> I'm afraid this does not make sense to me. An "interactive session" is > not some monolithic blob. It is a sequence of one or more (often very > many) discreet interactions. Any one of these may produce an > exceptional condition, and it seems to me that each individual > interaction that elicits an exception should be handled independently > and if it does produce an exceptional condition, that should be > reported as soon as it occurs. Things should then be cleaned up, as > necessary and feasible, and the interactive session should be allowed > to continue.
Sorry, indeed I was simplifying beyond recognition. Actually, I'm writing a GUI front-end to R, and what I called an "interactive session" is technically not an interactive session in R at all, but would only seem like one to the user. In the GUI, there are R expressions being run from some application code, but also there is a sort of console, where the user can interact with R more or less directly, like they would in a regular interactive session. What I mean by "special handling", then, is not so much reacting to the warnings/errors. Rather, the thing I'm really trying to do, is just to figure out, which output is "regular" output, which is a warning, and which is an error. All output gets displayed right in time, for the user to react to. However, one use case would be to mark up all warnings in a certain color and all errors in another. Also, for commands running for GUI-actions, it would be nice to inform the user - if applicable - "there have been the following warnings" or "an error occurred, see below". So basically, my programmer's dream would be to have three simple callbacks: one each for "regular output", warnings and errors. While it's not quite as simple, I have most things in place: I'm catching all that would be going to stdout via ptr_R_WriteConsole. I'm catching the errors using my own callback in options (error=...). Now all I need is a way to find out about warnings. It seemed to me, options (warning.expression=...) would be a neat way to do that, but I could not find out, how to use it. Hope that clarifies what I need this for. BTW, the project in question is this: http://rkward.sf.net Regards Thomas ______________________________________________ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
