Patches item #1930748, was opened at 2008-04-01 01:35
Message generated for change (Settings changed) made by xovo
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1930748&group_…
Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
>Category: puredata-dev
Group: bugfix
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Thomas Grill (xovo)
>Assigned to: Miller Puckette (millerpuckette)
Summary: fix type-punned pointers (part 3)
Initial Comment:
oh, forgot to say, those patches are against pd-0.41-4
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1930748&group_…
Patches item #1930747, was opened at 2008-04-01 01:33
Message generated for change (Settings changed) made by xovo
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1930747&group_…
Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
>Category: puredata-dev
Group: bugfix
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Thomas Grill (xovo)
>Assigned to: Miller Puckette (millerpuckette)
Summary: fix type-punned pointers (part 2)
Initial Comment:
part 1 continued
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1930747&group_…
Patches item #1930745, was opened at 2008-04-01 01:30
Message generated for change (Settings changed) made by xovo
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1930745&group_…
Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
>Category: puredata-dev
Group: bugfix
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Thomas Grill (xovo)
>Assigned to: Miller Puckette (millerpuckette)
Summary: fix type-punned pointers (part 1)
Initial Comment:
optimization with strict-aliasing chokes on type-punned pointers. Type-casting is replaced with unions.
Using static functions in m_pd.h is not nice, but those will be inlined by any decent compilers anyway.
Even better would be checking for the inline keyword with autoconf and use that throughout pd when available.
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1930745&group_…
Patches item #1930769, was opened at 2008-04-01 02:09
Message generated for change (Settings changed) made by xovo
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1930769&group_…
Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
>Category: puredata-dev
Group: bugfix
Status: Open
Resolution: None
>Priority: 7
Private: No
Submitted By: Thomas Grill (xovo)
>Assigned to: Miller Puckette (millerpuckette)
Summary: 0.41-4 fix memory leak with array sinesum/cosinesum
Initial Comment:
svecs are allocated twice....
--- ../src.ori/g_array.c 2008-03-15 01:03:00.000000000 +0100
+++ g_array.c 2008-04-01 04:05:50.000000000 +0200
@@ -1221,7 +1221,7 @@
static void garray_sinesum(t_garray *x, t_symbol *s, int argc, t_atom *argv)
{
- t_float *svec = (t_float *)t_getbytes(sizeof(t_float) * argc);
+ t_float *svec;
int npoints, i;
if (argc < 2)
{
@@ -1244,7 +1244,7 @@
static void garray_cosinesum(t_garray *x, t_symbol *s, int argc, t_atom *argv)
{
- t_float *svec = (t_float *)t_getbytes(sizeof(t_float) * argc);
+ t_float *svec;
int npoints, i;
if (argc < 2)
{
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1930769&group_…
Patches item #1930733, was opened at 2008-04-01 01:02
Message generated for change (Settings changed) made by xovo
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1930733&group_…
Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
>Category: puredata-dev
Group: bugfix
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Thomas Grill (xovo)
>Assigned to: Miller Puckette (millerpuckette)
Summary: 0.41-4: prevent buffer overrun in m_class.c
Initial Comment:
the array only has MAXPDARG elements.......
--- m_class.ori.c 2008-04-01 03:00:09.000000000 +0200
+++ m_class.c 2008-04-01 03:00:12.000000000 +0200
@@ -763,7 +763,7 @@
va_start(ap, fmt);
while (1)
{
- if (nargs > MAXPDARG)
+ if (nargs >= MAXPDARG)
{
pd_error(x, "pd_vmess: only %d allowed", MAXPDARG);
break;
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1930733&group_…
Patches item #1930681, was opened at 2008-03-31 23:44
Message generated for change (Settings changed) made by xovo
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1930681&group_…
Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
>Category: puredata-dev
Group: bugfix
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Thomas Grill (xovo)
>Assigned to: Miller Puckette (millerpuckette)
Summary: 0.41-4 s_path.c fixes
Initial Comment:
Two fixes for buffer overrun (buf[1000]=0 later) and rc file args:
--- s_path.ori.c 2008-04-01 01:37:03.000000000 +0200
+++ s_path.c 2008-04-01 01:38:02.000000000 +0200
@@ -379,7 +379,7 @@
int rcargc;
char* rcargv[NUMARGS];
char* buffer;
- char fname[MAXPDSTRING], buf[1000], *home = getenv("HOME");
+ char fname[MAXPDSTRING], buf[1001], *home = getenv("HOME");
int retval = 1; /* that's what we will return at the end; for now, let's think it'll be an error */
/* initialize rc-arg-array so we can safely clean up at the end */
@@ -423,7 +423,7 @@
fclose(file);
if (sys_verbose)
{
- if (rcargv)
+ if (rcargc)
{
post("startup args from RC file:");
for (i = 1; i < rcargc; i++)
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1930681&group_…
Patches item #1933679, was opened at 2008-04-03 19:19
Message generated for change (Tracker Item Submitted) made by Item Submitter
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1933679&group_…
Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: puredata-dev
Group: feature
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Thomas Grill (xovo)
Assigned to: Nobody/Anonymous (nobody)
Summary: FFTW3 interface for pd-0.41-4
Initial Comment:
The attached patch implements the necessary functionality to emulate the mayer_fft with FFTW3. It's a strange approach and not ultimately efficient, but it's better than the plain mayer_fft in any case.
