Re: from Danny's chat - 11/19
From: Jarle Stabell (jarle_at_remove_stuff_dlogikk_spam_kills_email.com)
Date: 12/21/04
- Next message: Nick Hodges [TeamB]: "Re: QC Report #: 7324 tops 500 votes"
- Previous message: Nick Hodges [TeamB]: "Re: QC Report #: 7324 tops 500 votes"
- In reply to: John Jacobson: "Re: from Danny's chat - 11/19"
- Next in thread: Jim Cooper: "Re: from Danny's chat - 11/19"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: Tue, 21 Dec 2004 21:11:17 +0100
John Jacobson wrote:
> "Kevin Berry" <kevin@berry_REM0VETH1S_ware.com> wrote in message
> news:41c83da1$1@newsgroups.borland.com...
>> Captain Jake wrote:
>>> Design by contract results in a nice clean syntax when it is built
>>> into the language, but it isn't anything you can not implement
>>> using asserts.
>>
>> One of the benefits of DBC is that unit testing can be automated
>> better. If you have a set of pre and post-conditions you can have
>> some kind of automated tool that uses this information to generate
>> test cases. At least that's the way I understand it. I've never
>> used DBC in practice so it's all just theory to me, but the theory
>> sounds good...
>
> This all sounds like something you would never want implemented in a
> release version. I can imagine the user delight upon encountering a
> cryptic error message like "Precondition x > 300 violated".
The alternative is undefined behaviour (which can result in your user
loosing valuable data), or (if you're lucky) your user getting an access
violation. With an assertion failure, a bug is "outed" and your user can
tell you the line number the problem was spotted.
Cheers,
Jarle
- Next message: Nick Hodges [TeamB]: "Re: QC Report #: 7324 tops 500 votes"
- Previous message: Nick Hodges [TeamB]: "Re: QC Report #: 7324 tops 500 votes"
- In reply to: John Jacobson: "Re: from Danny's chat - 11/19"
- Next in thread: Jim Cooper: "Re: from Danny's chat - 11/19"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|