gamedevlists-general — List for discussing general game development issues

Re: [GD-General] memory checking

[Timo H. <ma...@ti...>]

Hi Brett,

what about Paul Nettle's free Memory Manager:
http://www.flipcode.com/cgi-bin/msg.cgi?showThread=12September2000-PresentingAMemoryManager&forum=askmid&id=-1

it is:
- free
- simple to use (just put in the files into your project and include
  the header in all your files)

Timo

 Brett wrote:

> I recently ran into a serious memory bug I just can't track down and it seems it's time to get some sort of memory debugging tool. Ideally it would be available for Win32, PS2 and GameCube and work
> with CodeWarrior, but just Windows would be acceptable.  The only tools I can find so far are BoundsChecker and Purify, but neither looks like it will work with our CodeWarrior setup.  I also tried
> just using the crtdbg funcitons, but getting the MS libs into CodeWarrior is proving problematic.

> Does anybody know of any commercial product that might be useful? Are there any standalone memory checkers out there? Anybody using BoundsChecker that can comment on whether you must have Visual
> Studio or not?

> Thanks,
> Brett

RE: [GD-General] memory checking

[Ivan-Assen I. <as...@ha...>]

> does anyone have any suggestions for open-source applications for 
> debugging memory leaks on windows?  we can't afford the 
> license fees for 
> bounds checker etc unfortunately.

Hi,

I've seen Boehm's gargabage collector for C successfully used
for leaks hunting. (ifl://boehm+garbage+collector) 

But of course, it won't catch overruns or random memory damages.

We tried BoundsChecker once, but it was so unhappy with our
own debug memory manager that it was unusable, reporting
thousands of non-leaks as leaks etc.

Re: [GD-General] memory checking

[Mike W. <mi...@ge...>]

does anyone have any suggestions for open-source applications for=20
debugging memory leaks on windows?  we can't afford the license fees for=20
bounds checker etc unfortunately.

mike w
www.gekidodesigns.com

Brett Bibby wrote:
> Thanks guys for your notes. Since we have a memory corruption our memor=
y wrappers didn't pick up on the error. I'll look into some of the fancy =
stuff Brian mentioned. I heard good things about valgrind but alas it's o=
nly linux. Hmm, a port... maybe something to think about. I did find Heap=
Agent http://www.microquill.com/mqplatfm.htm interesting but the demo ver=
sion told me less than CodeWarrior about the problem :)=20
> Cheers,
> Brett
>=20
> ----- Original Message -----=20
> From: "Daniel Vogel" <vo...@ep...>
> To: <gam...@li...>
> Sent: Friday, March 05, 2004 12:32 PM
> Subject: RE: [GD-General] memory checking
>=20
>=20
>=20
>>Won't really help you (gotta love posts that start like this ;)) as it'=
s
>>Linux only though Valgrind is a godsend for this purpose -
>>http://valgrind.kde.org/. I wish someone would either port it to Window=
s or
>>create an equivalently useful product.=20
>>
>>I tried BoundsChecker in the past though wasn't happy with it as it req=
uires
>>instrumentation which didn't work in conjunction with our engine.
>>
>>-- Daniel, Epic Games Inc. =20
>>
>>
>>>-----Original Message-----
>>>From: gam...@li...=20
>>>[mailto:gam...@li...] On=20
>>>Behalf Of Brett Bibby
>>>Sent: Thursday, March 04, 2004 10:34 PM
>>>To: Gam...@li...
>>>Subject: [GD-General] memory checking
>>>
>>>I recently ran into a serious memory bug I just can't track=20
>>>down and it seems it's time to get some sort of memory=20
>>>debugging tool. Ideally it would be available for Win32, PS2=20
>>>and GameCube and work with CodeWarrior, but just Windows=20
>>>would be acceptable.  The only tools I can find so far are=20
>>>BoundsChecker and Purify, but neither looks like it will work=20
>>>with our CodeWarrior setup.  I also tried just using the=20
>>>crtdbg funcitons, but getting the MS libs into CodeWarrior is=20
>>>proving problematic.
>>>
>>>Does anybody know of any commercial product that might be=20
>>>useful? Are there any standalone memory checkers out there?=20
>>>Anybody using BoundsChecker that can comment on whether you=20
>>>must have Visual Studio or not?
>>>
>>>Thanks,
>>>Brett
>>>
>>>
>>>-------------------------------------------------------
>>>This SF.Net email is sponsored by: IBM Linux Tutorials
>>>Free Linux tutorial presented by Daniel Robbins, President and CEO of
>>>GenToo technologies. Learn everything from fundamentals to system
>>>administration.http://ads.osdn.com/?ad_id=1470&alloc_id638&op=CCk
>>>_______________________________________________
>>>Gamedevlists-general mailing list
>>>Gam...@li...
>>>https://lists.sourceforge.net/lists/listinfo/gamedevlists-general
>>>Archives:
>>>http://sourceforge.net/mailarchive/forum.php?forum_idU7
>>>
>>
>>
>>
>>-------------------------------------------------------
>>This SF.Net email is sponsored by: IBM Linux Tutorials
>>Free Linux tutorial presented by Daniel Robbins, President and CEO of
>>GenToo technologies. Learn everything from fundamentals to system
>>administration.http://ads.osdn.com/?ad_id=1470&alloc_id638&op=CCk
>>_______________________________________________
>>Gamedevlists-general mailing list
>>Gam...@li...
>>https://lists.sourceforge.net/lists/listinfo/gamedevlists-general
>>Archives:
>>http://sourceforge.net/mailarchive/forum.php?forum_idU7
>=20
>=20
>=20
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> administration.http://ads.osdn.com/?ad_id=1470&alloc_id638&op=CCk
> _______________________________________________
> Gamedevlists-general mailing list
> Gam...@li...
> https://lists.sourceforge.net/lists/listinfo/gamedevlists-general
> Archives:
> http://sourceforge.net/mailarchive/forum.php?forum_idU7
>=20
>=20

