#1202 0.10.2 fails to compile on macosx

Reporter Fabrice Ducos
Owner Zash
Created
Updated
Stars ★ (1)
Tags
  • Status-Fixed
  • Priority-Medium
  • Patch
  • Type-Defect
  • Milestone-0.10
  1. Fabrice Ducos on

    What steps will reproduce the problem? 1. pwd # :prosody-0.10.2 2. ./configure --ostype=macosx 3. make What is the expected output? Build completed successfully. What do you see instead? pposix.c:433:8: warning: implicit declaration of function 'initgroups' is invalid in C99 [-Wimplicit-function-declaration] ret = initgroups(lua_tostring(L, 1), gid); ^ pposix.c:523:10: error: use of undeclared identifier 'RLIMIT_MEMLOCK' return RLIMIT_MEMLOCK; ^ pposix.c:527:10: error: use of undeclared identifier 'RLIMIT_NPROC' return RLIMIT_NPROC; ^ pposix.c:531:10: error: use of undeclared identifier 'RLIMIT_RSS' return RLIMIT_RSS; What version of the product are you using? On what operating system? prosody 0.10.2 on MacOSX High Sierra (10.13.6) Please provide any additional information below. The same errors could be obtained with clang (llvm) 9.1.0 and gcc 8.2.0 Other reporters already noticed this problem: https://github.com/bjc/prosody/issues/16 https://github.com/prosody/homebrew-prosody/issues/11

  2. Fabrice Ducos on

    Zach's solution to https://github.com/prosody/homebrew-prosody/issues/11 fixes the errors with RLIMIT_XXX, but not the warning on initgroups. Strangely, <unistd.h> (where initgroups is declared) is properly included in pposix.c. Seems to be some conditional declaration issue.

  3. Zash on

    Thanks for the report and for testing that patch. Pushed it as https://hg.prosody.im/0.10/rev/8f9e18d4fe50 Could you check if adding #define _BSD_SOURCE affects that warning?

    Changes
    • tags Milestone-0.10 Status-Started
  4. Fabrice Ducos on

    Dear Zash, thank you for your quick reply. Defining _BSD_SOURCE has no effect. However, with the following patch in pposix.c, it works properly with clang and gcc (at least on High Sierra, not tested for older versions of OSX) #if defined(__APPLE__) #define _POSIX_C_SOURCE __DARWIN_C_FULL #else #define _POSIX_C_SOURCE 200809L #endif Cheers, Fabrice

  5. Zash on

    Is there some kind of documentation that explains what this does?

  6. Fabrice Ducos on

    I know no documentation about this, therefore I had to dive in the system headers. initgroups() is not POSIX.1 (reference: Advanced Programming in the Unix Environment, 3rd edition, p. 184). As you suggested, it should be made available with _BSD_SOURCE, but this macro doesn't seem to be tested in OSX system header files. In order to enable extensions, one needs to set up __DARWIN_C_LEVEL on Apple's systems. When one examines unistd.h (where initgroups is defined), one sees that it is defined conditionally within a snippet: /* Darwin extensions */ #if __DARWIN_C_LEVEL >= __DARWIN_C_FULL __DARWIN_C_FULL itself is defined in /usr/include/sys/cdefs.h (on OSX systems). /* * Set a single macro which will always be defined and can be used to determine * the appropriate namespace. For POSIX, these values will correspond to * _POSIX_C_SOURCE value. Currently there are two additional levels corresponding * to ANSI (_ANSI_SOURCE) and Darwin extensions (_DARWIN_C_SOURCE) */ #define __DARWIN_C_ANSI 010000L #define __DARWIN_C_FULL 900000L #if defined(_ANSI_SOURCE) #define __DARWIN_C_LEVEL __DARWIN_C_ANSI #elif defined(_POSIX_C_SOURCE) && !defined(_DARWIN_C_SOURCE) && !defined(_NONSTD_SOURCE) #define __DARWIN_C_LEVEL _POSIX_C_SOURCE #else #define __DARWIN_C_LEVEL __DARWIN_C_FULL #endif Regards

  7. Fabrice Ducos on

    A better way is not to hack the _POSIX_C_SOURCE macro (as I did in my first proposition) and to define _DARWIN_C_SOURCE instead. #if defined(__APPLE__) #define _DARWIN_C_SOURCE #endif #define _POSIX_C_SOURCE 200809L Has been tested and works.

  8. Zash on

    I committed that as https://hg.prosody.im/0.10/rev/37b796cc366a Thanks!

    Changes
    • owner Zash
    • tags Patch Status-Fixed

New comment

Not published. Used for spam prevention and optional update notifications.