[M3commit] CVS Update: cm3

[Jay]

Hey, this warning seems to be pretty important.

 

I have on birch a cm3 built with, I think, -O3, and it is very broken.

 

Like, I used to build m3cc, it exited after configure.

I cd'ed and make and ignored it.

Then I tried building m3core, it exited, roughly, after compiling

one file.

 

So, I break on exit and _exit and I end up here:

 

--- building in AMD64_LINUX ---

ignoring ../src/m3overrides

new source -> compiling RTHooks.i3
[Switching to Thread 47455981786832 (LWP 27346)]

Breakpoint 1, 0x00002b2934d52160 in _exit () from /lib/libc.so.6
(gdb) bt
#0  0x00002b2934d52160 in _exit () from /lib/libc.so.6
#1  0x00000000006390ee in ProcessPosixCommon__Create_ForkExec (M3_Bd56fi_cmd=Can
not access memory at address 0x8000761952a8
)
    at ../src/os/POSIX/ProcessPosixCommon.m3:90
#2  0x000000000063aa97 in Process__Create (M3_Bd56fi_cmd=Cannot access memory at


Now, I haven't tried -O3 in over a week, and never before that.

I don't use birch much.

I know I tried -O3 in the past few weeks.

I guess I tried it on birch.

So this might not even be -O3, but it probably is.

I'll go back and build with no optimization or maybe -O2 and come back to this..later.. no guarantee of when "later" is.


 - Jay
 


From: jay.krell at cornell.edu
To: hosking at cs.purdue.edu
Date: Mon, 11 May 2009 07:01:03 +0000
CC: m3commit at elegosoft.com
Subject: Re: [M3commit] CVS Update: cm3



Uh, any locals referenced in a finally block?
  And whatever locals are referenced by try/except/raise/lock?
If we have that capability easily.
Anyway, parse.c doesn't change much.
 
 - Jay
 


From: hosking at cs.purdue.edu
To: jay.krell at cornell.edu
Date: Mon, 11 May 2009 16:54:26 +1000
CC: m3commit at elegosoft.com
Subject: Re: [M3commit] CVS Update: cm3




On 11 May 2009, at 16:49, Jay wrote:


It is, you know, a reasonable message.


I know.


One might wonder how we avoided it?


I don't recall seeing it before and wonder if something has changed in the way locals are handled since last I tried -O3.


Mark all locals volatile in functions with try/raise/fatal?


I suppose, though I think we'd like to be smarter.



 
 - Jay

 


From: jay.krell at cornell.edu
To: hosking at cs.purdue.edu
Date: Mon, 11 May 2009 06:47:49 +0000
CC: m3commit at elegosoft.com
Subject: Re: [M3commit] CVS Update: cm3

I don't think so.
 
  - Jay
 


From: hosking at cs.purdue.edu
To: hosking at cs.purdue.edu
Date: Mon, 11 May 2009 15:19:31 +1000
CC: m3commit at elegosoft.com; jay.krell at cornell.edu
Subject: Re: [M3commit] CVS Update: cm3




Jay,


Have you changed something about how the compiler invokes setjmp/longjmp for exception handling?


-- Tony


On 11 May 2009, at 15:18, Tony Hosking wrote:

We didn't used to get these errors with -O3.  I don't know what has changed with the compiler backend, but something new is happening.  We should check the logs to see the last time I pushed through -O3, and figure out what changed since. 






On 11 May 2009, at 14:48, Jay wrote:

I got a lot of those warnings when I tried -O3.
And one error running the result that I didn't look into.
 
 - Jay

 
> Date: Mon, 11 May 2009 05:24:22 +0000
> To: m3commit at elegosoft.com
> From: hosking at elego.de
> Subject: [M3commit] CVS Update: cm3
> 
> CVSROOT: /usr/cvs
> Changes by: hosking at birch. 09/05/11 05:24:22
> 
> Modified files:
> cm3/m3-libs/m3core/src/runtime/POSIX/: RTPerfTool.m3 
> 
> Log message:
> Avoid warning:
> 
> ../src/runtime/POSIX/RTPerfTool.m3: In function 'RTPerfTool__StartTool':
> ../src/runtime/POSIX/RTPerfTool.m3:125: warning: variable 'M3_AcxOUs_execResult' might be clobbered by 'longjmp' or 'vfork'
> 



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://m3lists.elegosoft.com/pipermail/m3commit/attachments/20090511/f0701d5b/attachment-0002.html>