Bug#880011: csound: Macro system on csound make segfault

[Mickael Viey]

Le 10/11/2017 à 01:10, Felipe Sateler a écrit :
> Control: forwarded -1 https://github.com/csound/csound/issues/871
>
> 2017-11-08 16:45 GMT-03:00 Mickael Viey <m.viey at wanadoo.fr 
> <mailto:m.viey at wanadoo.fr>>:
>
>     Le 08/11/2017 à 14:19, Felipe Sateler a écrit :
>
>         Control: forcemerge -1 880010
>         Control: tags -1 moreinfo
>
>         On Sat, Oct 28, 2017 at 7:54 AM, Mickael Viey
>         <m.viey at wanadoo.fr <mailto:m.viey at wanadoo.fr>> wrote:
>
>             Package: csound
>
>             Version: 1:6.08.0~dfsg-1
>             Severity: normal
>
>             Dear Maintainer,
>
>
>             I have some csd files I made on another platforms which
>             generate
>             segfault when running with this version of csound. After
>             diagnostic It
>             appears that the macro system which is responsible.
>
>             By example, this macro works:
>             #define TIME # 0.25 #
>             #define freq # 146 #
>
>
>             #define T # $freq #
>             #define Am # $freq * 17 / 16 #
>             #define AM # $freq * 9/8 #
>
>             But If I had this after te last line:
>
>             #define E # $freq #
>
>             I get a segfault. I will join the complete csd file I used
>             for my tests.
>
>         Weird. I can reproduce on a stretch docker container, but I can't
>         reproduce on sid, either with 6.09 or rebuilding 6.08 . I ran
>         out of
>         time, but could you try rebuilding csound and see if the problem
>         persists?
>
>     Hi,
>
>     I am not sure what you mean by "rebuild" but this is what I done:
>
>
> Thanks, but I meant recompile the csound sources, not just reinstall.
>
> Anyway, I have debugged some more. Turns out the error in sid is still 
> present, it just doesn't cause a crash. I have forwarded upstream to 
> the above url.
>
> The error is not due to the macro, but due to the 3/2 p7. This is 
> invalid syntax, and that triggers the crash/malfunction (the crash is 
> in the error printing code, heh). You can work around the problem by 
> fixing the syntax error and passing 1.5.
>
>
> -- 
>
> Saludos,
> Felipe Sateler

That's right. In fact I can write expressions in score file but I must 
write them like this:

[  3 / 2 ]

On my previous install I doesn't have an error but I realize that the 
expression was interpreted like a 3 and not 3 / 2 .


If I reduce the number of space characters in the score I get the 
previous behavior (but I get 3 not 1.5).


I join the two versions of the score. One with a bad syntax but without 
seg fault when rendering, the other with the correct syntax.


Hoping they help you, (I really want have the time to debug this with you),


Thanks a lot for your help,

Best,



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.alioth.debian.org/pipermail/pkg-multimedia-maintainers/attachments/20171110/28ad5eed/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: bad_syntax_wihout_segfault.csd
Type: audio/csound
Size: 979 bytes
Desc: not available
URL: <http://lists.alioth.debian.org/pipermail/pkg-multimedia-maintainers/attachments/20171110/28ad5eed/attachment-0002.csd>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: correct_syntax.csd
Type: audio/csound
Size: 981 bytes
Desc: not available
URL: <http://lists.alioth.debian.org/pipermail/pkg-multimedia-maintainers/attachments/20171110/28ad5eed/attachment-0003.csd>