|
18537
|
1 |
|
|
|
2 |
text {*
|
|
|
3 |
The Isar toplevel works differently for interactive developments
|
|
|
4 |
vs.\ batch processing of theory sources. For example, diagnostic
|
|
|
5 |
commands produce a warning batch mode, because they are considered
|
|
|
6 |
alien to the final theory document being produced eventually.
|
|
|
7 |
Moreover, full @{text undo} with intermediate checkpoints to protect
|
|
|
8 |
against destroying theories accidentally are limited to interactive
|
|
|
9 |
mode. In batch mode there is only a single strictly linear stream
|
|
|
10 |
of potentially desctructive theory transformations.
|
|
|
11 |
*}
|
|
|
12 |
|
|
|
13 |
\item @{ML Toplevel.empty} is an empty transition; the Isar command
|
|
|
14 |
dispatcher internally applies @{ML Toplevel.name} (for the command)
|
|
|
15 |
name and @{ML Toplevel.position} for the source position.
|
|
|
16 |
|