tly call cilly.asm.exe. As far as I remember, "cilly" prints the
command line that executable's invoked with. This is a "raw"
cillyfier, which doesn't perform preprocessing and/or compilation.
--
Pavel Shved
website:
ructures should be suppressed when using --no-convert-field-
offsets, like it is in the patch.
However, I do not really understand in what cases splitting structures
happens. Is the situation like the one above feasible?
--
Pavel Shved
website: http://coldattic.info/
e-mail: pavel.sh...@gmail.com&
the conversion bug is still there!
So, it's a generic issue with --dosimplify, rather than a bug introduced
by the modification.
--
Pavel Shved
website: http://coldattic.info/
e-mail: pavel.sh...@gmail.com
--
rt-field-
offsets". Retaining behavior without the additional options is the
primary goal here.
--
Pavel Shved
website: http://coldattic.info/
e-mail: pavel.sh...@gmail.com
--
Get your Android app more play: Bri
elds.
> > Besides, the behavior without this command- line option is
> > retained.
>
> I do not like the idea of adding bugs on purpose, but I see your
> point. Need to think a bit more about it.
I do not approach this as "adding a bug on purpose". Rather, I think
or without this command-
line option is retained.
--
Pavel Shved
website: http://coldattic.info/
e-mail: pavel.sh...@gmail.com>From 6c3e51e7ae4882beae61e9600b6803414d9c9e23 Mon Sep 17 00:00:00 2001
From: Pavel Shved
Date: Fri, 30 Sep 2011 17:00:46 +0400
Subject: [PATCH] Add --no-convert-field-o
ons. We wanted to preserve direct calls
to these functions in order to run the generated code through different
static verification tools.)
--
Pavel Shved
ISPRAS
(Institute for System Programming
of Russian Academy of Sciences)
Operating Systems section
email: sh...@i
Here's a sample file to test the fix.
However, I'm not sure how and if it should be inserted into the CIL test
suite.
--
Pavel Shved
ISPRAS
(Institute for System Programming
of Russian Academy of Sciences)
Operating Systems section
email: sh...@ispras.ru
void good()
{
This option prevents converting direct function calls into the
calls-by-pointers if the address of the function has been taken
somewhere. Calls from inside of structures are converted anyway.
(A small debug print fix has been also applied)
---
src/ext/simplify.ml | 14 +++---
1 files c
The resultant main function looks like this:
int main(void)
{ void (*__cil_tmp1)(void) ;
{
{
good();
__cil_tmp1 = & bad;
(*__cil_tmp1)();
}
return (0);
}
}
--
Pavel Shved
ISPRAS
(Institute for System Programming
of Russian Academy of Sciences)
Operating Systems section
email:
ers who might have missed
> it:
>
> git clone --recursive git://github.com/kerneis/cil.git
No, we've actually set up our own Git mirror, and the commit IDs differ.
Would be nice, again, to have a designated upstream, so that everybody
could branch from it in a consistent
ody of function main falls-through. Adding a return
statement
Error: Cabs2cil had some errors
Fatal error: exception Errormsg.Error
After the patch, the error is gone, and the generated code contains the
original assignment.
--
Pavel Shved
ISPRAS
(Instutute for System Programming
of Russian Acad
Do not try to approach them as "field-to-union" casts
---
src/frontc/cabs2cil.ml | 19 ++-
1 files changed, 14 insertions(+), 5 deletions(-)
diff --git a/src/frontc/cabs2cil.ml b/src/frontc/cabs2cil.ml
index 6c3d692..7833095 100644
--- a/src/frontc/cabs2cil.ml
+++ b/src/frontc/c
ver? Or is our host banned (we had a nightly
mirroring script that updated from the repository each night)?
--
Pavel Shved
ISPRAS
(Instutute for System Programming
of Russian Academy of Sciences)
Operating Systems section
email:
14 matches
Mail list logo
