like it could be suspicious is the window
check (which I use to make sure the cursor is within the window). Did
anything change that I'm not aware of?

2013/4/2, Sik the hedgehog:
_______________________________________________
SDL mailing list

http://lists.libsdl.org/listinfo.cgi/sdl-libsdl.org

On 2013-04-02 22:12, Sam Lantinga wrote:

In SDL_string.c, lines 34 and 35, the definitions for SDL_atoi and
SDL_atof are obviously incorrect due to infinite recursion. I suspect
that SDL_atoi should call SDL_atoi_inline instead of itself, and the
same for SDL_atof.


--
Rainer Deyke
_______________________________________________
SDL mailing list

http://lists.libsdl.org/listinfo.cgi/sdl-libsdl.org

Suddenly the mouse wheel issue fixed itself after a reboot... Wait, what? o_O

2013/4/3, Rainer Deyke:
_______________________________________________
SDL mailing list

http://lists.libsdl.org/listinfo.cgi/sdl-libsdl.org

Good catch! Now fixed in hg changeset d4b70d310ade.

Most apps probably used the inline version in the headers, and probably
wouldn't have hit this bug; the ones in SDL_string.c are only there so
you can access the symbol if you want to link against it directly or
dlsym() it.

--ryan.


_______________________________________________
SDL mailing list

http://lists.libsdl.org/listinfo.cgi/sdl-libsdl.org

Aside from the 2 you mentioned, there are two things:

- Joystick Support for Android, I see that the patch was applied adn then removed because someone reported that the minimal android version was bumped up.
But that is not exactly true, you really need to build against a newer SDK version, but you can run the app on older versions... I suggest someone to check that (I would do that if I know how )
This one is related to the Sensor API because of the old Accelerometer inteface.
- The Sensor API would be used for iOS too, right?

from the Roadmap:
- "Create desktop video mode change notification"
This one should be reatively easy (at last on OSX is)

Also, is the Forcefeedback API well tested on OS X?
I am not sure if this is a problem with SDL or with the xbox 360 controller driver... (if this is the case, ignore this question and please recommend me something with forcefeedmack which works on OSX, thanks)

Cheers!
Rodrigo Rocha

Sam wrote:


This again I'm ready to test across OS X, windows, and iOS, except,
of course, for this missing bit. I know we went back and forth a lot
with this.

Has anything been decided internally? I'm pretty much willing to deal
with anything that fits within the needs of the APIs. I'd be more than
happy to pilot test stuff, too.

[>] Brian
_______________________________________________
SDL mailing list

http://lists.libsdl.org/listinfo.cgi/sdl-libsdl.org

While we're at it, can we pretty please get the CMake script fixed for pthreads? http://bugzilla.libsdl.org/show_bug.cgi?id=1709

I understand that it masks a bug in the pthread emulation code which needs further investigation, but as things stand now every default-setting Linux build of SDL2 I've tried (x86, x86_64 and armhf) deadlocks on init.



On 4 April 2013 02:07, Brian Barnes wrote:

Something that would be great before the release date are project templates for currently supported projects (visual studio, xcode, etc)
I am almost sure someone posted here xcode templates a few weeks ago.

If CMake has been chosen as build system replacement, would it be possible to integrate it in time for SDL2?

Vittorio



On Thu, Apr 4, 2013 at 5:57 AM, RodrigoCard wrote:

We have a GSOC project idea around a cross-platform build system (cmake or whatever) but won't know if SDL got accepted as mentoring organization until April 8th. If accepted, one would also need to find a student interested to work on this. The timeline for such a project would not be May ... but rather Sept.

On 4/4/2013 2:43 AM, Vittorio Giovara wrote:

This: http://bugzilla.libsdl.org/show_bug.cgi?id=1431

I think it must be easy to solve but I suck at autotools (yes, I'm for cmake).



2013/4/4 Andreas Schiffler

Shame on me, I left a comment in that ticket saying that I would fix it soon...almost a year ago :)

Anyway, better late than never! Let me know if the patch I attached there fixes the problem for you (I tested in my system and running with --disable-static or --disable-shared or none at all seem to work fine now).


