I'm trying to load an object named [mtx_*~]
I have the proper load path installed. I can load other objects in the same directory. If I look for the library in that directory (/usr/lib/pd/extra/iemmatrix), I find a file named mtx_0x2a0x7e.pd_linux. The hex translates to *~ in ASCII. So this should be the proper library to load for the [mtx_*~] object, right?
I'm unaware of how to debug this since it looks like I'm doing things right.
-lee
Heya, I switched this to IRC and got some help there. The solution for my particular problem was to compile iemmatrix as a single library and add it to the startup arguments. The underlying problem is that it's probably an incompatibility between the version of iemmatrix that is included in pd-extended and the version of hexloader, which should handle this kind of library naming.
-lee
On Sun, Apr 11, 2010 at 4:36 PM, Lee Azzarello lee@rockingtiger.com wrote:
I'm trying to load an object named [mtx_*~]
I have the proper load path installed. I can load other objects in the same directory. If I look for the library in that directory (/usr/lib/pd/extra/iemmatrix), I find a file named mtx_0x2a0x7e.pd_linux. The hex translates to *~ in ASCII. So this should be the proper library to load for the [mtx_*~] object, right?
I'm unaware of how to debug this since it looks like I'm doing things right.
-lee
In Pd-extended you can use [iemmatrix/mtx_mul~].
.hc
On Apr 11, 2010, at 11:55 PM, Lee Azzarello wrote:
Heya, I switched this to IRC and got some help there. The solution for my particular problem was to compile iemmatrix as a single library and add it to the startup arguments. The underlying problem is that it's probably an incompatibility between the version of iemmatrix that is included in pd-extended and the version of hexloader, which should handle this kind of library naming.
-lee
On Sun, Apr 11, 2010 at 4:36 PM, Lee Azzarello lee@rockingtiger.com wrote:
I'm trying to load an object named [mtx_*~]
I have the proper load path installed. I can load other objects in the same directory. If I look for the library in that directory (/usr/lib/pd/extra/iemmatrix), I find a file named mtx_0x2a0x7e.pd_linux. The hex translates to *~ in ASCII. So this should be the proper library to load for the [mtx_*~] object, right?
I'm unaware of how to debug this since it looks like I'm doing things right.
-lee
Pd-dev mailing list Pd-dev@iem.at http://lists.puredata.info/listinfo/pd-dev
----------------------------------------------------------------------------
Computer science is no more related to the computer than astronomy is related to the telescope. -Edsger Dykstra
Thanks. I know that but I'm not looking forward to find and replace on all the objects in pdx7~. I ended up compiling iemmatrix as a single library. Now I'm stuck on getting pool and memento to work, which is a wild journey of configuration and compiling and linking akin to recompiling the linux kernel to support a single piece of hardware that's really rad.
-lee
On Mon, Apr 12, 2010 at 11:44 AM, Hans-Christoph Steiner hans@at.or.at wrote:
In Pd-extended you can use [iemmatrix/mtx_mul~].
.hc
On Apr 11, 2010, at 11:55 PM, Lee Azzarello wrote:
Heya, I switched this to IRC and got some help there. The solution for my particular problem was to compile iemmatrix as a single library and add it to the startup arguments. The underlying problem is that it's probably an incompatibility between the version of iemmatrix that is included in pd-extended and the version of hexloader, which should handle this kind of library naming.
-lee
On Sun, Apr 11, 2010 at 4:36 PM, Lee Azzarello lee@rockingtiger.com wrote:
I'm trying to load an object named [mtx_*~]
I have the proper load path installed. I can load other objects in the same directory. If I look for the library in that directory (/usr/lib/pd/extra/iemmatrix), I find a file named mtx_0x2a0x7e.pd_linux. The hex translates to *~ in ASCII. So this should be the proper library to load for the [mtx_*~] object, right?
I'm unaware of how to debug this since it looks like I'm doing things right.
-lee
Pd-dev mailing list Pd-dev@iem.at http://lists.puredata.info/listinfo/pd-dev
Computer science is no more related to the computer than astronomy is related to the telescope. -Edsger Dykstra
Use sed or perl or something command line. You could also use sssad instead of memento, and skip pool.
.hc
On Apr 12, 2010, at 7:40 PM, Lee Azzarello wrote:
Thanks. I know that but I'm not looking forward to find and replace on all the objects in pdx7~. I ended up compiling iemmatrix as a single library. Now I'm stuck on getting pool and memento to work, which is a wild journey of configuration and compiling and linking akin to recompiling the linux kernel to support a single piece of hardware that's really rad.
