[Search for users]
[Overall Top Noters]
[List of all Conferences]
[Download this site]
Title: | Microsoft Visual C/C++ |
|
Moderator: | PLUGH::needle |
|
Created: | Tue Mar 16 1993 |
Last Modified: | Wed Jun 04 1997 |
Last Successful Update: | Fri Jun 06 1997 |
Number of topics: | 1121 |
Total number of notes: | 4385 |
Hi, the following is a note I have placed in the microsoft.public.vc.language
newsgroup. I thought it might be of interest here.
> Hi,
>
> Like everybody else I too have been battling moving a project from 4.2 to
> 5.0. I have a large framework library (200k lines) which makes extensive
> use of STL and standard C++ features. So far I have spent the last week and
> a half porting, I've finally got everything compiled but I'm yet to test. I
> should add that I introduced namespaces into my project for the first time,
> so a bit of time was spent going this, however most of my time was spent
> finding workarounds for compiler problems.
>
> A few of the problems I encountered..
>
> 1/ INTERNAL COMPILER ERROR
>
> This baby turned up quite frequently, fortunately it always told me which
> line in my source it was having trouble with.
>
> The problem was always to do with too many of the same namespace directives
> in the one source. Removing a few of them solved the problem, but this was
> not always easy as redundant directives were required in places to get the
> STL to compile.
>
> Talk about being caught between a rock and a hard place!
>
> 2/ COMPILER IS OUT OF HEAP SPACE (C1060)
>
> A few of my source files were large (2000 lines) and included a lot of STL,
> with the introduction of namespaces the compiler complained and told me to
> use the /Zm option to increase space. This did not help, these files
> required 1/2 a Gig of mem, so I had to arrange code in new source files.
>
> 3/ NAMESPACE-STL PROBLEMS
>
> These seem to be well known in this newsgroup now so I won't detail, except
> to say I lost most of my time here. When you have many different namspaces
> in a project and use STL as well this creates enormous handaches esp
> because the compiler only tells you about a problem within an STL source
> file and not where in your own source it is having problem with.
>
> The only way to figure out where the problem lies is to start commenting
> out source bottom-up, and this takes a lot of time.
>
> 4/ STL MAP BUG (Alpha Only)
>
> I reported this problem last week in this newsgroup, fortunately for intel
> users it does not apply so I've been told, it is only a bug on Alpha.
> However for Alpha users the fix I reported does work. (I'll attach my old
> note to this thread)
>
> 5/ LNK4078 WARNING
>
> Complains "multiple .xdata sections found with different attributes". I
> don't know what this means, this help system does provide much insight.
> Occasionally I don't get this warning - I don't understand??
>
> 6/ VIRTUAL FUNCTION BUG
>
> The following example code does not compile...
>
> class SET_OFBase : public AbstractData
> {
> public:
> // .....
>
> virtual std::vector< AbstractData*> elementIndex() const = 0;
> };
>
> template< class T >
> class SET_OF : public SET_OFBase, public std::set<T>
> {
> public:
> // ....
>
> virtual std::vector< AbstractData*> elementIndex() const;
> };
>
> the following does..
>
> typedef std::vector< AbstractData*> AbstractDataVector;
> class SET_OFBase : public AbstractData
> {
> public:
> // .....
>
> virtual AbstractDataVector elementIndex() const = 0;
> };
>
> template< class T >
> class SET_OF : public SET_OFBase, public std::set<T>
> {
> public:
> // ....
>
> virtual AbstractDataVector elementIndex() const;
> };
>
[Posted by WWW Notes gateway]
T.R | Title | User | Personal Name | Date | Lines |
---|
1093.1 | thanks | DECWET::PETERSON | | Tue Mar 18 1997 12:25 | 14 |
| thanks for catching the alpha bugs:
re 4/ STL MAP BUG
we posted a knowledge base article about the bug and its workaround to
Microsoft product support. Thanks for finding it and posting the
workaround yourself
re 5/ LNK4078 WARNING
fixed in the release version. the warning only occurred when code was
compiled with /GX (C++ exception handling support). .xdata section was
generated to be writable, when it should be read-only - the linker
handled it by making all .xdata section read-write
|
1093.2 | Release version? | NETRIX::"[email protected]" | Stephen Moore | Tue Mar 18 1997 22:44 | 7 |
| >> fixed in the release version.
where can one get this from please?
Thanks
Steve
[Posted by WWW Notes gateway]
|
1093.3 | | DECCXL::OUELLETTE | crunch | Wed Mar 19 1997 13:45 | 1 |
| The release version for Alpha is not yet available.
|