Re: [GD-General] memory checking

[Brett B. <res...@ga...>]

Thanks guys for your notes. Since we have a memory corruption our memory =
wrappers didn't pick up on the error. I'll look into some of the fancy =
stuff Brian mentioned. I heard good things about valgrind but alas it's =
only linux. Hmm, a port... maybe something to think about. I did find =
HeapAgent http://www.microquill.com/mqplatfm.htm interesting but the =
demo version told me less than CodeWarrior about the problem :)=20
Cheers,
Brett

----- Original Message -----=20
From: "Daniel Vogel" <vo...@ep...>
To: <gam...@li...>
Sent: Friday, March 05, 2004 12:32 PM
Subject: RE: [GD-General] memory checking


> Won't really help you (gotta love posts that start like this ;)) as =
it's
> Linux only though Valgrind is a godsend for this purpose -
> http://valgrind.kde.org/. I wish someone would either port it to =
Windows or
> create an equivalently useful product.=20
>=20
> I tried BoundsChecker in the past though wasn't happy with it as it =
requires
> instrumentation which didn't work in conjunction with our engine.
>=20
> -- Daniel, Epic Games Inc. =20
>=20
> > -----Original Message-----
> > From: gam...@li...=20
> > [mailto:gam...@li...] On=20
> > Behalf Of Brett Bibby
> > Sent: Thursday, March 04, 2004 10:34 PM
> > To: Gam...@li...
> > Subject: [GD-General] memory checking
> >=20
> > I recently ran into a serious memory bug I just can't track=20
> > down and it seems it's time to get some sort of memory=20
> > debugging tool. Ideally it would be available for Win32, PS2=20
> > and GameCube and work with CodeWarrior, but just Windows=20
> > would be acceptable.  The only tools I can find so far are=20
> > BoundsChecker and Purify, but neither looks like it will work=20
> > with our CodeWarrior setup.  I also tried just using the=20
> > crtdbg funcitons, but getting the MS libs into CodeWarrior is=20
> > proving problematic.
> >=20
> > Does anybody know of any commercial product that might be=20
> > useful? Are there any standalone memory checkers out there?=20
> > Anybody using BoundsChecker that can comment on whether you=20
> > must have Visual Studio or not?
> >=20
> > Thanks,
> > Brett
> >=20
> >=20
> > -------------------------------------------------------
> > This SF.Net email is sponsored by: IBM Linux Tutorials
> > Free Linux tutorial presented by Daniel Robbins, President and CEO =
of
> > GenToo technologies. Learn everything from fundamentals to system
> > administration.http://ads.osdn.com/?ad_id=1470&alloc_id638&op=CCk
> > _______________________________________________
> > Gamedevlists-general mailing list
> > Gam...@li...
> > https://lists.sourceforge.net/lists/listinfo/gamedevlists-general
> > Archives:
> > http://sourceforge.net/mailarchive/forum.php?forum_idU7
> >=20
>=20
>=20
>=20
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> administration.http://ads.osdn.com/?ad_id=1470&alloc_id638&op=CCk
> _______________________________________________
> Gamedevlists-general mailing list
> Gam...@li...
> https://lists.sourceforge.net/lists/listinfo/gamedevlists-general
> Archives:
> http://sourceforge.net/mailarchive/forum.php?forum_idU7

RE: [GD-General] memory checking

[Daniel V. <vo...@ep...>]

Won't really help you (gotta love posts that start like this ;)) as it's
Linux only though Valgrind is a godsend for this purpose -
http://valgrind.kde.org/. I wish someone would either port it to Windows =
or
create an equivalently useful product.=20

I tried BoundsChecker in the past though wasn't happy with it as it =
requires
instrumentation which didn't work in conjunction with our engine.

-- Daniel, Epic Games Inc. =20

> -----Original Message-----
> From: gam...@li...=20
> [mailto:gam...@li...] On=20
> Behalf Of Brett Bibby
> Sent: Thursday, March 04, 2004 10:34 PM
> To: Gam...@li...
> Subject: [GD-General] memory checking
>=20
> I recently ran into a serious memory bug I just can't track=20
> down and it seems it's time to get some sort of memory=20
> debugging tool. Ideally it would be available for Win32, PS2=20
> and GameCube and work with CodeWarrior, but just Windows=20
> would be acceptable.  The only tools I can find so far are=20
> BoundsChecker and Purify, but neither looks like it will work=20
> with our CodeWarrior setup.  I also tried just using the=20
> crtdbg funcitons, but getting the MS libs into CodeWarrior is=20
> proving problematic.
>=20
> Does anybody know of any commercial product that might be=20
> useful? Are there any standalone memory checkers out there?=20
> Anybody using BoundsChecker that can comment on whether you=20
> must have Visual Studio or not?
>=20
> Thanks,
> Brett
>=20
>=20
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> administration.http://ads.osdn.com/?ad_id=1470&alloc_id638&op=CCk
> _______________________________________________
> Gamedevlists-general mailing list
> Gam...@li...
> https://lists.sourceforge.net/lists/listinfo/gamedevlists-general
> Archives:
> http://sourceforge.net/mailarchive/forum.php?forum_idU7
>=20

