Re: [lmi] Stylistic question about constructing error messages

[Greg Chicares]

On 02/12/2016 02:01 PM, Vadim Zeitlin wrote:
> 
>  There are many checks for the input data in the SOA tables code that I'm
> writing and each of them throws an exception if it fails. Most of them also
> try to provide as much information as possible about what exactly went
> wrong because nothing is more aggravating than running a command merging
> in a hundred of files and getting "invalid data" error without any more
> details. So a typical check looks like this:
> 
>     if(num > max_num)
>         {
>         std::ostringstream oss;
>         oss << "value for numeric field '"
>             << name
>             << "' is out of range (maximum allowed is "
>             << max_num
>             << ") at line "
>             << line_num
>             ;
>         throw std::runtime_error(oss.str());
>         }

Ultimately I anticipate integrating this into lmi, where I'd prefer to use
lmi's fatal_error() instead...

    fatal_error()
        << "Value for numeric field '"
        << name
        << "' is out of range (maximum allowed is "
        << max_num
        << ")."
        << LMI_FLUSH // A macro, yes, but a minimal, unavoidable one.
        ;

>  So far so good, but there are _many_ of them. I was writing all the checks
> in full so far because I wanted to keep things simple, but this is getting
> really tiresome, so I'd like to ask you what do you think about some
> alternatives, i.e. to choose from

Of the ideas presented, I like the macro-based ones least, and I think
you do, too. Your favorite was this:

>         throw make_error<std::runtime_error>
>             ("value for numeric field '"
>             ,name
>             ,"' is out of range (maximum allowed is "
>             ,max_num
>             ,") at line "
>             ,line_num
>             );

But isn't that almost identical to to the fatal_error() version above?