chuck/[mini]audicle vs. gcc 4.4.x
This tiny patch makes chuck compilable under gcc4.4 (Fedora 11):
-------
--- chuck-1.2.1.2/src/util_string.h~ 2008-07-16 02:55:17.000000000 -0700
+++ chuck-1.2.1.2/src/util_string.h 2009-06-11 18:24:55.349093274 -0700
@@ -36,6 +36,7 @@
#include "chuck_def.h"
#include <string>
#include <vector>
+#include
Fernando;
I'm also hitting this error in miniaudicle and audicle, any help?:
For the Mini it's most likely caused by the Mini still using 1.2.1.1 which doesn't play nicely with newer and more strict versions of GCC. The trick is to take your Mini source, look fo rthe ChucK dir and replace all of the files in that with the files from the source of ChucK 1.2.1.2. That will make it compile and you'll end up with a Mini that uses the latest version of ChucK. I don't think it's worth patching as the result of this workaround is superior to what you'd get with a patch for ChucK 1.2.1.1. For the Audicle this -sadly- doesn't work. I do like the Audicle and the ideas behind it but right now the only version available is based on a version of ChucK that's so badly out of date that I have a hard time reccomending using the Audicle for anything more serious than demonstrating experimental interfaces for programming. I hope that helps, Yours, Kas.
On Fri, 2009-06-12 at 19:25 +0200, Kassen wrote:
Fernando;
I'm also hitting this error in miniaudicle and audicle, any help?:
For the Mini it's most likely caused by the Mini still using 1.2.1.1 which doesn't play nicely with newer and more strict versions of GCC. The trick is to take your Mini source, look fo rthe ChucK dir and replace all of the files in that with the files from the source of ChucK 1.2.1.2. That will make it compile and you'll end up with a Mini that uses the latest version of ChucK. I don't think it's worth patching as the result of this workaround is superior to what you'd get with a patch for ChucK 1.2.1.1.
I did that and it seems to work fine. It does require patching the makefiles as well to include -fno-strict-aliasing to bypass an optimization problem (feature?) of the newer gcc (see another message in this thread).
For the Audicle this -sadly- doesn't work. I do like the Audicle and the ideas behind it but right now the only version available is based on a version of ChucK that's so badly out of date that I have a hard time reccomending using the Audicle for anything more serious than demonstrating experimental interfaces for programming.
Hmmm, I managed to get a hybrid version of the audicle going. It _seems_ to work but I'm not yet a chucker so I can't really test. I added chuck 1.2.1.2 (with a couple of patches), then some more patches to deal with const char ** issues (argv related) in the audicle codebase and voila, compiled fine and started. As before no _real_ testing.
I hope that helps,
Thanks... So there are now chuck/miniaudicle/audicle versions on fc11 that appear (maybe) to work. -- Fernando
Fernando Lopez-Lezcano wrote:
This tiny patch makes chuck compilable under gcc4.4 (Fedora 11):
Great, thanks. However after a successful build chuck segfaults on any files I throw at it. --loop works fine, but still segfaults when I add files. This is on arch-linux (everything was fine under my old ubuntu 8.04, but I'm trying to move away from that for various reasons). Anything I can do to narrow this down? -- Atte http://atte.dk http://modlys.dk http://virb.com/atte
On Mon, 2009-06-15 at 12:48 +0200, Atte Andre Jensen wrote:
Fernando Lopez-Lezcano wrote:
This tiny patch makes chuck compilable under gcc4.4 (Fedora 11):
Great, thanks.
However after a successful build chuck segfaults on any files I throw at it. --loop works fine, but still segfaults when I add files.
This is on arch-linux (everything was fine under my old ubuntu 8.04, but I'm trying to move away from that for various reasons).
Anything I can do to narrow this down?
I got a message off list regarding this, R. Steven Rainwater discovered that with gcc4.4 (ie: fc11) the optimization flags make a difference. -O3 is too much, either use -O2 or -O3 -fno-strict-aliasing (I chose the later for Planet CCRMA packages). It appears to work. -- Fernando
participants (3)
-
Atte Andre Jensen -
Fernando Lopez-Lezcano -
Kassen