-lee
On Mon, Apr 12, 2010 at 11:44 AM, Hans-Christoph Steiner <hans@at.or.at
wrote:
In Pd-extended you can use [iemmatrix/mtx_mul~].
.hc
On Apr 11, 2010, at 11:55 PM, Lee Azzarello wrote:
Heya, I switched this to IRC and got some help there. The solution for my particular problem was to compile iemmatrix as a single library and add it to the startup arguments. The underlying problem is that it's probably an incompatibility between the version of iemmatrix that is included in pd-extended and the version of hexloader, which should handle this kind of library naming.
-lee
On Sun, Apr 11, 2010 at 4:36 PM, Lee Azzarello lee@rockingtiger.com wrote:
I'm trying to load an object named [mtx_*~]
I have the proper load path installed. I can load other objects in the same directory. If I look for the library in that directory (/usr/lib/pd/extra/iemmatrix), I find a file named mtx_0x2a0x7e.pd_linux. The hex translates to *~ in ASCII. So this should be the proper library to load for the [mtx_*~] object, right?
I'm unaware of how to debug this since it looks like I'm doing things right.
-lee
Pd-dev mailing list Pd-dev@iem.at http://lists.puredata.info/listinfo/pd-dev
Computer science is no more related to the computer than astronomy is related to the telescope. -Edsger Dykstra
----------------------------------------------------------------------------
Using ReBirth is like trying to play an 808 with a long stick. - David Zicarelli
I just noticed sssad, though I was led in another direction by documentation at footils.org. I searched that site for the keyword sssad and there were zero results, so I figured that memento and pool was the better documented. It also looks like sssad is not included in the pd-extended binary for debian squeeze, though it is in the source archive.
Since both systems have the same author, would you recommend ssad over memento and pool?
Also, if Frank Barknecht is on this list, I have been updating your pdx7 V2 synthesizer to run without errors in polyphonic mode for the latest pd-extended release. The state saving is the last step before I can publish the repository I'm working in.
-lee
On Mon, Apr 12, 2010 at 8:39 PM, Hans-Christoph Steiner hans@at.or.at wrote:
Use sed or perl or something command line. You could also use sssad instead of memento, and skip pool.
.hc
On Apr 12, 2010, at 7:40 PM, Lee Azzarello wrote:
Thanks. I know that but I'm not looking forward to find and replace on all the objects in pdx7~. I ended up compiling iemmatrix as a single library. Now I'm stuck on getting pool and memento to work, which is a wild journey of configuration and compiling and linking akin to recompiling the linux kernel to support a single piece of hardware that's really rad.
-lee
On Mon, Apr 12, 2010 at 11:44 AM, Hans-Christoph Steiner hans@at.or.at wrote:
In Pd-extended you can use [iemmatrix/mtx_mul~].
.hc
On Apr 11, 2010, at 11:55 PM, Lee Azzarello wrote:
Heya, I switched this to IRC and got some help there. The solution for my particular problem was to compile iemmatrix as a single library and add it to the startup arguments. The underlying problem is that it's probably an incompatibility between the version of iemmatrix that is included in pd-extended and the version of hexloader, which should handle this kind of library naming.
-lee
On Sun, Apr 11, 2010 at 4:36 PM, Lee Azzarello lee@rockingtiger.com wrote:
I'm trying to load an object named [mtx_*~]
I have the proper load path installed. I can load other objects in the same directory. If I look for the library in that directory (/usr/lib/pd/extra/iemmatrix), I find a file named mtx_0x2a0x7e.pd_linux. The hex translates to *~ in ASCII. So this should be the proper library to load for the [mtx_*~] object, right?
I'm unaware of how to debug this since it looks like I'm doing things right.
-lee
Pd-dev mailing list Pd-dev@iem.at http://lists.puredata.info/listinfo/pd-dev
Computer science is no more related to the computer than astronomy is related to the telescope. -Edsger Dykstra
Using ReBirth is like trying to play an 808 with a long stick. -David Zicarelli
Could it possibly be that iemmatrix is not [hexloader]-ready yet?
