I've had a couple of messages to the list blocked, apparently
sourceforge does not like comcast....
It seems to be working tonight so I'll try this again.
--------------------------------------------------------------------------------
Yes, with just %d some 64 bit compilers will warn too.
codec2-dev/unittest/ofdm_mem.c:46:12: warning: format ‘%d’ expects argument of
type ‘int’,
but argument 2 has type ‘long unsigned int’ [-Wformat=]
printf("struct OFDM.................: %d\n", sizeof(struct OFDM));
I have updated ofdm_mem.c to use %zd.
Do you have warnings from fdmdv_mem.c too?
--------
I put the -no-pie flag in for the new ofdm_stack program. David has a
compiler that produces "shared object" files by default even for a plain
executable program. Apparently that is the trend now as they can be run
as well as dynamically linked.
This program is used with a utility which converts function addresses to
their name (addr2line). Shared object files get put into mapped
(virtual) memory and the addresses don't match anymore.
I don't know cmake very well so I may not have found the best way
to deal with this.
--------
Regards
Don (new guy)
On Thu, Jun 21, 2018 at 04:10:02AM +0930, David Rowe wrote:
> Don (new contributor who is working on the OFDM port to the stm32) added
> this option to get unittest/ofdm_stack to run. This program is a clever
> unit test to measure stack usage. So -no-pie is only really needed on
> x86 for this single unittest.
>
> Don can you pls comment on this linker option?
>
> Thanks
>
> David
>
> On 20/06/18 22:09, Richard Shaw wrote:
> > On Tue, Jun 19, 2018 at 7:51 PM Alan Beard <bear...@unixservice.com.au
> > <mailto:bear...@unixservice.com.au>> wrote:
> >
> > Hi all,
> >
> > unittest/CMakeLists.txt on lines 91 and 92 has the added option
> > -no-pie
> >
> > Which is not recognised by the linker here, Fedora 23 on armhf
> > GNU ld version 2.25-15.fc23
> >
> >
> > If David can chime and say whether the -no-pie option is required that
> > would be the quickest fix. I know upgrades aren't always fun but the
> > oldest version of Fedora not EOL'd is 27. Can your hardware run on the
> > armv7hl version? I confess I don't understand the differences in arm
> > hardware and targets so I'm not sure what the differences are in armhf
> > vs armv7hl....
> >
> >
> > Also, on a 32 bit machine the function "sizeof(thing)" returns
> > an int so a printf statement should use %d not %ld
> >
> > unittest/ofdm_mem.c
> >
> >
> > I tried Jeroen's suggestion to change to %zd and the output of ofdm_mem
> > looked fine on my Fedora 28 x86_64 machine. Can you try the same change
> > and see if it address your problem?
> >
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Freetel-codec2 mailing list
Freetel-codec2@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/freetel-codec2