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1933679&group_…
Bugs item #1933444, was opened at 2008-04-03 15:02
Message generated for change (Tracker Item Submitted) made by Item Submitter
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478070&aid=1933444&group_…
Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: externals
Group: v0.41.0
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Julian Villegas (julovi)
Assigned to: Nobody/Anonymous (nobody)
Summary: sigmund~ freezes Pd in Mac OS X v.10.5.2
Initial Comment:
I've found that when using sigmund~ in Mac OS X v.10.5.2 and Pd-0.41-4, the patch freezes if I stop computing the audio. I mean, I can only close Pd from the Pd menu, and nothing else seems to work. I attached a simple patch to show my problem.
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478070&aid=1933444&group_…
Patches item #1931686, was opened at 2008-04-01 21:54
Message generated for change (Comment added) made by xovo
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1931686&group_…
Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: puredata
Group: feature
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Thomas Grill (xovo)
Assigned to: Miller Puckette (millerpuckette)
Summary: functionality for 128-bit-aligned memory (against pd-0.41-4)
Initial Comment:
This patch implements public functions for allocation of 128-bit aligned memory.
For allocation of all signals and buffers within PD these functions are consequently used.
128-bit alignment can speed up vector calculations, esp. if used with SIMD (also with autovectorization techniques).
The infrastructure has been part of devel_0_39 and has now been ported to vanilla pd.
----------------------------------------------------------------------
>Comment By: Thomas Grill (xovo)
Date: 2008-04-03 14:15
Message:
Logged In: YES
user_id=350252
Originator: YES
Yes, you are right, but the current implementation wastes a bit of memory
which might not be acceptable (although it's only 4-19 bytes per block).
The implementation is portable because it relies on the existing
functions, but there's also the possibility to rewrite getbytes() in a way
that the system provides allocated memory right from the start.
Under OS X it seems that plain malloc is ok for that
(http://developer.apple.com/documentation/Darwin/Reference/ManPages/man3/mal…)
under general POSIX there seems to be memalign
(http://man-wiki.net/index.php/3:posix_memalign) , under Windows there's
_aligned_malloc (http://msdn2.microsoft.com/en-us/library/8z34s9c6.aspx).
However, i didn't want to deal with platform issues.
----------------------------------------------------------------------
Comment By: Hans-Christoph Steiner (eighthave)
Date: 2008-04-03 13:45
Message:
Logged In: YES
user_id=27104
Originator: NO
This sounds like a very useful addition. I have one minor question about
the patch implementation. It seems that it would be more manageable if the
new functions had the same name as the originals, then the original and the
optimized versions could be swapped for each other with a single #ifdef
which could be controlled by a ./configure flag.
Is there some code where you don't want to use the aligned functions?
Redefining the original names would automatically give all the externals
aligned memory as well.
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1931686&group_…
Patches item #1931686, was opened at 2008-04-01 17:54
Message generated for change (Comment added) made by eighthave
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1931686&group_…
Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: puredata
Group: feature
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Thomas Grill (xovo)
>Assigned to: Miller Puckette (millerpuckette)
Summary: functionality for 128-bit-aligned memory (against pd-0.41-4)
Initial Comment:
This patch implements public functions for allocation of 128-bit aligned memory.
For allocation of all signals and buffers within PD these functions are consequently used.
128-bit alignment can speed up vector calculations, esp. if used with SIMD (also with autovectorization techniques).
The infrastructure has been part of devel_0_39 and has now been ported to vanilla pd.
----------------------------------------------------------------------
>Comment By: Hans-Christoph Steiner (eighthave)
Date: 2008-04-03 09:45
Message:
Logged In: YES
user_id=27104
Originator: NO
This sounds like a very useful addition. I have one minor question about
the patch implementation. It seems that it would be more manageable if the
new functions had the same name as the originals, then the original and the
optimized versions could be swapped for each other with a single #ifdef
which could be controlled by a ./configure flag.
Is there some code where you don't want to use the aligned functions?
Redefining the original names would automatically give all the externals
aligned memory as well.
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1931686&group_…