From the hexloader README:
«CAVEAT: C also forbids to start function names with numeric values. therefore, this is a nono: [>]'s setupfunction would be "0x3e_setup()" (ILLEGAL) we can simply fix this by putting the "setup()" in front: [>] has a setupfunction "setup_0x3e()"»
Looking at the source, for instance mtx_mul.c, I don't see any function with prefix 'setup', only with 'setup' as suffix. While having another look it appears that there is even no setup-function name with hexadecimals in the name.
If it is only a matter of someone doing the boring uninspiring uninteresting painful work of renaming those functions, I _might_ be able to help, if there is interest (IOhannes?). If so, I would probably need some advice on how to do it correctly.
Roman
On Sun, 2010-04-11 at 16:36 -0400, Lee Azzarello wrote:
I'm trying to load an object named [mtx_*~]
I have the proper load path installed. I can load other objects in the same directory. If I look for the library in that directory (/usr/lib/pd/extra/iemmatrix), I find a file named mtx_0x2a0x7e.pd_linux. The hex translates to *~ in ASCII. So this should be the proper library to load for the [mtx_*~] object, right?
I'm unaware of how to debug this since it looks like I'm doing things right.
-lee
Pd-dev mailing list Pd-dev@iem.at http://lists.puredata.info/listinfo/pd-dev
On 2010-04-14 00:19, Roman Haefeli wrote:
If it is only a matter of someone doing the boring uninspiring uninteresting painful work of renaming those functions, I _might_ be able to help, if there is interest (IOhannes?). If so, I would probably need some advice on how to do it correctly.
run Pd in verbose mode and load hexloader. try to instantiate [mtx_*~] and watch out for the exact name of the setup fun. create a stub function of that very name that points to the real setup function (look at "iemtx_mul__setup()" for how this is done)
obviously this could be done in a more intelligent way, by manually calculating the setup-function name.
also be sure to add stubs for the hexloader built into Pd whenever they differ (just repeat the above without hexloader loaded, and see what's going on; you probably need to provide dummy libs like "mtx_*~.pd_linux" in order to see the requested name) IIRC, the hexloader built-into Pd does everything but filename-mangling; therefore you still need the "hexloader" lib in order to load libraries with "illegal" filenames (e.g. you cannot have a file named "mtx_>.dll" on an NTFS/FAT32 filesystem)
fgmasr IOhannes
Roman
On Sun, 2010-04-11 at 16:36 -0400, Lee Azzarello wrote:
I'm trying to load an object named [mtx_*~]
I have the proper load path installed. I can load other objects in the same directory. If I look for the library in that directory (/usr/lib/pd/extra/iemmatrix), I find a file named mtx_0x2a0x7e.pd_linux. The hex translates to *~ in ASCII. So this should be the proper library to load for the [mtx_*~] object, right?
I'm unaware of how to debug this since it looks like I'm doing things right.
-lee
Pd-dev mailing list Pd-dev@iem.at http://lists.puredata.info/listinfo/pd-dev
Pd-dev mailing list Pd-dev@iem.at http://lists.puredata.info/listinfo/pd-dev
On Wed, 2010-04-14 at 10:12 +0200, IOhannes m zmoelnig wrote:
On 2010-04-14 00:19, Roman Haefeli wrote:
If it is only a matter of someone doing the boring uninspiring uninteresting painful work of renaming those functions, I _might_ be able to help, if there is interest (IOhannes?). If so, I would probably need some advice on how to do it correctly.
run Pd in verbose mode and load hexloader. try to instantiate [mtx_*~] and watch out for the exact name of the setup fun. create a stub function of that very name that points to the real setup function (look at "iemtx_mul__setup()" for how this is done)
obviously this could be done in a more intelligent way, by manually calculating the setup-function name.
also be sure to add stubs for the hexloader built into Pd whenever they differ (just repeat the above without hexloader loaded, and see what's going on; you probably need to provide dummy libs like "mtx_*~.pd_linux" in order to see the requested name) IIRC, the hexloader built-into Pd does everything but filename-mangling; therefore you still need the "hexloader" lib in order to load libraries with "illegal" filenames (e.g. you cannot have a file named "mtx_>.dll" on an NTFS/FAT32 filesystem)
Thanks for the help. Actually the main work was already done by Georg Holzmann. Only very few were missing (probably new classes?). Of course, I initially tried one of those, which made me assume, there is no hexloader support in iemmatrix.
I submitted a patch to tracker. Hopefully this fixes all aliases.
Roman
-
Hans-Christoph Steiner -
IOhannes m zmoelnig -
Lee Azzarello -
Roman Haefeli