Re: Decimal: Formal review begins.

From: Christopher Kormanyos (e_float_at_[hidden])
Date: 2025-01-20 18:33:59

>> one potential recipe on how>> to deal with this in generic>> code above.
> I'm not sure I understand this.> How can you have "generic code"> with different suffixes where there> are compilation failures that> would not be there in non-generic> code?
> From "one data table for a type T"> I imagine something like
> `T table[] = {1.25_df, 3.2_df, 51_df}`
> So while your table type is generic> the constants used are not.
You are right and my rhetoricwas flawed.
In fact, that is exactly what I meant.My description got all confusedand was not quite right. You mightcast each table menber to T viastatic_cast, but that is the generalidea.
Now one thing I am not confused aboutis that I do not want compilationerrors nor warnings if the clientputs not-enough/too-many digitsin the string literals. That wasthe main issue I was attemptingto debate.
Thank you and good catch.
- Chris

    On Monday, January 20, 2025 at 11:17:15 AM GMT+1, Alexander Grund via Boost <boost_at_[hidden]> wrote:  
 
 
Am 18.01.25 um 12:48 schrieb Christopher Kormanyos via Boost:
> > Could you illustrate such a use case?
> > Concretely, what do you mean by
> > several widths of types? Do you
> > mean decimal32 vs decimal64?
> Yes. Let's say you made a program
> that uses one data table for a type T that could be decimal32, 64 or 128.
> But you only wanted one table.
> You would make the table entries with the highest width suffix and decoratee ach table entry with a static-cast to T in order to get all the table entries into type T.
> To be fair, everyone already has todo something like this with normal float, double and long double.
>> With the syntax you have today,
>> you already have distinct suffixes
>> for each type (i.e. _df always constructs
>> decimal32, _dd always constructs decimal64,
>> and so on).
> Yes. And I gave one potential recipe on how to deal with this in generic code above. My judgement may have been hurried and might seem unfair.
> But my intuition tells me, a whole bunch of clients would run into compilation failures. Until they come up with their own recipes for generic literal values.
> And the result would be more frustration than had we simply done nothing.
I'm not sure I understand this.
How can you have "generic code" with different suffixes where there are 
compilation failures that would not be there in non-generic code?
 From "one data table for a type T" I imagine something like

`T table[] = {1.25_df, 3.2_df, 51_df}`

So while your table type is generic the constants used are not.
If the above errors due to using `1.25352397826723672375_df` you ought 
to use `_dd` to be able to use it in `decimal64 table[]` anyway.
I don't see a scenario where it would be ok to truncate a constant and 
then store them in a type that would have been able to store the 
nun-truncated value.
The other way round is more reasonable to use rounding/truncation when 
really required by the target type.

Can you explain which case you had in mind?

_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost