Kurok PSP errors

[revelator]

ok this explains it The MIPS family has a wide variety of implementations by a wide range of vendors. Consequently, there are many, many CPU models within it. ouch no wonder its hard to find a default assembler.
and yes rtems got one
and bingo http://courses.missouristate.edu/KenVollmar/MARS/ heres the assembler

[Max_Salivan]

may be i miss something


http://rghost.ru/36549862

[Baker]

UPD:i am compile it
added
#define FL_MONSTERCLIP 8192 // Only solid to entities with FL_MONSTER

but doors and buttons dont work,i am pass through them

I'm glad you got it to compile. I can't help you with debugging your own code, but sv_physics.c and world.c are 2 places that might involve that kind of thing.

You might consider using WinMerge to compare the old and new files (especially sv_physics.c and world.c) to look through and try to determine what change is making that happen.

[Max_Salivan]

UPD:i am compile it
added
#define FL_MONSTERCLIP 8192 // Only solid to entities with FL_MONSTER

but doors and buttons dont work,i am pass through them

I'm glad you got it to compile. I can't help you with debugging your own code, but sv_physics.c and world.c are 2 places that might involve that kind of thing.

You might consider using WinMerge to compare the old and new files (especially sv_physics.c and world.c) to look through and try to determine what change is making that happen.

thanks,i cant find errors
i am place sv_physics.c and world.c from kurok
but doors and buttons dont work=\

[Baker]

There are physics changes in pr_cmds.c, sv_phys.c and world.c with stuff like this

Code: Select all

/*
==================
SV_ClipMoveToEntity

Handles selection or creation of a clipping hull, and offseting (and
eventually rotation) of the end points
==================
*/
trace_t SV_ClipMoveToEntity (edict_t *ent, vec3_t start, vec3_t mins, vec3_t maxs, vec3_t end)
{
	trace_t	trace;
	vec3_t	offset, temp;
	vec3_t	start_l, end_l;
	hull_t	*hull;
	qboolean	rotated;
	matrix4x4	matrix;

// fill in a default trace
	memset (&trace, 0, sizeof(trace_t));
	trace.fraction = 1;
	trace.allsolid = true;
	VectorCopy (end, trace.endpos);

// get the clipping hull
	hull = SV_HullForEntity (ent, mins, maxs, offset);

// g-cont. get support for rotational brush-models
	if( ent != EDICT_NUM( 0 ) && ent->v.solid == SOLID_BSP && !VectorIsNull( ent->v.angles ))
		rotated = true;
	else rotated = false;

	if( rotated )
	{
		Matrix4x4_CreateFromEntity( matrix, ent->v.angles, offset, 1.0f );
		Matrix4x4_VectorITransform( matrix, start, start_l );
		Matrix4x4_VectorITransform( matrix, end, end_l );
	}
	else
	{
		VectorSubtract( start, offset, start_l );
		VectorSubtract( end, offset, end_l );
	}

// trace a line through the apropriate clipping hull
	SV_RecursiveHullCheck (hull, hull->firstclipnode, 0, 1, start_l, end_l, &trace);

	if( trace.fraction != 1.0f )
	{
		// compute endpos (generic case)
		VectorLerp( start, trace.fraction, end, trace.endpos );

		if( rotated )
		{
			// transform plane
			VectorCopy( trace.plane.normal, temp );
			Matrix4x4_TransformPositivePlane( matrix, temp, trace.plane.dist, trace.plane.normal, &trace.plane.dist );
		}
		else
		{
			// make dist adjustments
			trace.plane.dist = DotProduct( trace.endpos, trace.plane.normal );
		}
	}

// did we clip the move?
	if (trace.fraction < 1.0f || trace.startsolid  )
		trace.ent = ent;

	return trace;
}

And this ...