2013/4/4 Carles Pagès




--
Gabriel.

The patch spacing was broken and didn't apply to the unpacked zip (perhaps you used the repo and it has changed?) Anyway, after some minor modifications it applied and it fixes the issue, so thanks!

I attach the reworked patch in case it is of use.



2013/4/5 Gabriel Jacobo

Hi,

quite a while ago (before I also subscribed to the mail list), I
submitted a small patch on the forum which allows one to disconnect the
event loop from the video loop, in case one does not need the video
capabilities of SDL. Since SDL is one of the few crossplatform
libraries for HID access, it would be great to have this option. More
arguments for it are in the forum post.

http://forums.libsdl.org/viewtopic.php?t=8855

Unfortunately the patch does not fix it for mouse and keyboard access
yet...
For fixing that, I can highly recommend looking at Chuck's
implementation of HID, which was based on SDL (some years back), but
then adapted for additional features - and disconnecting from the
videoloop.

In other words... it would be great if the input device access could be
decoupled in such a way that it can be used as a separate module.

sincerely,
Marije

On Tue, 2 Apr 2013 13:12:55 -0700
Sam Lantinga wrote:


_______________________________________________
SDL mailing list

http://lists.libsdl.org/listinfo.cgi/sdl-libsdl.org

2013/4/3 Scott Percival




Turns out there was no bug being masked, the semaphore emulation uses pthread mutexes, and those need the -pthread flag.

Fixed in http://hg.libsdl.org/SDL/rev/a2bddc1fb02f 




--
Gabriel.

Fantastic work. Just tested builds with and without semaphore emulation, that magic juju -pthreads flag seems to have done the trick. Thanks!




On 20 April 2013 00:32, Gabriel Jacobo wrote:

Umm, it's May now...
Where is SDL 2.0 ?

They never said *which* day of May :P

Somebody should set up a countdown ala Exodus that ends at the last
day of the month XD

2013/5/1, JeZ-l-Lee:
_______________________________________________
SDL mailing list

http://lists.libsdl.org/listinfo.cgi/sdl-libsdl.org

Please be patient, a delayed software is better than a buggy one.
If you are in a hurry, feel free to submit patches and running tests, it'll definitely speed up development.
Cheers,
Vittorio



On Thu, May 2, 2013 at 12:44 AM, JeZ-l-Lee wrote:

Where can I find precompiled shared libs to test? If possible, for Windows and Linux.

Hi,

I agree!!!

Staff should release a compiled BETA for Windows and Linux for testing purposes!
So we can use it now and test it before official release...

There aren't precompiled libraries as far as I know.

On Linux it's easy if you just take the Mercurial snapshot (although
it's a bit out of date) because you only need to do this (CMake
support is still a bit rough but this should work still):

./configure
make
sudo make install
sudo ldconfig

You can replace that last one with a reboot :v Also I'd suggest
getting some development libraries (like OpenGL) if you don't want to
be stuck with horrible performance.

Windows is more tricky, can't say anything about it for the moment.

2013/5/2, REVERSE:
_______________________________________________
SDL mailing list

http://lists.libsdl.org/listinfo.cgi/sdl-libsdl.org

Oh, you guys thought May 2013? No no, there must have been a confusion, it's May *2014*









¿Kidding? aside, Sam told me he wanted to get a release candidate in the first (firsts?) week/s of the month with a final release after that, though as they say "the best laid plans of mice and men often go astray".


I for one have a few pending bug fixes in the tracker that I would like to get reviewed and incorporated before that happens.



2013/5/2 Vittorio Giovara




--
Gabriel.

Windows is not tricky, just download the source and build from the visual project provided.
I could do that and post the libs and dlls somewhere if it's helpful.

1) I'm not sure if any extra libraries need to be downloaded or if SDL
can work as-is (albeit not fully optimal).

2) That only works for Visual Studio. If you're using MinGW it's a
whole different story.

