signed/unsigned integer conversion for right shift seems

Peter T. Breuer ptb@inv.it.uc3m.es
Wed Feb 7 18:50:00 GMT 2018


"Also sprach Tadeus Prastowo:"
> Good.  Now since you claimed that you were a mathematical logician in
> one of the previous e-mails, I will let you know as a newcomer in this
> mailing list that the set of axioms in this mailing list that you keep
> posting to includes the proposition that said that "a discussion on
> the C language issue is disallowed", another proposition that said

I am not discussing the C language, let alone any "issue" you may or
may not have with it. 

You see no opinion on C from me at all.  I have not expressed whether or
not I think the gcc computed goto is a good idea, for example (I don't
like it but have never used it, and it surely has its place - it's
semantically horrible, however, and gcc statements as expressions
is also bad).

It was quite a specific question: is gcc right in this instance.

If you thought it was off topic for gcc-help, I would have been glad to
ask elsewhere, at your suggestion.

For the rest, I merely reply to what is emailed to me.

My best guess from the weird replies is indeed that I've overtaxed the
natives, unwittingly, and I apologise for that. 

I recall now, belatedly, that the "help" newsgroups for an area are
generally for dispensing somewhat platitudinous advice to newbies.  

That explains a lot.

> that "if a discussion is about a C language issue, then the discussion
> is off-topic", and another proposition that said that "an off-topic

What "C language issue" do you claim this is about? We want to know
if gcc is right, since on the face of it it seemed to contradict the
standard's pithy words on conversions.

The correct response was "no it doesn't because 6.5 says that
conversions are not applied for >> and 6.3.18 does not list
 >> among what things conversion is applied to and 6.1 says that nothing
else than 6.5 defaulting to 6.3.18 is allowed".

Fin.

Instead, a hail of rubbish ensued, to which I conscienciously replied,
nicely pointing out what was wrong with it, because I am a nice and
kind person, and when somebody writes nonsense, I am nice and kind
enough to correct them, less they continue to do so and hurt
themselves.

I should have been clever enough to realise I was talking to the
janitor.

> discussion shall not be discussed any longer."

Nobody is discussing anything. The correct statement is as above:

  "signed >> unsigned"

  in gcc should not convert to

  "unsigned >> unsigned"

  even though that IS an "automatic" conversion in the standard because
  $6.5 says that conversions are not applied for >> and $6.3.18 does not
  list >> among things conversion is applied to and $6.1 says that
  nothing else than $6.5 defaulting to $6.3.18 SHALL BE allowed

QED.

It should not have been hard for you.

> So, can't you stop now, please, you who claimed to be
> http://dblp.uni-trier.de/pers/hd/b/Breuer:Peter_T= ?

I did not claim it, and I insist you spell correctly, if at all.

PTB



More information about the Gcc-help mailing list