[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Help-smalltalk] ZeroDivide bug in float printing
From: |
Holger Freyther |
Subject: |
Re: [Help-smalltalk] ZeroDivide bug in float printing |
Date: |
Fri, 29 Mar 2019 16:32:28 +0000 |
> On 29. Mar 2019, at 10:15, Derek Zhou <address@hidden> wrote:
>
>
>> http://git.savannah.gnu.org/cgit/smalltalk.git/commit/libgst/interp.inl?id=72ada189aba0283c551ead16635c1983968080b8
> That's good; however there are poor poeple still stuck with old
> compilers
Is this true for mainstream OS and CPU platforms? Do you have an example?
When I added the builtin (supported by then gcc and now present day clang)
my assumption was sooner than later everyone will be able to use a modern
compiler.
Even for old CentOS there is support for newer compilers. Do you mind
sharing your specific example?
>> Overflow checking requires lots of thinking. E.g. have you considered
>> where abs(a) < 0? I would prefer if we can exclusively use the compiler
>> primitive.
> compiler primitives are nice; but there should be fallback path that
> also work
I think it depends. Undefined behavior is difficult to grok and it feels
we are walking very close to the edge.
By skimming the code I feel we have to look very carefully at it. Not
because I think it is wrong but because it is not trivial. The cognitive
load for using the builtin is a lot lower.
Let me try to review this in depth.
holger