Re: [GD-General] memory checking

[brian s. <pud...@po...>]

You can probably get what you need writing your own - I imagine that's 
what most cross-platform and console people end up doing.  It's what I 
did.  It's easy.

Assuming you have all the C allocation functions inside your own 
wrappers, you adjust all allocations to allocate a larger block, add 
guard bytes at the start and end of the block, and return a pointer to 
the inside of the block (you'll want to add enough bytes at the front to 
guarantee the same alignment that your platform's malloc delivers).  
You'll want to put a reference to the source file and line in that 
header block too, so you can see where the trashed pointer comes from.

Then when the pointer's freed, you just check those guard bytes and see 
if they're trashed, before calling the actual free function (with the 
pointer adjusted downward, of course). 

You can also keep a list of all the pointers you've alloc'ed, and write 
a function that walks over all of them looking for damage.  And if you 
have a list of all outstanding pointers, with file/line pairs, you can 
even start checking for leaks, if you're so inclined :)

You can do fancier stuff on Win32 and Xbox, like putting allocations in 
separate pages and using page protections to detect overruns 
immediately.  BoundsChecker probably does that (along with all sorts of 
other stuff).  But it's usually enough to be able to see which pointer 
got nuked.  If it's reproducible, you can watch a particular address in 
the debugger to see when it gets hit.

Hope this helps,

--brian

Brett Bibby wrote:

>I recently ran into a serious memory bug I just can't track down and it seems it's time to get some sort of memory debugging tool. Ideally it would be available for Win32, PS2 and GameCube and work with CodeWarrior, but just Windows would be acceptable.  The only tools I can find so far are BoundsChecker and Purify, but neither looks like it will work with our CodeWarrior setup.  I also tried just using the crtdbg funcitons, but getting the MS libs into CodeWarrior is proving problematic.
>
>Does anybody know of any commercial product that might be useful? Are there any standalone memory checkers out there? Anybody using BoundsChecker that can comment on whether you must have Visual Studio or not?
>
>Thanks,
>Brett
>
>
>-------------------------------------------------------
>This SF.Net email is sponsored by: IBM Linux Tutorials
>Free Linux tutorial presented by Daniel Robbins, President and CEO of
>GenToo technologies. Learn everything from fundamentals to system
>administration.http://ads.osdn.com/?ad_id70&alloc_id638&op=click
>_______________________________________________
>Gamedevlists-general mailing list
>Gam...@li...
>https://lists.sourceforge.net/lists/listinfo/gamedevlists-general
>Archives:
>http://sourceforge.net/mailarchive/forum.php?forum_idU7
>
>  
>

[GD-General] memory checking

[Brett B. <res...@ga...>]

I recently ran into a serious memory bug I just can't track down and it =
seems it's time to get some sort of memory debugging tool. Ideally it =
would be available for Win32, PS2 and GameCube and work with =
CodeWarrior, but just Windows would be acceptable.  The only tools I can =
find so far are BoundsChecker and Purify, but neither looks like it will =
work with our CodeWarrior setup.  I also tried just using the crtdbg =
funcitons, but getting the MS libs into CodeWarrior is proving =
problematic.

Does anybody know of any commercial product that might be useful? Are =
there any standalone memory checkers out there? Anybody using =
BoundsChecker that can comment on whether you must have Visual Studio or =
not?

Thanks,
Brett

RE: [GD-General] Re: file synchronization software

[Magnus A. <mag...@st...>]

I sent that link to the people at Subversion and they say that it isn't true
at all. Many people use large repositories with subversion.

Read this: http://www.red-bean.com/sussman/svn-anti-fud.html
It clears up a lot of claims that isn't true regarding Subversion.
 
-----Original Message-----
From: Parveen Kaler [mailto:pk...@al...] 
Sent: den 24 februari 2004 19:09
To: gam...@li...
Subject: Re: [GD-General] Re: file synchronization software

On Tue, 2004-02-24 at 01:31, Enno Rehling wrote:
> > What specifically about SVN right now do you feel makes it not mature?
> 
> It's just gone to version 1.0 - I accept there's going to be a surge of 
> people trying it out now, and a couple of them may find a few more faults.

> I'm in a position where I can wait for them.
> 

The Gentoo team recently tried to test it with their portage tree.  It
seems there are still a couple issues with big repositories.  May have
been worked out in 1.0 though.
http://article.gmane.org/gmane.linux.gentoo.devel/15374

> 
> We've got a lot of experience with CVS, and while we like what we hear
about 
> SVN, we're not under any pressure to get rid of CVS. It comes down to the 
> old "if it works, don't fix it". There's a zillion other things that need 
> fixing first.

