• "Walter" <walter digitalmars.com> May 20 2003
• gf <mz_y2k yahoo...com> May 21 2003
• "Walter" <walter digitalmars.com> May 21 2003
• John Fletcher <J.P.Fletcher aston.ac.uk> May 22 2003
• John Fletcher <J.P.Fletcher aston.ac.uk> May 22 2003
• Richard Grant <fractal clark.net> May 22 2003
• John Fletcher <J.P.Fletcher aston.ac.uk> May 28 2003
• John Fletcher <J.P.Fletcher aston.ac.uk> May 29 2003
• Richard Grant <fractal clark.net> May 29 2003
• John Fletcher <J.P.Fletcher aston.ac.uk> Jun 02 2003
• "Walter" <walter digitalmars.com> Jun 02 2003

"Walter" <walter digitalmars.com> writes:

http://www.digitalmars.com/download/freecompiler.html

gf <mz_y2k yahoo...com> writes:

"Walter" <walter digitalmars.com> wrote in news:badt7c$3uc$1
 digitaldaemon.com:

 http://www.digitalmars.com/download/freecompiler.html



What's new/corrected in the new beta (and the previous)?

GF

"Walter" <walter digitalmars.com> writes:

 What's new/corrected in the new beta (and the previous)?


Mostly fixes to help boost compile.

John Fletcher <J.P.Fletcher aston.ac.uk> writes:

Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit



Walter wrote:

 What's new/corrected in the new beta (and the previous)?


 Mostly fixes to help boost compile.


I have tested the new beta against some code from the lambda library.

Here is an example which does not compile, extracted from this code,
together with output from the compiler.

John Fletcher

John Fletcher <J.P.Fletcher aston.ac.uk> writes:

Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Walter wrote:

 What's new/corrected in the new beta (and the previous)?


 Mostly fixes to help boost compile.


Here's another more complicated example from the same source.  I had to
copy quite a lot of code to be sure that everything is defined, but the
error is a oneliner at the very end.

John

Richard Grant <fractal clark.net> writes:

In article <3ECCB530.1C4DAF1E aston.ac.uk>, John Fletcher says...
This is a multi-part message in MIME format.
--------------DEC9756CD53B9448383DF924
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Walter wrote:

 What's new/corrected in the new beta (and the previous)?


 Mostly fixes to help boost compile.


Here's another more complicated example from the same source.  I had to
copy quite a lot of code to be sure that everything is defined, but the
error is a oneliner at the very end.


template <class HT, class TT> struct cons -- is not defined.

That does not mean that something wasn't causing a problem, just that your
example needs a little more work. What were you doing that generated your
original error?

Richard

John Fletcher <J.P.Fletcher aston.ac.uk> writes:

Richard Grant wrote:

 In article <3ECCB530.1C4DAF1E aston.ac.uk>, John Fletcher says...
This is a multi-part message in MIME format.
--------------DEC9756CD53B9448383DF924
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Walter wrote:

 What's new/corrected in the new beta (and the previous)?


 Mostly fixes to help boost compile.


Here's another more complicated example from the same source.  I had to
copy quite a lot of code to be sure that everything is defined, but the
error is a oneliner at the very end.


 template <class HT, class TT> struct cons -- is not defined.

 That does not mean that something wasn't causing a problem, just that your
 example needs a little more work. What were you doing that generated your
 original error?

 Richard


Sorry for delay - away sick over the weekend.  Yes, cons is another of the
macros. It may be in the other file I uploaded.  I will check it out - and
also run the examples on the next beta.

I was running the complete file and had exactly the same error as I reported.

John

John Fletcher <J.P.Fletcher aston.ac.uk> writes:

Richard Grant wrote:

 In article <3ECCB530.1C4DAF1E aston.ac.uk>, John Fletcher says...
This is a multi-part message in MIME format.
--------------DEC9756CD53B9448383DF924
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Walter wrote:

 What's new/corrected in the new beta (and the previous)?


 Mostly fixes to help boost compile.


Here's another more complicated example from the same source.  I had to
copy quite a lot of code to be sure that everything is defined, but the
error is a oneliner at the very end.


 template <class HT, class TT> struct cons -- is not defined.

 That does not mean that something wasn't causing a problem, just that your
 example needs a little more work. What were you doing that generated your
 original error?

 Richard


Adding the definition of cons does not change the error message at all.

The material comes from an old version of the lambda library before it was
added to the boost libraries.  The error message is reported by the complete
file.

John

Richard Grant <fractal clark.net> writes:

In article <3ED5CA49.F92ACED0 aston.ac.uk>, John Fletcher says...

The material comes from an old version of the lambda library before it was
added to the boost libraries.  The error message is reported by the complete
file.


Since it is older sources, I'm guessing it was a little buggy. Most folks expect
to see:

inline void fn() { }

as oppossed to:

void inline fn() { }

Richard

John Fletcher <J.P.Fletcher aston.ac.uk> writes:

Richard Grant wrote:

 In article <3ED5CA49.F92ACED0 aston.ac.uk>, John Fletcher says...

The material comes from an old version of the lambda library before it was
added to the boost libraries.  The error message is reported by the complete
file.


 Since it is older sources, I'm guessing it was a little buggy. Most folks
expect
 to see:

 inline void fn() { }

 as oppossed to:

 void inline fn() { }

 Richard


Richard

I am not clear whether the position of the inline at the start of the statement
is
compulsory or conventional.  As it compiled with another compiler in cannot be
said
to be a bug in the source code, so much as a compiler dependence issue, unless
the
other compiler went beyond the standard.

The code concerned (lambda library) is one which I started to use in my
programs and
it has prevented me using those versions with DM because the compiler didn't
compile
it.  That is quite a separate issue from the subsequent migration of that code
into
the Boost library in a different form.

It isn't my code so I didn't decide where to put the keywords.  I  have been
attempting to help improve DM by isolating the reasons why code which worked
e.g.
with gcc 2.95.n didn't work with DM.  Between us we have just found out one
reason,
and know a work around, for which thanks.

I hope sometime soon the whole of Boost will work with DM, but I know that is a
difficult target as Boost keeps growing.

John

"Walter" <walter digitalmars.com> writes:

"John Fletcher" <J.P.Fletcher aston.ac.uk> wrote in message
news:3EDB3502.66AB4724 aston.ac.uk...
 It isn't my code so I didn't decide where to put the keywords.  I  have


 attempting to help improve DM by isolating the reasons why code which


 with gcc 2.95.n didn't work with DM.  Between us we have just found out


 and know a work around, for which thanks.


I am grateful for the work you, Richard, and Christof have done to isolate
problems and thereby improve the compiler.

 I hope sometime soon the whole of Boost will work with DM, but I know that


 difficult target as Boost keeps growing.


We're not there 100% yet, but there have been big improvements.