[Spark2014-discuss] Plans for I/O

Arnaud Charlet charlet at adacore.com
Thu Jan 24 16:47:55 CET 2013

> > We will also need to come up with a way of telling the flow analyser how
> > to represent the Text_IO functions. This could be by adding global and
> > derives annotations to the library functions or building the knowledge
> > into the flow analyser.
> The first option would be preferrable.

Agreed, much more flexible and preferable maintenance wise to be able to add
these in the sources directly rather than build into the tool itself.


More information about the Spark2014-discuss mailing list