Re: [dpdk-dev] Build error with mingw in debug mode

2020-11-27 Thread Nick Connolly
I've created a patch for this and also for a couple of build warnings that showed up doing a similar test with clang. On 27/11/2020 10:07, Nick Connolly wrote: Errata: the definition is in rte_os.h rather than rte_windows.h On 27/11/2020 10:05, Nick Connolly wrote: The root cause here is that

Re: [dpdk-dev] Build error with mingw in debug mode

2020-11-27 Thread Nick Connolly
Errata: the definition is in rte_os.h rather than rte_windows.h On 27/11/2020 10:05, Nick Connolly wrote: The root cause here is that rte_windows.h shouldn't be injecting 'missing' POSIX functionality into the environment.  The header is included indirectly in application code via rte_os.h, but

Re: [dpdk-dev] Build error with mingw in debug mode

2020-11-27 Thread Nick Connolly
The root cause here is that rte_windows.h shouldn't be injecting 'missing' POSIX functionality into the environment.  The header is included indirectly in application code via rte_os.h, but deciding how to implement the 'missing' functionality should be the applications responsibility. There a

[dpdk-dev] Build error with mingw in debug mode

2020-11-27 Thread David Marchand
This is introduced by e8428a9d89f1 ("eal/windows: add some basic functions and macros"). So sending to you guys: Caught while compiling all my envs with --buildtype=debug rather than debugoptimized: ninja: Entering directory `/home/dmarchan/builds/build-x86-mingw' ... Build targets in project: 70