patch to add POSIX fsync and fdatasync to fs module
billywhizz
fsync and fdatasync system calls to the fs module. patch is here:
http://gist.github.com/339815.
Ryan Dahl
Can you add a test which at least calls these functions? Also test
with node_g (./configure --debug && make)
billywhizz
on my setup which is Fedora 12 (2.6.32.9) on x86_64.
patch is here:
http://gist.github.com/339869
the test script is here (couldn't figure out how to get this included
in the patch):
http://gist.github.com/339870
this goes in the test/simple/ directory in the node source tree. i
used file name test-fs-fsync.js.
if you need anything else, let me know.
On Mar 22, 6:12 am, Ryan Dahl <coldredle...@gmail.com> wrote:
Ryan Dahl
> have created a test for this and have tested both debug and non debug
> on my setup which is Fedora 12 (2.6.32.9) on x86_64.
The patch looks fine, but it appears as if deps/libeio/wscript is not
properly checking for fdatasync. It doesn't exist on OSX yet the
configure script still thinks it does. This is causing problems.
billywhizz
the check into wscript i imagine...
On Mar 23, 3:56 pm, Ryan Dahl <coldredle...@gmail.com> wrote:
Ryan Dahl
> do you want me to have a look at this? should be pretty easy to put
> the check into wscript i imagine...
Would be nice. Otherwise I can have a look at it, but it will take a week or so.
billywhizz
conf.check_cc(msg="Checking for fdatasync(2)",
define_name="HAVE_FDATASYNC", fragment="""
#include <unistd.h>
int main(void)
{
int fd = 0;
fdatasync (fd);
return 0;
}
""")
do you know of any documentation explaining how these tests are
supposed to work? from looking at the others i presume the build
engine is compiling/running that snippet and then catching errors if
the function doesn't exist. is OSX not throwing errors during the
build for this snippet? afraid i don't have a mac to test on...
On Mar 25, 12:52 am, Ryan Dahl <coldredle...@gmail.com> wrote:
billywhizz
sure why this isn't caught during the build on those platforms...
On Mar 25, 12:52 am, Ryan Dahl <coldredle...@gmail.com> wrote:
billywhizz
the libeio wscript should fix the mac OSX issue.basically i changed
the script to always return a failure code for fdatasync on OSX so
that libeio will alias fdatasync to fsync and everything should build
and run. unfortunately, this is only a theory, as i don't have a mac
to test on. if anyone has a mac and can test this, i would be
grateful. i can fork the latest node on github and apply my patches if
you want to pull from there and do a build...
conf.check_cc(msg="Checking for fdatasync(2)",
define_name="HAVE_FDATASYNC", fragment="""
#include <unistd.h>
int main(void)
{
int fd = 0;
fdatasync (fd);
return -1;
#else
return 0;
#endif
}
""")
On Mar 31, 12:35 pm, billywhizz <apjohn...@gmail.com> wrote:
> also, it looks like OSX/FreeBSD do not supportfdatasyncso i'm not
>
> On Mar 25, 12:52 am, Ryan Dahl <coldredle...@gmail.com> wrote:
>
>
>
> > On Wed, Mar 24, 2010 at 3:58 PM, billywhizz <apjohn...@gmail.com> wrote:
> > > do you want me to have a look at this? should be pretty easy to put
> > > the check into wscript i imagine...
>
> > Would be nice. Otherwise I can have a look at it, but it will take a week or so.
You received this message because you are subscribed to the Google Groups "nodejs" group.
To post to this group, send email to nod...@googlegroups.com.
To unsubscribe from this group, send email to nodejs+un...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/nodejs?hl=en.
billywhizz
applications - it just catches errors so i call a function that
doesn't exist to make sure the app craps out on OSX:
conf.check_cc(msg="Checking for fdatasync(2)",
define_name="HAVE_FDATASYNC", fragment="""
#include <unistd.h>
int main(void)
{
int fd = 0;
//make the application crash
sdskdfjhskjhsdf();
#endif
return 0;
}
""")
