From: tr.17687@z991.linuxsc.com   
      
   James Kuyper  writes:   
      
   > On 2025-11-01 17:39, Tristan Wibberley wrote:   
   >   
   >> Follow-up set comp.lang.c   
   >>   
   >> On 01/11/2025 02:45, Richard Damon wrote:   
   >>   
   >>> .. [stuff I think is about the invalidity of not completing   
   >>> execution] isn't a valid definition ...   
   >>   
   >> Let's suppose C disallows abort due to external events... Then no   
   >> conforming C implementation exists.   
   >>   
   >> I think we must suppose that susceptibility to power-cuts do not   
   >> render C implementations non-conforming.   
   >   
   > Correct.  What a power cut does is make a conforming implementation   
   > of C fail to continue qualifying as a implementation.   
   >   
   > In 1.6 of C89, an implementation is defined as "a particular set of   
   > software, running in a particular translation environment under   
   > particular control options, that performs translation of programs   
   > for, and supports execution of functions in, a particular execution   
   > environment."   
   >   
   > Thus, when the power is cut, the software ceases to run, so it   
   > ceases to be an implementation, and in particular, ceases to be a   
   > conforming implementation.   
      
   The description above is wrong.  The C standard clearly shows   
   that (a) program execution is carried about by a data-processing   
   system, about which the C standard does not impose any conditions   
   or requirements, and (b) that implementations are separate and   
   distinct from the data-processing system(s) in which programs   
   run.  Implementations /support/ execution of functions (and   
   programs) in an execution environment, but implementations do not   
   /carry out/ the execution of functions and programs in such an   
   environment.  Execution is carried out by the data-processing   
   system used to run programs.   
      
   The points above may be found in section 1, paragraph 2.   
      
   > There's important wording that wasn't added to the standard until   
   > C99 that clarified something that was already considered to be true:   
   > "A program that is correct in all other aspects, operating on   
   > correct data, containing unspecified behavior shall be a correct   
   > program and act in accordance with 5.1.2.3."  (4p3).  The peculiar   
   > wording is intended to make clear that unspecified behavior does not   
   > relieve an implementation of this requirement.   
   >   
   > Section 5.1.2.3 indirectly references most of the rest of the   
   > standard, while defining that only the observable behavior of a   
   > program has to match what the standard says - that behavior doesn't   
   > have to be generated in the fashion that the standard describes.   
   > Section 5.1.2.3 in C99 corresponds to 2.1.2.3 in C89.   
   >   
   > The rest of the standard defines, among other things, how a program   
   > exits.  If the behavior of the program is not undefined, it can only   
   > exit by reaching the end of the main() function, or by calling   
   > certain library functions.  If there is any observable behavior that   
   > allows you determine that it ended by any other method, the   
   > implementation is non-conforming.   
      
   The comments above reflect a misunderstanding of what is meant by   
   the word "behavior".  In the C standard, "behavior" refers to an   
   abstract description of program semantics, not what actually   
   occurs when a program is run by a data-processing system.  This   
   distinction is made evident in section 5.1.2.3 paragraph 1, which   
   says "The semantic descriptions in this International Standard   
   describe the behavior of an abstract machine in which issues of   
   optimization are irrelevant."  If when we run a program it runs   
   out of memory and crashes, that doesn't suddenly make the compiler   
   non-conforming;  it could mean just that some other programs are   
   running and at the moment there is not enough memory available.   
   The C standard specifically does not concern itself with such   
   circumstances;  see section 1, paragraph 2, the last two marked   
   sub-paragraphs.   
      
   The idea that what happens during program execution can change   
   whether an implementation is conforming would mean that there are   
   /no/ conforming implementations, because it's possible to write a   
   strictly conforming program that will exceed the capacity of any   
   machine in existence (and indeed, that will exceed the capacity   
   of any machine that ever will be in existence).  Surely the C   
   standard does not mean to rule out the possibility of any   
   conforming implementations at all.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)