2013/5/2, li:
_______________________________________________
SDL mailing list

http://lists.libsdl.org/listinfo.cgi/sdl-libsdl.org

Sik, you're right. I'm using Code::Blocks and MingW. Is there an easy way to compile libs?

this is why we need an officially released Release Candidate...

This is why instead of complaining, you should really learn how to compile software...
Also Release Candidate is one of the latest stages of a release cycle, before there are alphas and betas and a varying number of them.


Anyway enough on this topic from me...
Vittorio

On 02/mag/2013, at 21:45, JeZ-l-Lee wrote:

With MSys or Cygwin, the first two steps are the same.
With Visual C++, there's already a Visual Studio project.

(PS: If you're familiar with Linux dev and don't have MSys or Cygwin on Windows, you've made a massive mistake).

Use Minimal System (an optional part of MingW): http://www.mingw.org/wiki/MSYS

Then literally all you do is:
cd C:/MySDLDirectory/
./configure
make

Also make sure that this directory does not have mercurial information (either delete it or copy the contents to another location and build there), or else config will fail towards the end. It's incredibly obnoxious, but it explains the reason for failure in plain English.

Quoth Nathaniel J Fries, on 2013-05-05 04:35:34 -0700:

As far as I've been able to get with this, things like DirectSound rather than
WinMM audio backend don't work with MinGW because the needed DirectX development
files aren't available in that environment. Nor can you just blindly pull the
headers in, because they seem to depend on Microsoft-compiler-specific pieces,
which might be rectifiable some other way---that's roughly where I am WRT SDL
on Windows locally, and at the limit of my Windows toolchain experience (my
next step is a hunch that I'm missing the "Windows SDK" stuff).

(I haven't gotten SDL to compile with VS either, but I think that's just some
local bogosity. I'd rather have a MinGW compile due to the VS redist stuff,
though even that is on shaky ground since MSVCRT isn't intended for non-OS use;
the C runtime library situation on Windows is just hairy everywhere, ugh.)

---> Drake Wilson
_______________________________________________
SDL mailing list

http://lists.libsdl.org/listinfo.cgi/sdl-libsdl.org

Yeah, I haven't gotten dsound enabled in my MinGW build either and the combination of sound lag and another symptom cause some awful results with multiple sounds playing.  I was trying with CMake, but it doesn't pick up the DirectX SDK properly.

I noticed that the compiler flag character is wrong for gcc ('/' vs '-') in this line:set(CMAKE_REQUIRED_FLAGS "/I\"$ENV{DXSDK_DIR}\\Include\"")


...and the environment variable has an extra '\' at the end.  I don't know if either matters.



Either way, the check_include_file lines fail and when I override them, DirectX still doesn't seem to get linked. :-/


Jonny D



On Sun, May 5, 2013 at 8:49 AM, Drake Wilson wrote:

Well, maybe it is an issue with missing headers...  It looks like Visual Studio is necessary at least for some system headers.

Jonny D



On Sun, May 5, 2013 at 9:18 AM, Jonathan Dearborn wrote:

Alright, I think I finally got it.  I disabled dinput and xaudio2 so I didn't need the VS headers.  Here's a build:http://www.dinomage.com/temp/SDL2_mingw32_5-5-13.zip


Jonny D




On Sun, May 5, 2013 at 9:29 AM, Jonathan Dearborn wrote:

2013/5/5, Nathaniel J Fries:

The problem is that if you install Code::Blocks, you get MinGW without
MSYS which completely screws up the idea of using configure and
make...

2013/5/5, Drake Wilson:

MinGW-w64 seems to come with the DirectX SDK (at least the headers and
the libraries are there, couldn't get around doing anything with it
yet). Vanilla MinGW doesn't seem to, though, but last I installed
vanilla MinGW was some time ago so maybe that changed.

Sure, one could try to use the multimedia API instead of DirectSound
(it's possible to build SDL without DirectX at all), but then you'll
be stuck with 500ms of audio latency. That's very far from acceptable.
I know because I tried that...
_______________________________________________
SDL mailing list

http://lists.libsdl.org/listinfo.cgi/sdl-libsdl.org

I just borrow headers from the Wine project when they're missing from MinGW and I need them. You need to patch in a couple macros defined by wine, but that's no biggie. I also had to patch dsound.h to include mmsystem.h to pass the dsound test in configure for some reason.

