----- Original Message ----- From: "Ian MacLean" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Cc: <[EMAIL PROTECTED]> Sent: Friday, September 24, 2004 10:36 AM Subject: Re: [Nant-users] RE: can"t overwrite property
> [EMAIL PROTECTED] wrote: > > > I think the change in .85 to make overwrite a read-only property an > > exception is going to be a huge problem. This is because NAnt works > > exactly opposite in the way users expect, and it has gone unnoticed > > until this change. > > > > Commonly programs are configured such that the command line arguments > > override anything specified in a config file. NAnt works the other > > way; command line args are registered first, then the config file > > properties. People don't realize this previous to .85 because NAnt > > just logged an error instead of halting the build. It appeared the > > command line args overrode the config file, but what really happens is > > that the config file tries to override the command line and fails. > > > > Furthermore, the workaround is also confusing. People want to be able > > to override properties from the command line, and the solution is to > > add ' overwrite="false"' to the property tag? From a usability > > standpoint this is just bizzare. Only by realizing the config file > > overrides the command line you understand the why on setting it to > > false, but it still trips me up every time. > > > > I think this behavior/design needs to re-examined. I'd like to hear > > the devs and other users comment. > > > we have commented - and then some. Take a look at: > > > http://www.mail-archive.com/[EMAIL PROTECTED]/msg04761.html > http://www.mail-archive.com/[EMAIL PROTECTED]/msg04639.html > > and > > http://www.mail-archive.com/[EMAIL PROTECTED]/msg04601.html > there is probably more that I've missed. > > the consensus right now is that the param solution is probably the best > way to go but it will be more than a quick hack and probably won't make > the next release. I'm still not sure it really is the best way to go, as we need to be able to execute (quite) some logic (meaning tasks, targets and/or functions) in order to determine the default value of a given property. The <param> construct as proposed, does not allow this. We'll definitely need to give this more thought. > As I've said previously I'm almost in favour of > reverting to the previous behaviour until we have a clean solution in > place. Its seeming like the user confusion around this issue is > outweighing the benefits of correctly throw the exception when > overwriting readonly properties. Ian, for now we might just output a warning instead of having the build fail. Gert ------------------------------------------------------- This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170 Project Admins to receive an Apple iPod Mini FREE for your judgement on who ports your project to Linux PPC the best. Sponsored by IBM. Deadline: Sept. 24. Go here: http://sf.net/ppc_contest.php _______________________________________________ Nant-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/nant-users
