I think the LZX compiler could help out in this particular case, when it knows the type of the attribute. If it knows the type cannot sensibly take an empty value, it should warn.
I can maybe do this under the auspicies of LPP-7106... On Thu, Apr 1, 2010 at 1:59 PM, <[email protected]> wrote: > Thanks for the replies folks. I finally found it. I had unknowingly > picked up a *<acceptencodings="">* in a dataset tag, no doubt from the > editor auto-completion. I'm using Eclipse with Spket, and when inside the > tag, if you hit <ctrl-space>, the first item in the suggestion list is > acceptencodings. > > So, the question is, can we get the compiler to detect invalid values? > Is this worthy of filing a bug/improvement? > > > On 4/1/2010 10:28 AM, Suburban Antihero wrote: > > I most often see this when you set a type="boolean" and don't set a value. > > If that's not it, you literally are looking for incomplete syntax in a > constraint. The best bet here is usually to do a diff with your local > history to see what you may have changed. I use Eclipse and this is a > simple task. > > > > > On Apr 1, 2010, at 9:38 AM, [email protected] wrote: > > Of course the fault for the error lies entirely with me. However, I must > say, the following type of compiler error is about extremely unhelpful. > > Syntax error: the attribute value ended in mid-expression. > Look for an unclosed '(', '{', or '['. > > > Would there be any way to give some more information to narrow the scope > for the user? > > -- Henry Minsky Software Architect [email protected]
