19 messages in com.mysql.lists.plusplusRe: Link errors with resetdb example ...| From | Sent On | Attachments |
|---|---|---|
| Dale Smith | 08 Jun 2005 12:20 | |
| Chris Frey | 08 Jun 2005 13:22 | |
| Dale Smith | 08 Jun 2005 13:31 | |
| Chris Frey | 08 Jun 2005 13:44 | |
| Dale Smith | 08 Jun 2005 14:00 | |
| Dale Smith | 08 Jun 2005 14:23 | |
| Chris Frey | 08 Jun 2005 16:38 | |
| Warren Young | 09 Jun 2005 07:28 | |
| Dale Smith | 09 Jun 2005 07:31 | |
| Dale Smith | 10 Jun 2005 12:03 | |
| Warren Young | 10 Jun 2005 13:10 | |
| Dale Smith | 10 Jun 2005 13:26 | |
| Warren Young | 10 Jun 2005 14:31 | |
| Chris Frey | 12 Jun 2005 02:19 | |
| Dale Smith | 13 Jun 2005 07:40 | |
| Dale Smith | 13 Jun 2005 08:25 | |
| Warren Young | 13 Jun 2005 12:45 | |
| Chris Frey | 13 Jun 2005 20:24 | |
| Warren Young | 14 Jun 2005 11:30 |
| Subject: | Re: Link errors with resetdb example program![]() |
|---|---|
| From: | Warren Young (mysq...@etr-usa.com) |
| Date: | 06/13/2005 12:45:06 PM |
| List: | com.mysql.lists.plusplus |
Chris Frey wrote:
The iostream classes were designed with inheritance in mind,
To an extent, yes. Clearly IOStreams is a class hierarchy, with base classes designed to be derived from. What I dispute is whether it's a good idea to derive from the leaf classes. For one thing, stringstream has no virtual functions.
What we're inheriting here is primarily interface, not implementation. We could start inheriting far up the tree -- from basic_ostream perhaps -- and still work with all existing end-user code. Query could maintain a basic_stringbuf member, and would work just the same.
I admit, I latched onto Dale's problem report just hoping that it was more reason to make this change, and it turns out that it isn't. But the recent VC++ problems alone are enough to keep me nervous about continuing on this path.
The reason to get rid of stream inheritance would be the copying that Query objects might do,
Can you clarify that, please?