Code: Select all

/*
==================
SV_ClipMoveToEntity

Handles selection or creation of a clipping hull, and offseting (and
eventually rotation) of the end points
==================
*/
trace_t SV_ClipMoveToEntity (edict_t *ent, vec3_t start, vec3_t mins, vec3_t maxs, vec3_t end)
{
	trace_t	trace;
	vec3_t	offset, temp;
	vec3_t	start_l, end_l;
	hull_t	*hull;
	qboolean	rotated;
	matrix4x4	matrix;

// fill in a default trace
	memset (&trace, 0, sizeof(trace_t));
	trace.fraction = 1;
	trace.allsolid = true;
	VectorCopy (end, trace.endpos);

// get the clipping hull
	hull = SV_HullForEntity (ent, mins, maxs, offset);

// g-cont. get support for rotational brush-models
	if( ent != EDICT_NUM( 0 ) && ent->v.solid == SOLID_BSP && !VectorIsNull( ent->v.angles ))
		rotated = true;
	else rotated = false;

	if( rotated )
	{
		Matrix4x4_CreateFromEntity( matrix, ent->v.angles, offset, 1.0f );
		Matrix4x4_VectorITransform( matrix, start, start_l );
		Matrix4x4_VectorITransform( matrix, end, end_l );
	}
	else
	{
		VectorSubtract( start, offset, start_l );
		VectorSubtract( end, offset, end_l );
	}

// trace a line through the apropriate clipping hull
	SV_RecursiveHullCheck (hull, hull->firstclipnode, 0, 1, start_l, end_l, &trace);

	if( trace.fraction != 1.0f )
	{
		// compute endpos (generic case)
		VectorLerp( start, trace.fraction, end, trace.endpos );

		if( rotated )
		{
			// transform plane
			VectorCopy( trace.plane.normal, temp );
			Matrix4x4_TransformPositivePlane( matrix, temp, trace.plane.dist, trace.plane.normal, &trace.plane.dist );
		}
		else
		{
			// make dist adjustments
			trace.plane.dist = DotProduct( trace.endpos, trace.plane.normal );
		}
	}

// did we clip the move?
	if (trace.fraction < 1.0f || trace.startsolid  )
		trace.ent = ent;

	return trace;
}

It isn't unheard of for floating point functions on the PSP to work differently, poorly, unexpectedly on a PSP. And there are physics changes and matrix transformations and so forth in all those files. You'd need to find someone dedicated to porting that and debugging that on a PSP, depending on what the issue could be it might take them a lot of time (weeks?) to identify why that is happening. Sorry.

I do hope you get it to work.

[Max_Salivan]

There are physics changes in pr_cmds.c, sv_phys.c and world.c with stuff like this

Code: Select all

/*
==================
SV_ClipMoveToEntity

Handles selection or creation of a clipping hull, and offseting (and
eventually rotation) of the end points
==================
*/
trace_t SV_ClipMoveToEntity (edict_t *ent, vec3_t start, vec3_t mins, vec3_t maxs, vec3_t end)
{
	trace_t	trace;
	vec3_t	offset, temp;
	vec3_t	start_l, end_l;
	hull_t	*hull;
	qboolean	rotated;
	matrix4x4	matrix;

// fill in a default trace
	memset (&trace, 0, sizeof(trace_t));
	trace.fraction = 1;
	trace.allsolid = true;
	VectorCopy (end, trace.endpos);

// get the clipping hull
	hull = SV_HullForEntity (ent, mins, maxs, offset);

// g-cont. get support for rotational brush-models
	if( ent != EDICT_NUM( 0 ) && ent->v.solid == SOLID_BSP && !VectorIsNull( ent->v.angles ))
		rotated = true;
	else rotated = false;

	if( rotated )
	{
		Matrix4x4_CreateFromEntity( matrix, ent->v.angles, offset, 1.0f );
		Matrix4x4_VectorITransform( matrix, start, start_l );
		Matrix4x4_VectorITransform( matrix, end, end_l );
	}
	else
	{
		VectorSubtract( start, offset, start_l );
		VectorSubtract( end, offset, end_l );
	}

