From: mutazilah@gmail.com   
      
   On 21/02/24 10:51, Lew Pitcher wrote:   
   > On Wed, 21 Feb 2024 05:37:45 +0800, Paul Edwards wrote:   
      
   >>> Choice (b) is ... useless. It recognizes a non-issue, and does nothing.   
   >>> It would be a header change that reserves a flag that would be ignored   
   >>> by everyone.   
   >>   
   >> Almost everyone.   
   >>   
   >> PDOS/386 would use it.   
   >   
   > If PDOS/386 needs it, then PDOS/386 probably should use it.   
      
   Not so much "needs" as "would be desirable".   
      
   > However, the needs of PDOS/386 aren't the needs of Linux or POSIX, and   
   > neither Linux in particular, nor POSIX in general, need bother with   
   > it.   
      
   Sure. As I said in a previous message - "different goals".   
      
   > If you are concerned with PDOS/386 being able to use source code written   
   > to the POSIX standard,   
      
   Not just source - that is easy to deal with.   
   The executable.   
      
   > then I suggest that you select the value and   
   > treatment of your O_TEXT flag to work in absence, as POSIX code won't   
   > have it or use it.   
      
   That's exactly what I'm doing. I'm just asking   
   which bit would be good to use. I showed which   
   bit that Cygwin was using, and I made a suggestion   
   that I could work downwards on the assumption   
   that Linux is working upwards.   
      
   > The alternative is to specify that such a flag /is/ required for PDOS/386,   
      
   That's the alternative I am after.   
      
   > which restricts you to using code written specifically for PDOS/386, and   
      
   It is only code that uses the new flag that will   
   work nicely on PDOS/386, yes.   
      
   The other code would still run, it just won't   
   automatically put the CR before the NL, so it   
   won't look good on (this) Windows-like environment.   
      
   > not the more general-purpose   
      
   I consider adding O_TEXT to be more general-purpose   
   than one that assumes the code will only be used on   
   a Unix-like environment, never a Windows-like   
   environment.   
      
   > and widely available POSIX standard.   
      
   It still follows the rest of that, so it will still   
   work on any POSIX environment.   
      
   The *application* code will need to have:   
      
   #ifndef O_TEXT   
   /* uh oh, we're on a standard POSIX environment,   
   missing the PDOS/386-inspired Linux enhancement   
   that hasn't YET (as of 2024-02-21) made it to a   
   formal POSIX standard */   
   #define O_TEXT 0   
   #endif   
      
   ... open(... | O_TEXT ...)   
      
   BFN. Paul.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)