Re: Newbie question: What is the importance for a compiler to be able to compile itself?

>> You can turn this around, and make it an argument against
>> self-residency in the compiler. There is a danger that you come to
>> rely on non-standard extensions to the language, and worse : implicit
>> non-standard extensions to the semantic definition of the language.
>>
> And why wouldn't you get reliant on similar feature with an arbitrary
> different compiler used? (like gccisms) Only actively compiling with
> multiple compilers avoids that situation, so self-residency or not is
> irrelevant in this case.


You quoted me selectively. I actually said these problems are avoided if
one regularly uses different compilers. Self-residency is highly relevant
in this case, because you only ever use one compiler.


> And then it is of course a different question all together if this is
> that bad at all. Do you really want your compiler and runtime (which
> is a bigger problem than the compiler, since you are more likely to
> use extensions there) to be written and limited by the lowest common
> denomitor?
>
> No language extensions, no platform extensions, no ability to use
> inline assembler features etc.


I prefer to think of it as sticking to the lanaguage standard. And
whilst I have had need to code routines in assembler, I have never had
need to inline assembler in a high level language. If you work on
multiple Unix platforms this is not viable anyway.

Return to the comp.compilers page.
Search the comp.compilers archives again.