Csound Csound-dev Csound-tekno Search About

Re: Strange segment violation with 5.07

Date2007-10-26 15:29
FromMichael Gogins
SubjectRe: Strange segment violation with 5.07
There are some MSVC DLLs that fluidOpcodes.dll now needs, which I omitted from the installer. I will be uploading a 5.07.1 installer soon to fix this.

Regards,
Mike

-----Original Message-----
>From: Davis Pyon 
>Sent: Oct 26, 2007 7:12 AM
>To: csound@lists.bath.ac.uk
>Subject: Re: [Csnd] Strange segment violation with 5.07
>
>I am also experiencing crashes with the 5.07 C API, but with different
>crash behavior.  I removed fluidOpcodes.dll from Csound's plugin
>directory and now I no longer have crashes.  Could you try the same and
>let us know what happens?
>
>Davis
>
>----- Original Message ----
>From: Jean-Pierre Lemoine 
>To: csound@lists.bath.ac.uk
>Sent: Thursday, October 25, 2007 11:12:32 AM
>Subject: [Csnd] Strange segment violation with 5.07
>
>
>My Java code using the Csound API is running without any problems with 
>5.06. I use the program for long work sessions, and it's perfect.
>The same code is now causing segment violation when running with 5.07 
>double precision on Vista. It is possible that my logic in using the 
>csound APIs is not so good, but it is so strange as you can see below 
>that ... ???
> 
>Case 1
>I create a Csound Instance, compile a CSD file with this instance, then
> 
>start a Java thread which performs the CSD file in its run method, and 
>at the end of the Perform(), call Reset().
>When I quit the program, I check that the thread is not alive, and then
> 
>I destroy resources linked to the OpenGL library (LWJGL). At this step,
> 
>I got a segment violation in LWJGL (it's was ok with 5.06)
>I change Reset() by Cleanup(), and ... OK, no more segment voilation! 
>Thus I think this is in the way I use the APIs, but
>
>Case 2
>When I switch the program in full screen mode, and come back in
> windowed 
>mode, I get also a segment violation when quitting the program (during 
>the same step at case 1). In this case, the error log tells me that 
>there is an error in [sfont.dll+0x89d5].
>This is in when launching the application inside Eclipse. If I launch 
>the application standalone, outside of Eclipse, I get a segment 
>violation with a different DLL.
>
>Any ideas?
>Thanks
>Jean-Pierre
>
>
>
>
>
>
>
>-- 
>Send bugs reports to this list.
>To unsubscribe, send email to csound-unsubscribe@lists.bath.ac.uk
>
>
>
>--
>Send bugs reports to this list.
>To unsubscribe, send email to csound-unsubscribe@lists.bath.ac.uk


Date2007-10-26 18:04
FromJean-Pierre Lemoine
SubjectRe: Strange segment violation with 5.07
Thanks Mike for looking at this problem.

Jean-Pierre

Michael Gogins a écrit :
> There are some MSVC DLLs that fluidOpcodes.dll now needs, which I omitted from the installer. I will be uploading a 5.07.1 installer soon to fix this.
>
> Regards,
> Mike
>
> -----Original Message-----
>   
>> From: Davis Pyon 
>> Sent: Oct 26, 2007 7:12 AM
>> To: csound@lists.bath.ac.uk
>> Subject: Re: [Csnd] Strange segment violation with 5.07
>>
>> I am also experiencing crashes with the 5.07 C API, but with different
>> crash behavior.  I removed fluidOpcodes.dll from Csound's plugin
>> directory and now I no longer have crashes.  Could you try the same and
>> let us know what happens?
>>
>> Davis
>>
>> ----- Original Message ----
>> From: Jean-Pierre Lemoine 
>> To: csound@lists.bath.ac.uk
>> Sent: Thursday, October 25, 2007 11:12:32 AM
>> Subject: [Csnd] Strange segment violation with 5.07
>>
>>
>> My Java code using the Csound API is running without any problems with 
>> 5.06. I use the program for long work sessions, and it's perfect.
>> The same code is now causing segment violation when running with 5.07 
>> double precision on Vista. It is possible that my logic in using the 
>> csound APIs is not so good, but it is so strange as you can see below 
>> that ... ???
>>
>> Case 1
>> I create a Csound Instance, compile a CSD file with this instance, then
>>
>> start a Java thread which performs the CSD file in its run method, and 
>> at the end of the Perform(), call Reset().
>> When I quit the program, I check that the thread is not alive, and then
>>
>> I destroy resources linked to the OpenGL library (LWJGL). At this step,
>>
>> I got a segment violation in LWJGL (it's was ok with 5.06)
>> I change Reset() by Cleanup(), and ... OK, no more segment voilation! 
>> Thus I think this is in the way I use the APIs, but
>>
>> Case 2
>> When I switch the program in full screen mode, and come back in
>> windowed 
>> mode, I get also a segment violation when quitting the program (during 
>> the same step at case 1). In this case, the error log tells me that 
>> there is an error in [sfont.dll+0x89d5].
>> This is in when launching the application inside Eclipse. If I launch 
>> the application standalone, outside of Eclipse, I get a segment 
>> violation with a different DLL.
>>
>> Any ideas?
>> Thanks
>> Jean-Pierre
>>
>>
>>
>>
>>
>>
>>
>> -- 
>> Send bugs reports to this list.
>> To unsubscribe, send email to csound-unsubscribe@lists.bath.ac.uk
>>
>>
>>
>> --
>> Send bugs reports to this list.
>> To unsubscribe, send email to csound-unsubscribe@lists.bath.ac.uk
>>     
>
>
>
>