On 09/27/2018 03:57 AM, Daniel wrote: > On 27/09/2018 02:04, Richard Kimberly Heck wrote: >> On 9/26/18 2:51 AM, Daniel wrote: >>> On 25/09/2018 18:16, Richard Kimberly Heck wrote: >>>> On 9/25/18 10:21 AM, Daniel wrote: >>>>> Hi >>>>> >>>>> When a document with a child document with different textclass is >>>>> typeset LyX warns about this. I am wondering why that is since as far >>>>> as I understood the child document does not incorporate any class >>>>> information into the master document. So, is this warning because I >>>>> could have used commands incompatible with the master document, like >>>>> chapter when the master is an article? >>>> >>>> Yes. >>>> >>>>> But then LyX would warn me anyway, or? >>>> >>>> What do you mean? When might it warn you otherwise? >>> >>> For example, if I compile an article master with a child book that has >>> a chapter, then LyX will throw an "undefined control sequence" error. >> >> True. The 'different textclasses' error is meant to be a pre-warning in >> a way about that. The LaTeX warning might be harder for some users to >> trace. > > I had the opposite reaction thinking there must be something > particular to LyX why I should not have different textclasses over and > beyond what concerns LaTeX. So to me it was rather confusing than > helpful. But that might just be me. Maybe add some more explanatory > text to the message box as to why this might be a problem, e.g. > layouts being unavailable etc.? (see attachment of the original)
Our error messages are often a bit terse, so feel free to propose an improvment. > >>> Also, I find LyX's extra warning about the textclass a bit misleading. >>> There seems to be nothing special about having another textclass >>> compared to, say, using other modules. LyX does not create an extra >>> warning in the latter case but the problem might be basically the same. >> >> It does, actually. See InsetInclude.cpp, around line 760. > > Yes, as I wrote in the other email, I stand corrected. There is a > message box. It even comes up independently of whether one has > actually used the module in the child document document. But it in > contrast to the textclass warning it cannot be silenced kind of > forcing one to make changes to the child document (see attachment). > But maybe that's by design. Probably just an oversight. The silencing functionality was added much later. Riki