// trace a line through the apropriate clipping hull
	SV_RecursiveHullCheck (hull, hull->firstclipnode, 0, 1, start_l, end_l, &trace);

	if( trace.fraction != 1.0f )
	{
		// compute endpos (generic case)
		VectorLerp( start, trace.fraction, end, trace.endpos );

		if( rotated )
		{
			// transform plane
			VectorCopy( trace.plane.normal, temp );
			Matrix4x4_TransformPositivePlane( matrix, temp, trace.plane.dist, trace.plane.normal, &trace.plane.dist );
		}
		else
		{
			// make dist adjustments
			trace.plane.dist = DotProduct( trace.endpos, trace.plane.normal );
		}
	}

// did we clip the move?
	if (trace.fraction < 1.0f || trace.startsolid  )
		trace.ent = ent;

	return trace;
}

And this ...

Code: Select all

/*
==================
SV_ClipMoveToEntity

Handles selection or creation of a clipping hull, and offseting (and
eventually rotation) of the end points
==================
*/
trace_t SV_ClipMoveToEntity (edict_t *ent, vec3_t start, vec3_t mins, vec3_t maxs, vec3_t end)
{
	trace_t	trace;
	vec3_t	offset, temp;
	vec3_t	start_l, end_l;
	hull_t	*hull;
	qboolean	rotated;
	matrix4x4	matrix;

// fill in a default trace
	memset (&trace, 0, sizeof(trace_t));
	trace.fraction = 1;
	trace.allsolid = true;
	VectorCopy (end, trace.endpos);

// get the clipping hull
	hull = SV_HullForEntity (ent, mins, maxs, offset);

// g-cont. get support for rotational brush-models
	if( ent != EDICT_NUM( 0 ) && ent->v.solid == SOLID_BSP && !VectorIsNull( ent->v.angles ))
		rotated = true;
	else rotated = false;

	if( rotated )
	{
		Matrix4x4_CreateFromEntity( matrix, ent->v.angles, offset, 1.0f );
		Matrix4x4_VectorITransform( matrix, start, start_l );
		Matrix4x4_VectorITransform( matrix, end, end_l );
	}
	else
	{
		VectorSubtract( start, offset, start_l );
		VectorSubtract( end, offset, end_l );
	}

// trace a line through the apropriate clipping hull
	SV_RecursiveHullCheck (hull, hull->firstclipnode, 0, 1, start_l, end_l, &trace);

	if( trace.fraction != 1.0f )
	{
		// compute endpos (generic case)
		VectorLerp( start, trace.fraction, end, trace.endpos );

		if( rotated )
		{
			// transform plane
			VectorCopy( trace.plane.normal, temp );
			Matrix4x4_TransformPositivePlane( matrix, temp, trace.plane.dist, trace.plane.normal, &trace.plane.dist );
		}
		else
		{
			// make dist adjustments
			trace.plane.dist = DotProduct( trace.endpos, trace.plane.normal );
		}
	}

// did we clip the move?
	if (trace.fraction < 1.0f || trace.startsolid  )
		trace.ent = ent;

	return trace;
}

It isn't unheard of for floating point functions on the PSP to work differently, poorly, unexpectedly on a PSP. And there are physics changes and matrix transformations and so forth in all those files. You'd need to find someone dedicated to porting that and debugging that on a PSP, depending on what the issue could be it might take them a lot of time (weeks?) to identify why that is happening. Sorry.

I do hope you get it to work.

oh thanks for help,Baker,I cant find one who will do it...
this is the end for Portal to PSP(i think)
without this physics,mod has many bugs with cubes and turrets-_-

[Baker]

oh thanks for help,Baker,I cant find one who will do it...
this is the end for Portal to PSP(i think)
without this physics,mod has many bugs with cubes and turrets-_-

That sucks

The PSP really hates some floating point stuff. I had to kill matrix functions once for an engine build to run. It wasn't that complex either, it mostly involved gluUnproject and a few matrix functions that requires. I think I also had to remove MH's chasecam fix for similar reasons if I recall correctly. I never achieved an understanding of why or what the issue was nor a work around, but I #ifdef'd the code out of the PSP build. The PSP and some types of floating point don't seem to mix.

And your code is using a lot of matrix functions.