Simon wrote: > qvx wrote: > > I guess that I have nothing against ticket #258. However, to > > automagically call the same function twice (in hope that the next time > > it will fare better) seems like "overdesign". > > It's called twice under the assumption, or hope if you prefer, that the > error will be handled in some manner.
I didn't know I was handling an error. How can I detect that I am handling an error? > On the other hand, I suppose you may have a point. Would it be better to > require explicit declaration in form of parameter tg_errors before a > method is deemed an error handler (for itself)? If we go this route, how > should failed validation be handled than? This way (via tg_errors), at least I can tell that I am handling an error. But, don't expect me to give you the answer. It would mean that I know something about the innards of TG, but the truth is that I don't know much about TG. The best that I can suggest to you is not to call the same function twice at all. qvx
