This is the mail archive of the gsl-discuss@sourceware.org mailing list for the GSL 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: containers tentative design summary


On Tue, 2009-10-27 at 19:28 -0400, James Bergstra wrote:
> Sorry to bug you, but could you point me to a link that would explain
> why you put both of those structures into the union?

[ I assume you meant to send this to the list; so I have replied there ]


Here is an interesting (and very long) reference:

http://www.coding-guidelines.com/cbook/cbook1_2.pdf

The relevant section is 6.5.2.3, the discussion around sentences
1037, 1038 of the standard. Sentence 1037 introduces the union
construct. Sentence 1038 is a definition for
"common initial sequence", which is used in 1037.

Of course, we all believe it works right without the union.
But the presence of the union seems to guarantee that it
works right (by the new standard; the old standard is
apparently mute). I may be misreading this, and the
union may not actually be necessary, but there is
some confusion about this, if you search the web,
including a couple of formal "defect reports" for
the standard, regarding these sentences.

--
G. Jungman



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