On Sat, Nov 07, 2009 at 04:36:22AM +0100, Uwe Stöhr wrote: > > The moral is that what you do in the preamble defies LyX and we can't do > > much about that. Or are you proposing that we should give the > > possibility to disable any package automatically loaded by LyX? Because > > I don't see why mhchem should be special (note that I personally don't > > use that package). > > I understand your point. But for example the esint handling was > introduced because this package redefines commands that are also used > by other packages. The two-letter commands defined by mhchem are error > prone as well. Besides this, mhchem is our third package that is > automatically loaded when formulas are edited. Especially formulas are > often redefined - we therefore have the math macro feature. So to have > the feature to disable mhchem is an add on and not a must have. I just > thought it doesn't harm and is useful for the mentioned reasons.
So, you claim that mhchem should be treated specially because its two-letter commands are easy to clash with. I think I can agree, but this doesn't solve the need for a general solution, and, most importantly, if the commands are not defined in a LyX macro, the handling of \ce and \cf remains the same, such that if they are defined as not taking a parameter, the same problem still occurs, as demonstrated by the attached documents (compare their output, which should be the same). This means that not only the loading of mhchem should be disabled, but even the behavior of LyX should be changed when mhchem is disabled, and your patch doesn't do this. -- Enrico
ce-1.lyx
Description: application/lyx
ce-2.lyx
Description: application/lyx
