On Thu, Apr 06, 2006 at 11:36:52AM +0200, Lars Gullik Bjønnes wrote: > Andre Poenitz <[EMAIL PROTECTED]> writes: > > | On Wed, Apr 05, 2006 at 01:20:48AM +0200, Lars Gullik Bjønnes wrote: > | > "Bo Peng" <[EMAIL PROTECTED]> writes: > | > > | > | > Could you please post a use case or two where you think exceptions > | > | > will improve LyX significantly? > | > | > | > | When I get my hands dirty with output_latex, paragraphlist, cursor > | > | etc, lyx will simply crash for a misuse of function, and leave me no > | > | clue what goes wrong. I would vote for much more assert and throw in > | > | the code. Attached is something I use in my own project: > | > > | > You need a good reason to not be able to go with BOOST_ASSERT > | > | A hard crash is nothing I can sell _my_ customers. Under no > | circumstances. The best thing is to have a message box and > | let the user decide to let two weeks worth of simulation die > | or give him a chance to make sense of the data collected so far. > > Note that we can make BOOST_ASSERT throw if we want it to.
Ah. I did not know taht. Andre'
