From: craigberry@nospam.mac.com   
      
   On 7/3/25 3:07 PM, Dan Cross wrote:   
   > In article <1046hc5$62th$2@dont-email.me>,   
   > Arne Vajhøj   wrote:   
   >> On 7/3/2025 1:40 PM, Dan Cross wrote:   
      
   >>> You refer to `_PATH_DEVNULL` but do not `#include `, as   
   >>> required by POSIX.   
   >>   
   >> This must support non *nix systems (as an example   
   >> VMS !!) - config.h is expected to provide that with the rest.   
   >   
   > If copyright dates are anything to judge by, `` has   
   > been a thing since 1989, but I was wrong in that it is not   
   > actually in POSIX: it is a BSD extension, though common.   
   >   
   > In general, it is good practice and good hygiene in C programs   
   > to `#include` the header files that are documented to define the   
   > symbols and types that you use in your program, instead of   
   > relying on transitive includes to populate things ambiantly via   
   > e.g. `config.h`.   
   >   
   > Regardless, in that case, you shouldn't use `_PATH_DEVNULL`.   
   > Note the leading `_`: that signifies a reserved identifier,   
   > which is not something you should be defining yourself.  If you   
   > want to punt this to e.g. config.h, better would be to define a   
   > new name (say, `TDS_PATH_DEVNULL`) and use that.   
      
   The code ya'll are talking about replacing:   
      
   https://github.com/FreeTDS/freetds/blob/Branch-1_5/src/replacements/vasprintf.c   
      
   does the following:   
      
   #if HAVE_PATHS_H   
   #include    
   #endif /* HAVE_PATHS_H */   
   . . .   
   #ifndef _PATH_DEVNULL   
   #define _PATH_DEVNULL "/dev/null"   
   #endif   
      
   FreeTDS is relying on autoconf to sort out whether paths.h is available   
   and will only include it if HAVE_PATHS_H is defined in config.h, and   
   thus only defines _PATH_DEVNULL if it's known not to exist. Obviously   
   one should avoid making up one's own reserved identifiers and clobbering   
   documented and well-known names, but that's not really the same thing as   
   providing an implementation for one that's known to be missing. This   
   fallback approach could of course be added to Arne's implementation if   
   desired; to me it seems a lot cleaner than defining yet another macro,   
   which would need to be defined in terms of _PATH_DEVNULL when it exists   
   but otherwise not.   
      
   The problem for VMS is that the workaround doesn't work.  Or at least it   
   didn't when I was first porting FreeTDS 20+ years ago because the CRTL   
   at the time did not recognize '/dev/null' as a valid path.  I believe   
   recentish CRTLs (maybe VMS 8.x and later?) do have special case code   
   that translates it to the native null device.  But that wasn't   
   available, so I added the following line to the template from which   
   config.h is generated on VMS:   
      
   #define _PATH_DEVNULL "_NLA0:"   
      
   The main advantage of this approach is that it worked, and it did so   
   with a one-line change to a file I was already maintaining. I didn't   
   have to sprinkle '#ifdef __VMS' in files that already existed and were   
   already working on other platforms. I didn't have to create a lot of new   
   files, the presence of which in the repository I couldn't test without   
   autoconf/automake, which I didn't have access to at the time.  All of   
   which meant upstream maintainers were far more likely to accept my changes.   
      
   I'm sure I'll get a lecture about everything I did wrong, but at least   
   now the context is out there for why it was reasonable for Arne to   
   depend on prior art, i.e., the presence of _PATH_DEVNULL in config.h.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)