Also, why exactly does dsound support require ddraw.h and dinput.h?

Here, I stuck the required headers for dsound support in my dropbox: http://db.tt/3t5ckfRH

Can you submit it to bugzilla? We'll take a look at it for release.


http://bugzilla.libsdl.org



On Sun, Apr 7, 2013 at 2:41 AM, Marije Baalman wrote:

The best laid plans of mice...


We moved this past weekend and I won't have internet access for another week.  We' re unpacking and so on. I'm still working on the release, but this may delay things, FYI.

Cheers!



On Tue, Apr 2, 2013 at 1:12 PM, Sam Lantinga wrote:

In the meantime, please download and try out the current snapshot here:
http://www.libsdl.org/tmp/SDL-2.0.zip

... and please report issues and provide patches on bugzilla:

http://bugzilla.libsdl.org/

See ya!





On Mon, May 6, 2013 at 10:29 AM, Sam Lantinga wrote:

In the meantime, please download and try out the current snapshot here:
http://www.libsdl.org/tmp/SDL-2.0.zip

... and please report issues and provide patches on bugzilla:

http://bugzilla.libsdl.org/

See ya!





On Mon, May 6, 2013 at 10:29 AM, Sam Lantinga wrote:

In case folks didn't notice, I'm back online!

I'm busy catching up on e-mail and will be working on building packages this weekend for people to try out.

Cheers!



On Mon, May 6, 2013 at 10:29 AM, Sam Lantinga wrote:

Very nice to hear that Sam!


Before SDL is officially released, do you think you could "clean up" the source code files?

I've been using http://mercurial.selenic.com/wiki/CheckFilesExtension to remove any tabs and get rid of trailing whitespaces and I've been quite happy with it.



Vittorio



On Thu, May 16, 2013 at 10:22 AM, Sam Lantinga wrote:

+1 on the cleanup. Not all editors are created equal :P


On Sat, May 18, 2013 at 12:40 PM, Vittorio Giovara wrote:

Hi,

Will full documentation be released in the upcoming SDL2 Release Candidate?
I can't really learn SDL2, SDL2_Image, SDL2_Mixer, SDL2_TTF without proper documentation.

I know LazyFoo is working on SDL2 use examples, but not sure what the status is on that.

Thanks!

SDL is community documented.  Anyone is welcome to create an account on the wiki and contribute to the documentation, just send me your wiki username and request access!

http://wiki.libsdl.org/



On Sat, May 18, 2013 at 9:01 AM, JeZ-l-Lee wrote:

You got it!http://hg.libsdl.org/SDL/rev/75360622e65f




On Sat, May 18, 2013 at 5:40 AM, Vittorio Giovara wrote:

Great, thanks Sam!


On Sat, May 18, 2013 at 11:20 PM, Sam Lantinga wrote:

I know it's far too late on making any API changes, buT what is the case about SDL_Init(SDL_INIT_EVERYTHING) failing because of SDL_INIT_HAPTIC??

Or SDL_INIT_EVERYTHING should be not used anymore?

I know it's far too late on making any API changes, buT what is the case about SDL_Init(SDL_INIT_EVERYTHING) failing because of SDL_INIT_HAPTIC??

Or SDL_INIT_EVERYTHING should be not used anymore?

SDL_INIT_EVERYTHING should really be only used if you actually need all of the subsystems provided by SDL.


On Thu, May 23, 2013 at 1:47 AM, neoaggelos wrote:

On 05/23/2013 11:42 AM, Sam Lantinga wrote:

I would go further and say "don't ever use SDL_INIT_EVERYTHING," but
that's just me.

Do we know why SDL_INIT_HAPTIC fails in this case? Is it something we
can reasonably fix (like, not fail, but report zero haptic devices
instead, or something)?

--ryan.



_______________________________________________
SDL mailing list

http://lists.libsdl.org/listinfo.cgi/sdl-libsdl.org