[m-rev.] for review: standalone interfaces

[Jonathan Morgan]

On 2/5/07, Julien Fischer <juliensf at csse.unimelb.edu.au> wrote:
> Also, while working on this one of the problems I ran into is that the compiler
> sets the standard output stream to be different things at different points.
> This is quite confusing and extremely annoying to debug if it goes wrong.
> I would like to (a) change the compiler so that all streams are explicit
> arguments and (b) change the coding standards to say that implicit streams
> shouldn't be used.  (It's quite large change but it should be fairly trivial.)
> Objections?

Logically, this is the wrong thing to do.  However, not having to pass
streams as arguments is quite convenient in a lop-sided sort of way.
I strongly support this suggested change, so long as the coding
standard is updated as suggested.  However, I find this a problem in
many bits of Mercury code, not just the compiler.  I observed that, as
most code that I saw used see() and tell(), that this was considered
the correct way to do file I/O in Mercury.  Though it annoyed me
exceedingly I have been using it in my own code.  This is why I
suggest that the coding standard is updated.

Jon
--------------------------------------------------------------------------
mercury-reviews mailing list
Post messages to:       mercury-reviews at csse.unimelb.edu.au
Administrative Queries: owner-mercury-reviews at csse.unimelb.edu.au
Subscriptions:          mercury-reviews-request at csse.unimelb.edu.au
--------------------------------------------------------------------------