The migration tools need some work.  The svn team is using one of the
BSD trees to test with.  (Don't remember if it was NetBSD, FreeBSD, or
OpenBSD)

The subversion team does use svn themselves.

I use it at home.  I like it a whole lot more than cvs.

-- Parveen



-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click
_______________________________________________
Gamedevlists-general mailing list
Gam...@li...
https://lists.sourceforge.net/lists/listinfo/gamedevlists-general
Archives:
http://sourceforge.net/mailarchive/forum.php?forum_id=557

Re: [GD-General] Re: file synchronization software

[Parveen K. <pk...@al...>]

On Tue, 2004-02-24 at 01:31, Enno Rehling wrote:
> > What specifically about SVN right now do you feel makes it not mature?
> 
> It's just gone to version 1.0 - I accept there's going to be a surge of 
> people trying it out now, and a couple of them may find a few more faults. 
> I'm in a position where I can wait for them.
> 

The Gentoo team recently tried to test it with their portage tree.  It
seems there are still a couple issues with big repositories.  May have
been worked out in 1.0 though.
http://article.gmane.org/gmane.linux.gentoo.devel/15374

> 
> We've got a lot of experience with CVS, and while we like what we hear about 
> SVN, we're not under any pressure to get rid of CVS. It comes down to the 
> old "if it works, don't fix it". There's a zillion other things that need 
> fixing first.

The migration tools need some work.  The svn team is using one of the
BSD trees to test with.  (Don't remember if it was NetBSD, FreeBSD, or
OpenBSD)

The subversion team does use svn themselves.

I use it at home.  I like it a whole lot more than cvs.

-- Parveen

Re: [GD-General] Re: file synchronization software

[Brian H. <ho...@py...>]

> comes down to the old "if it works, don't fix it". 

CVS works? =3D)

Brian

[GD-General] Re: file synchronization software

[Enno R. <en...@de...>]

Brian Hook wrote:

>>We are using a source management system (CVS, because we want to
>>give SVN some time to mature), but our artists don't work this way.
> 
> 
> What specifically about SVN right now do you feel makes it not mature?

It's just gone to version 1.0 - I accept there's going to be a surge of 
people trying it out now, and a couple of them may find a few more faults. 
I'm in a position where I can wait for them.

We've got a lot of experience with CVS, and while we like what we hear about 
SVN, we're not under any pressure to get rid of CVS. It comes down to the 
old "if it works, don't fix it". There's a zillion other things that need 
fixing first.

Enno.
-- 
"As a PC gamer, you need to respect yourself. Part of that respect is making 
sure that your clock speed isn't /one third/ of the Goddamn processors out 
now."  -- Tycho

RE: [GD-General] Re: file synchronization software

[Magnus A. <mag...@st...>]

Just to put in a good word for subversion. Here is a bit of their mentality.

[snapped from http://www.red-bean.com/sussman/svn-anti-fud.html]
When Subversion hit "alpha" it was already being used by dozens of private
developers and shops for real work. Any other project probably would have
called the product "1.0" at that point, but we deliberately decided to delay
that label as long as possible. Because we're talking managing people's
irreplaceable data, the project was extremely conservative about labeling
something 1.0. We were aware that many people were waiting for that label
before using Subversion, and had very specific expectations about the
meaning of that label. So we stuck to that standard. All it takes is one
high-profile case of data loss to destroy an SCM's reputation.
[/snap]

1.0 was released today (Monday).

-----Original Message-----
From: Brian Hook [mailto:ho...@py...] 
Sent: den 24 februari 2004 01:22
To: gam...@li...
Subject: Re: [GD-General] Re: file synchronization software

> We are using a source management system (CVS, because we want to
> give SVN some time to mature), but our artists don't work this way.

What specifically about SVN right now do you feel makes it not mature?

I've switched over to it, and it absolutely cleans the floor with CVS.

Brian





-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id56&alloc_id438&op=click
_______________________________________________
Gamedevlists-general mailing list
Gam...@li...
https://lists.sourceforge.net/lists/listinfo/gamedevlists-general
Archives:
http://sourceforge.net/mailarchive/forum.php?forum_idU7

Re: [GD-General] Re: file synchronization software

[Brian H. <ho...@py...>]

> We are using a source management system (CVS, because we want to
> give SVN some time to mature), but our artists don't work this way.

What specifically about SVN right now do you feel makes it not mature?

I've switched over to it, and it absolutely cleans the floor with CVS.

Brian

[GD-General] Re: file synchronization software

[Enno R. <en...@de...>]

Magnus Auvinen wrote:
> Do you mean a complete SCM (perforce, source safe, cvs, subversion, arch,
> etc) system or a something more simple?
> 
> I would recommend Subversion. Quite easy to setup and free for any use. It
> went 1.0 today as well. Look at http://subversion.tigris.org/ if it's
> something for you. You could install a hook that updates remote copies
> whenever a change is made.

I was looking for unison. thanks to everyone who replied.

We are using a source management system (CVS, because we want to give 
SVN some time to mature), but our artists don't work this way. CVS is 
really only the gateway between them and the rest of the production 
team, meaning when they've got something they want to pass off to the 
rest of us, they check it in.

