This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Ada.Characters.{Wide_}Latin_9 should be deleted

> I agree.  At least Wide_Latin_9 is a serious mistake and has to be
> removed.  I'll submit a patch.
> I'm not so sure about Latin_9; maybe a distinct character type can be
> used?

I find it absurd to remove this functionality, when there are serious Ada
users who requested it and use it. These packages are there because Ada
users need them. What possible gain is there in creating incompatibilities
between ACT's version of GNAT and the FSF version by removing useful features
based on theoretical concerns that have very little to do with practical
Ada use.

If someone wants to add a 32-bit wide_wide_character type, that's a reasonable
though probably in practice not very useful addition. We would probably adopt
such a change if it was clean, since we have liked to have very flexible
character support (there are certainly no other Ada compilers that support
Latin-2 identifiers :-)

I would definitely be strongly opposed to such a patch, which seems to have
no value whatsoever, and will be damaging to the GCC project. I hate to see
discussion on CLA about how the FSF version does not include all the
functoinality of GNAT because, over ACT's objections, useful stuff has been
removed. Note that "useful" here is not a value judgment, if users say
something is useful, then it is useful. I agree that sometimes users think
something is useful and it is too horrible to put in anyway. But the
addition of Latin-9 is entirely consistent with the kind of addition
envisioned in the RM, and is harmless from a language coherency point of view.

Indeed, the bizarre viewpoint here seems to be something like "you should not
be using Latin-9, even though it is an approved standard, instead we think
you should use 32-bit characters". This kind of stylistic judgment (which
is not even universally agreed on, I think it is plain wrong) has no place 
in language design and implementation.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]