Georg Baum
2014-05-26 20:07:01 UTC
Hi,
I am currently cleaning up my stash pile and finishing native support for
\smash[t] and \smash[b] (this came from bug 8967) and \notag (same as
\nonumber, but uses amsmath, this is helpful if you want to import AMS
example documents with tex2lyx).
These commands cause now amsmath to be automatically loaded, and before they
did not. This could in theory cause LaTeX erros, if a user used an own macro
with a name used by amsmath. Therefore, we could say that this is a file
format change, and disable amsmath in lyx2lyx if it was previously auto, one
of the new commands is used and no other amsmath command is used.
Alternatively, we could consider the current status as a bug: Since we have
the automatic amsmath loading already we could say that it should be
complete, consider the current behaviour as a bug and no file format change
is needed (this is different to the case where a new package is
automatically loaded: In that case it is always a file format change).
The disadvantage of the first option is that the non-loading of amsmath is
set for the future of this document which could cause similar issues as in
bug 9069. The disadvantage of the second option is that it could create
uncompilable documents.
What do you think? We cannot get a 100% correct solution, and I slightly
tend towards the second option, since IMHO amsmath is needed for any serious
math document anyway.
Georg
I am currently cleaning up my stash pile and finishing native support for
\smash[t] and \smash[b] (this came from bug 8967) and \notag (same as
\nonumber, but uses amsmath, this is helpful if you want to import AMS
example documents with tex2lyx).
These commands cause now amsmath to be automatically loaded, and before they
did not. This could in theory cause LaTeX erros, if a user used an own macro
with a name used by amsmath. Therefore, we could say that this is a file
format change, and disable amsmath in lyx2lyx if it was previously auto, one
of the new commands is used and no other amsmath command is used.
Alternatively, we could consider the current status as a bug: Since we have
the automatic amsmath loading already we could say that it should be
complete, consider the current behaviour as a bug and no file format change
is needed (this is different to the case where a new package is
automatically loaded: In that case it is always a file format change).
The disadvantage of the first option is that the non-loading of amsmath is
set for the future of this document which could cause similar issues as in
bug 9069. The disadvantage of the second option is that it could create
uncompilable documents.
What do you think? We cannot get a 100% correct solution, and I slightly
tend towards the second option, since IMHO amsmath is needed for any serious
math document anyway.
Georg