Most of the time, they work on stuff more losely, and the large 
filesizes and frequent update cycles would blow up CVS (and a lot of 
other systems) real quick. They also don't want it - what they want is 
to be able to have another guy review some of their stuff, or share a 
texture quick and painless, so stuff floats around a lot. And that's 
where I want to get unison into play, if it does what I expect it to do.

Enno.
-- 
Almost nothing in Perl serves a single purpose.
       (Larry Wall in <199...@wa...>)

[GD-General] Re: file synchronization software

[Enno R. <en...@de...>]

Enno Rehling wrote:

> Some time last year, someone on this list suggested a piece of software. 
> It was used to keep assets in sync over several machines, like rsync but 
> with a few extra features (like keeping up to n backup copies).

As usual, I spent 30 minutes of searching through the list archives before I 
posted, then 5 minutes later, the name pops into my head: unison.

Enno.
-- 
"Christianity is the most perverted system that ever shone on man"
-- Thomas Jefferson

RE: [GD-General] file synchronization software

[Magnus A. <mag...@st...>]

Do you mean a complete SCM (perforce, source safe, cvs, subversion, arch,
etc) system or a something more simple?

I would recommend Subversion. Quite easy to setup and free for any use. It
went 1.0 today as well. Look at http://subversion.tigris.org/ if it's
something for you. You could install a hook that updates remote copies
whenever a change is made.

-----Original Message-----
From: Enno Rehling [mailto:en...@de...] 
Sent: den 23 februari 2004 16:22
To: gam...@li...
Subject: [GD-General] file synchronization software

Some time last year, someone on this list suggested a piece of software. It 
was used to keep assets in sync over several machines, like rsync but with a

few extra features (like keeping up to n backup copies).

I cannot remember the name. It was a command line tool, it was free, and 
there was a linux client and server that I tried out back then.

Can anyone help me remember?

Enno.
-- 
"Christianity is the most perverted system that ever shone on man"
-- Thomas Jefferson



-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click
_______________________________________________
Gamedevlists-general mailing list
Gam...@li...
https://lists.sourceforge.net/lists/listinfo/gamedevlists-general
Archives:
http://sourceforge.net/mailarchive/forum.php?forum_id=557

Re: [GD-General] file synchronization software

[Brian H. <ho...@py...>]

> I cannot remember the name. It was a command line tool, it was
> free, and there was a linux client and server that I tried out back
> then.

Unison?

http://www.cis.upenn.edu/~bcpierce/unison/

Re: [GD-General] file synchronization software

[brian s. <pud...@po...>]

I think it may have been Unison: http://www.cis.upenn.edu/~bcpierce/unison/

--brian

Enno Rehling wrote:

> Some time last year, someone on this list suggested a piece of 
> software. It was used to keep assets in sync over several machines, 
> like rsync but with a few extra features (like keeping up to n backup 
> copies).
>
> I cannot remember the name. It was a command line tool, it was free, 
> and there was a linux client and server that I tried out back then.
>
> Can anyone help me remember?
>
> Enno.

[GD-General] file synchronization software

[Enno R. <en...@de...>]

Some time last year, someone on this list suggested a piece of software. It 
was used to keep assets in sync over several machines, like rsync but with a 
few extra features (like keeping up to n backup copies).

I cannot remember the name. It was a command line tool, it was free, and 
there was a linux client and server that I tried out back then.

Can anyone help me remember?

Enno.
-- 
"Christianity is the most perverted system that ever shone on man"
-- Thomas Jefferson

Re: [GD-General] time

[Brett B. <res...@ga...>]

Forgot to mention another doc. This one at =
http://www.codeproject.com/datetime/dstbugs.asp is a really nice =
introduciton to higher level timekeeping and it helped me wade through =
some of the issues, especially wrt integrating global timestamped stuff.
Brett

----- Original Message -----=20
From: "Brett Bibby" <res...@ga...>
To: <gam...@li...>
Sent: Saturday, February 21, 2004 12:06 PM
Subject: Re: [GD-General] time


