This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: document 3.4 changes
- From: Gerald Pfeifer <gerald at pfeifer dot com>
- To: Nathan Sidwell <nathan at codesourcery dot com>
- Cc: gcc-patches at gcc dot gnu dot org,Mark Mitchell <mitchell at codesourcery dot com>
- Date: Thu, 14 Aug 2003 02:17:01 +0200 (CEST)
- Subject: Re: document 3.4 changes
- References: <3F3A09D8.5040108@codesourcery.com>
On Wed, 13 Aug 2003, Nathan Sidwell wrote:
> here is a patch to document some 3.4 changes. I've folded in
> those changes described in cp/NEWS, as well as augmenting the gcov
> information.
This looks good, in principle, thanks! Just a few notes below, plus
please also consider Gaby's comments.
> It occurs to me that we should remove cp/NEWS - no point duplicating
> things.
I think that indeed it's much better to have a single place where changes
are documented, and that cannot really be cp/NEWS (not the least because
HTML or texinfo can be converted to plain text, but not the other way
around).
May I suggest that for now you only _migrate_ changes from cp/NEWS to
gcc-3.4/changes.html (that is, remove everything that your patch adds
to gcc-3.4/changes.html from cp/NEWS) and add a reference from cp/NEWS
to the web page?
> ok?
Yes, modulo the minor comments below, and thanks for tackling this!
+ <li>Coverage file format has been redesigned.
+ <li>The <code>gcov</code> coverage tool has been improved.
Please add </li> at the end of all list items. Do we miss a "The" at
the beginning of the first entry above? For program names such as gcov,
I would omit the <code>...</code> markup.
+ <li>You must now use the `typename' and `template' keywords to
For code (fragments) like the two above, I suggest to use <code>...</code>.
IIRC, Joseph had strong arguments against TeX quotes in web pages, so I
suggest to simply omit the quotes or use double quotes or single quotes
instead.
+ You must explicitly qualify the name. Here is the corrected
+ definition of <code>C<T>::g</code>,
<p>...</p>?
+ <li>When -pedantic is used, G++ now issues errors about spurious
+ semicolons. For example,
+ <pre>
+ namespace N {}; // Invalid semicolon.
+ void f() {}; // Invalid semicolon.</pre>
Should G++ really issue an error, and not a warning? (Cf. the current
discussion on the gcc list.)
! <li>You must use "template <>" to introduce template
<code>template...</code>
! This behavior is not mandated by the standard. Now G++ issues
! an error about this code. To avoid the error, you must move
! the declaration of `g' before the declaration of `f'. The
! default arguments for `g' must be visible at the point where
! it is called.
Please use <code>g</code> and <code>f</code>.
Gerald