On 11/15/2014 12:20 PM, Brandon Allbery wrote:
On Sat, Nov 15, 2014 at 12:17 PM, René J.V. rjvber...@gmail.com
mailto:rjvber...@gmail.com wrote:
Have you asked your users to start the executable in the debugger
to see if the SIGTERM results from an abort function being called
or
On Nov 16, 2014, at 9:56 AM, Qianqian Fang wrote:
any more suggestions?
Seems like the solution is: don't use statically linked binaries. Link
dynamically, like Apple recommends. If you build for the right SDK and
deployment target, it should be compatible with older OS X versions. At worst,
hi everyone
I have compiled a set of binaries on previous versions of Mac (mostly
OSX 10.6) with macports gcc/g++ using static linking, and thought that
these binaries will be supported in the future versions of MacOS.
However, several users of my software package reported failure on
On Sat, Nov 15, 2014 at 12:17 PM, René J.V. rjvber...@gmail.com wrote:
Have you asked your users to start the executable in the debugger to see
if the SIGTERM results from an abort function being called or something
else?
9 is SIGKILL, not SIGTERM.
I am under the impression that Apple
Apple has never guaranteed support for fully static binaries, rather
it has explicitly discouraged them.
The reason is that Apple wants to be free to change the binary
interface for system calls. My understanding is that they are, in
general, implemented by sending Mach messages from a dylib
On 11/15/2014 12:17 PM, René J.V. Bertin wrote:
Have you asked your users to start the executable in the debugger to
see if the SIGTERM results from an abort function being called or
something else?
I asked my users to try valgrind, but one of them mentioned
valgrind has not yet been
valgrind has supported OS X for a few years now.
If you can't find a prepackaged valgrind build I'd be happy to build
it for you then send you the binary.
If you don't want to actually install Yosemite on your own Mac, there
is a documented procedure for imaging a bootable drive from the OS X
On Sat, Nov 15, 2014 at 12:53 PM, Qianqian Fang fan...@gmail.com wrote:
On 11/15/2014 12:17 PM, René J.V. Bertin wrote:
Have you asked your users to start the executable in the debugger to see
if the SIGTERM results from an abort function being called or something
else?
I asked my users
On Sat, Nov 15, 2014 at 11:51 AM, Qianqian Fang fan...@gmail.com wrote:
I have compiled a set of binaries on previous versions of Mac (mostly OSX
10.6) with macports gcc/g++ using static linking, and thought that these
binaries will be supported in the future versions of MacOS.
I should note
On Sat, 15 Nov 2014, Michael Crawford wrote:
If you can't find a prepackaged valgrind build I'd be happy to build it
for you then send you the binary.
--- Fetching distfiles for valgrind
Error: valgrind @3.10.0 is only compatible with Mac OS X 10.5, 10.6, 10.7, 10.8
and 10.9
Error:
http://stackoverflow.com/questions/26564125/yosemite-and-valgrind
From the above, Valgrind is notorious for taking their sweet time supporting
new versions of OSX. You're just going to have to wait.
On Nov 15, 2014, at 7:11 PM, Dave Horsfall wrote:
Hmmm... The port for 10.10 not yet
valgrind does two main things: it emulates the Instruction Set
Architecture of the microprocessor (x86_64 in Yosemite's case), and it
also wraps all - or at least most - of the system and library calls
with its own.
That is, if you call, say, fopen(), when you run your code under
valgrind, you'll
OS X has vast numbers of APIs. Some of these, while strictly speaking
they may be well-documented, it can be quite difficult to actually
_find_ the documentation.
...
Back in the day I figured out how to talk directly to the audio
driver, without going through QuickTime. It actually
13 matches
Mail list logo