> That's a great document and covers pretty much everything we discussed =
internally, nice reference! The doc's discussion of different =
resolutions combined with independent running (some time keeps running =
even when paused) is something I hadn't thought of and I will look into =
that. Meanwhile, anybody doing an online game care to comment how you =
handle time? As a player of MMORPG I always am wondering how they =
timestamp and integrate time into their engines, and if feasible I would =
like to make sure we can support arbitrary lengths of time too.
> Brett
>=20
> ----- Original Message -----=20
> From: "Tom Spilman" <to...@pi...>
> To: <gam...@li...>
> Sent: Saturday, February 21, 2004 11:14 AM
> Subject: RE: [GD-General] time
>=20
>=20
> >=20
> >   http://www.dallasigda.org/notes.htm
> >=20
> >   You'll see a document called "Time in games" which are notes on a
> > presentation done by Squirrel Eiserloh.  It covers a lot about time =
in
> > Win32 and the different methods he's used to tackle it.  People may =
find
> > it helpful.
> >=20
> >   Tom
> >=20
> >=20
> > > -----Original Message-----
> > > From: gam...@li...=20
> > > [mailto:gam...@li...] On=20
> > > Behalf Of Brett Bibby
> > > Sent: Friday, February 20, 2004 8:59 PM
> > > To: Gam...@li...
> > > Subject: [GD-General] time
> > >=20
> > >=20
> > > Howdy,
> > > I have been wrestling with time in my engine for some time=20
> > > now (no pun intended). I wondered how other people are=20
> > > handling time? Our game engine is specifically designed for=20
> > > PS2/Gcn, but it runs under Windows too. Specifically...
> > >=20
> > > What are the different components that people are using? We=20
> > > currently have a single 32-bit time variable initialized to=20
> > > zero at startup and that's it. Each frame we either count=20
> > > v-blanks, use system ticks/timers or OS facilities custom for=20
> > > each platform to increment the global time value, usually=20
> > > stored in microseconds and rounded up to millisecond=20
> > > resolution for use in the engine. Works well and I have done=20
> > > this for every game engine I have developed so far.
> > >=20
> > > Feels like a change may be warranted though as online gaming=20
> > > becomes more popular, since if the game was to run for many=20
> > > days it would overflow and timestamping may be problematic.=20
> > > So now I'm thinking of adding a time struct (different from=20
> > > ANSI C though) whereby we store the year, month, day, hour=20
> > > and minute, and second and also a conversion struct that says=20
> > > how many mins->hour->day->month->year so we can convert=20
> > > "real" time into game time. I can pack this into 64-bits, but=20
> > > then again a simple 64-bit timer using the old way to simply=20
> > > count ticks would also eliminate the overflow problem anyway=20
> > > and so I'm questioning the right thing to do. Also, I hoping=20
> > > to be able to use absolute time for certain things where we=20
> > > used to use delta times. What are people using for timestamps=20
> > > for online games? Does each game session reset the server=20
> > > session time to prevent eventual overflow? Or is a more=20
> > > robust timing mechanism used? Are  people using microseconds?=20
> > > milliseconds?
> > >=20
> > > Thanks,
> > > Brett
> > >=20
> > >=20
> > > -------------------------------------------------------
> > > SF.Net is sponsored by: Speed Start Your Linux Apps Now.
> > > Build and deploy apps & Web services for Linux with
> > > a free DVD software kit from IBM. Click Now!=20
> > > http://ads.osdn.com/?ad_id=1356&alloc_id438> &op=3Dick
> > >=20
> > > _______________________________________________
> > >=20
> > > Gamedevlists-general mailing list=20
> > > Gam...@li...
> > > https://lists.sourceforge.net/lists/listinfo/gamedevlists-general
> > > Archives: http://sourceforge.net/mailarchive/forum.php?forum_idU7
> > >=20
> > >=20
> >=20
> >=20
> >=20
> > -------------------------------------------------------
> > SF.Net is sponsored by: Speed Start Your Linux Apps Now.
> > Build and deploy apps & Web services for Linux with
> > a free DVD software kit from IBM. Click Now!
> > http://ads.osdn.com/?ad_id=1356&alloc_id438&op=CCk
> > _______________________________________________
> > Gamedevlists-general mailing list
> > Gam...@li...
> > https://lists.sourceforge.net/lists/listinfo/gamedevlists-general
> > Archives:
> > http://sourceforge.net/mailarchive/forum.php?forum_idU7
>=20
>=20
> -------------------------------------------------------
> SF.Net is sponsored by: Speed Start Your Linux Apps Now.
> Build and deploy apps & Web services for Linux with
> a free DVD software kit from IBM. Click Now!
> http://ads.osdn.com/?ad_id=1356&alloc_id438&op=CCk
> _______________________________________________
> Gamedevlists-general mailing list
> Gam...@li...
> https://lists.sourceforge.net/lists/listinfo/gamedevlists-general
> Archives:
> http://sourceforge.net/mailarchive/forum.php?forum_idU7

Re: [GD-General] time

[Brett B. <res...@ga...>]

That's a great document and covers pretty much everything we discussed =
internally, nice reference! The doc's discussion of different =
resolutions combined with independent running (some time keeps running =
even when paused) is something I hadn't thought of and I will look into =
that. Meanwhile, anybody doing an online game care to comment how you =
handle time? As a player of MMORPG I always am wondering how they =
timestamp and integrate time into their engines, and if feasible I would =
like to make sure we can support arbitrary lengths of time too.
Brett

----- Original Message -----=20
From: "Tom Spilman" <to...@pi...>
To: <gam...@li...>
Sent: Saturday, February 21, 2004 11:14 AM
Subject: RE: [GD-General] time


>=20
>   http://www.dallasigda.org/notes.htm
>=20
>   You'll see a document called "Time in games" which are notes on a
> presentation done by Squirrel Eiserloh.  It covers a lot about time in
> Win32 and the different methods he's used to tackle it.  People may =
find
> it helpful.
>=20
>   Tom
>=20
>=20
> > -----Original Message-----
> > From: gam...@li...=20
> > [mailto:gam...@li...] On=20
> > Behalf Of Brett Bibby
> > Sent: Friday, February 20, 2004 8:59 PM
> > To: Gam...@li...
> > Subject: [GD-General] time
> >=20
> >=20
> > Howdy,
> > I have been wrestling with time in my engine for some time=20
> > now (no pun intended). I wondered how other people are=20
> > handling time? Our game engine is specifically designed for=20
> > PS2/Gcn, but it runs under Windows too. Specifically...
> >=20
> > What are the different components that people are using? We=20
> > currently have a single 32-bit time variable initialized to=20
> > zero at startup and that's it. Each frame we either count=20
> > v-blanks, use system ticks/timers or OS facilities custom for=20
> > each platform to increment the global time value, usually=20
> > stored in microseconds and rounded up to millisecond=20
> > resolution for use in the engine. Works well and I have done=20
> > this for every game engine I have developed so far.
> >=20
> > Feels like a change may be warranted though as online gaming=20
> > becomes more popular, since if the game was to run for many=20
> > days it would overflow and timestamping may be problematic.=20
> > So now I'm thinking of adding a time struct (different from=20
> > ANSI C though) whereby we store the year, month, day, hour=20
> > and minute, and second and also a conversion struct that says=20
> > how many mins->hour->day->month->year so we can convert=20
> > "real" time into game time. I can pack this into 64-bits, but=20
> > then again a simple 64-bit timer using the old way to simply=20
> > count ticks would also eliminate the overflow problem anyway=20
> > and so I'm questioning the right thing to do. Also, I hoping=20
> > to be able to use absolute time for certain things where we=20
> > used to use delta times. What are people using for timestamps=20
> > for online games? Does each game session reset the server=20
> > session time to prevent eventual overflow? Or is a more=20
> > robust timing mechanism used? Are  people using microseconds?=20
> > milliseconds?
> >=20
> > Thanks,
> > Brett
> >=20
> >=20
> > -------------------------------------------------------
> > SF.Net is sponsored by: Speed Start Your Linux Apps Now.
> > Build and deploy apps & Web services for Linux with
> > a free DVD software kit from IBM. Click Now!=20
> > http://ads.osdn.com/?ad_id=1356&alloc_id438> &op=3Dick
> >=20
> > _______________________________________________
> >=20
> > Gamedevlists-general mailing list=20
> > Gam...@li...
> > https://lists.sourceforge.net/lists/listinfo/gamedevlists-general
> > Archives: http://sourceforge.net/mailarchive/forum.php?forum_idU7
> >=20
> >=20
>=20
>=20
>=20
> -------------------------------------------------------
> SF.Net is sponsored by: Speed Start Your Linux Apps Now.
> Build and deploy apps & Web services for Linux with
> a free DVD software kit from IBM. Click Now!
> http://ads.osdn.com/?ad_id=1356&alloc_id438&op=CCk
> _______________________________________________
> Gamedevlists-general mailing list
> Gam...@li...
> https://lists.sourceforge.net/lists/listinfo/gamedevlists-general
> Archives:
> http://sourceforge.net/mailarchive/forum.php?forum_idU7

RE: [GD-General] time

[Tom S. <to...@pi...>]

  http://www.dallasigda.org/notes.htm

  You'll see a document called "Time in games" which are notes on a
presentation done by Squirrel Eiserloh.  It covers a lot about time in
Win32 and the different methods he's used to tackle it.  People may find
it helpful.

  Tom


> -----Original Message-----
> From: gam...@li...=20
> [mailto:gam...@li...] On=20
> Behalf Of Brett Bibby
> Sent: Friday, February 20, 2004 8:59 PM
> To: Gam...@li...
> Subject: [GD-General] time
>=20
>=20
> Howdy,
> I have been wrestling with time in my engine for some time=20
> now (no pun intended). I wondered how other people are=20
> handling time? Our game engine is specifically designed for=20
> PS2/Gcn, but it runs under Windows too. Specifically...
>=20
> What are the different components that people are using? We=20
> currently have a single 32-bit time variable initialized to=20
> zero at startup and that's it. Each frame we either count=20
> v-blanks, use system ticks/timers or OS facilities custom for=20
> each platform to increment the global time value, usually=20
> stored in microseconds and rounded up to millisecond=20
> resolution for use in the engine. Works well and I have done=20
> this for every game engine I have developed so far.
>=20
> Feels like a change may be warranted though as online gaming=20
> becomes more popular, since if the game was to run for many=20
> days it would overflow and timestamping may be problematic.=20
> So now I'm thinking of adding a time struct (different from=20
> ANSI C though) whereby we store the year, month, day, hour=20
> and minute, and second and also a conversion struct that says=20
> how many mins->hour->day->month->year so we can convert=20
> "real" time into game time. I can pack this into 64-bits, but=20
> then again a simple 64-bit timer using the old way to simply=20
> count ticks would also eliminate the overflow problem anyway=20
> and so I'm questioning the right thing to do. Also, I hoping=20
> to be able to use absolute time for certain things where we=20
> used to use delta times. What are people using for timestamps=20
> for online games? Does each game session reset the server=20
> session time to prevent eventual overflow? Or is a more=20
> robust timing mechanism used? Are  people using microseconds?=20
> milliseconds?
>=20
> Thanks,
> Brett
>=20
>=20
> -------------------------------------------------------
> SF.Net is sponsored by: Speed Start Your Linux Apps Now.
> Build and deploy apps & Web services for Linux with
> a free DVD software kit from IBM. Click Now!=20
> http://ads.osdn.com/?ad_id=1356&alloc_id438> &op=3Dick
>=20
> _______________________________________________
>=20
> Gamedevlists-general mailing list=20
> Gam...@li...
> https://lists.sourceforge.net/lists/listinfo/gamedevlists-general
> Archives: http://sourceforge.net/mailarchive/forum.php?forum_idU7
>=20
>=20

Re: [GD-General] time

[Mike W. <mi...@ge...>]

for multiplayer games like counter-strike or bf9142, i think that the=20
engine completely 'restarts itself' on every map change to help avoid=20
these types of errors.

i recall a bug in quake that would crash it every 'x' hours on clockwork=20
if left running in 'demo mode' for example...

for a massively multiplayer game though you would want to use something=20
more robust indeed.

other than relying on the system clock itself (or a system-specific=20
high-performance multi-media timer), i'm not sure how i'd approach this.

we use the windows internal clock for our own timing, and have routines=20
that check to see if we've 'fallen off the track' at any point, if so,=20
we do a sync between our internal game 'clock' and the windows system clo=
ck.

for multiplayer games, you may want all games (client or server) to rely=20
on the 'master server' clock possibly, not on a frame-by-frame basis,=20
but on a larger 'sync' basis, where you can make sure that all clients=20
are synced to the same timestamp?

i'm not sure how else you'd approach this, aside from just forcing the=20
clients to sync to the server's time, whatever it happens to=20
be...clients would add or drop frames as neceessary to stay in sync...

mike w
www.gekidodesigns.com

Brett Bibby wrote:
> Howdy, I have been wrestling with time in my engine for some time now
> (no pun intended). I wondered how other people are handling time? Our
> game engine is specifically designed for PS2/Gcn, but it runs under
> Windows too. Specifically...
>=20
> What are the different components that people are using? We currently
> have a single 32-bit time variable initialized to zero at startup and
> that's it. Each frame we either count v-blanks, use system
> ticks/timers or OS facilities custom for each platform to increment
> the global time value, usually stored in microseconds and rounded up
> to millisecond resolution for use in the engine. Works well and I
> have done this for every game engine I have developed so far.
>=20
> Feels like a change may be warranted though as online gaming becomes
> more popular, since if the game was to run for many days it would
> overflow and timestamping may be problematic. So now I'm thinking of
> adding a time struct (different from ANSI C though) whereby we store
> the year, month, day, hour and minute, and second and also a
> conversion struct that says how many mins->hour->day->month->year so
> we can convert "real" time into game time. I can pack this into
> 64-bits, but then again a simple 64-bit timer using the old way to
> simply count ticks would also eliminate the overflow problem anyway
> and so I'm questioning the right thing to do. Also, I hoping to be
> able to use absolute time for certain things where we used to use
> delta times. What are people using for timestamps for online games?
> Does each game session reset the server session time to prevent
> eventual overflow? Or is a more robust timing mechanism used? Are
> people using microseconds? milliseconds?
>=20
> Thanks, Brett
>=20
>=20
> ------------------------------------------------------- SF.Net is
> sponsored by: Speed Start Your Linux Apps Now. Build and deploy apps
> & Web services for Linux with a free DVD software kit from IBM. Click
> Now! http://ads.osdn.com/?ad_id=1356&alloc_id438&op=CCk=20
> _______________________________________________ Gamedevlists-general
> mailing list Gam...@li...=20
> https://lists.sourceforge.net/lists/listinfo/gamedevlists-general=20
> Archives: http://sourceforge.net/mailarchive/forum.php?forum_idU7
>=20
>=20

[GD-General] time

[Brett B. <res...@ga...>]

Howdy,
I have been wrestling with time in my engine for some time now (no pun =
intended). I wondered how other people are handling time? Our game =
engine is specifically designed for PS2/Gcn, but it runs under Windows =
too. Specifically...

What are the different components that people are using? We currently =
have a single 32-bit time variable initialized to zero at startup and =
that's it. Each frame we either count v-blanks, use system ticks/timers =
or OS facilities custom for each platform to increment the global time =
value, usually stored in microseconds and rounded up to millisecond =
resolution for use in the engine. Works well and I have done this for =
every game engine I have developed so far.

Feels like a change may be warranted though as online gaming becomes =
more popular, since if the game was to run for many days it would =
overflow and timestamping may be problematic. So now I'm thinking of =
adding a time struct (different from ANSI C though) whereby we store the =
year, month, day, hour and minute, and second and also a conversion =
struct that says how many mins->hour->day->month->year so we can convert =
"real" time into game time. I can pack this into 64-bits, but then again =
a simple 64-bit timer using the old way to simply count ticks would also =
eliminate the overflow problem anyway and so I'm questioning the right =
thing to do. Also, I hoping to be able to use absolute time for certain =
things where we used to use delta times. What are people using for =
timestamps for online games? Does each game session reset the server =
session time to prevent eventual overflow? Or is a more robust timing =
mechanism used? Are  people using microseconds? milliseconds?

Thanks,
Brett

[GD-General] New Physics middleware

[Pierre T. <pie...@no...>]

Hi guys,

I'm happy to announce the release of our new physics SDK. You can grab it
here:
http://www.novodex.com/downloads/NovodeX_SDK_Personal_2.02.exe

...the demos are worth watching... :)

- Pierre

PS: I think this is not off-topic for GD-General, is it ?