| 1 |
Syntax: bf(%expect) tt(number) |
| 2 |
|
| 3 |
B() normally warns if there are any conflicts in the grammar (see section |
| 4 |
ref(SHIFTREDUCE)), but many real grammars have harmless em(shift/reduce |
| 5 |
conflicts) which are resolved in a predictable way and which would be |
| 6 |
difficult to eliminate. It is desirable to suppress the warning about these |
| 7 |
conflicts unless the number of conflicts changes. You can do this with the |
| 8 |
tt(%expect) declaration. |
| 9 |
|
| 10 |
The argument tt(number) is a decimal integer. The declaration says there |
| 11 |
should be no warning if there are tt(number) shift/reduce conflicts and no |
| 12 |
em(reduce/reduce conflicts). The usual warning is given if there are either |
| 13 |
em(more) or em(fewer) conflicts, em(or) if there are em(any) reduce/reduce |
| 14 |
conflicts. |
| 15 |
|
| 16 |
In general, using tt(%expect) involves these steps: |
| 17 |
itemization( |
| 18 |
it() Compile your grammar without tt(%expect). Use the `tt(-V)' option to |
| 19 |
get a verbose list of where the conflicts occur. B() will also print the |
| 20 |
number of conflicts. |
| 21 |
it() Check each of the conflicts to make sure that b()'s default |
| 22 |
resolution is what you really want. If not, rewrite the grammar and go back to |
| 23 |
the beginning. |
| 24 |
it() Add an tt(%expect) declaration, copying the number of (shift-reduce) |
| 25 |
conflict printed by b(). |
| 26 |
) |
| 27 |
Now b() will stop annoying you about the conflicts you have checked, but |
| 28 |
it will warn you again if changes in the grammar result in a different number |
| 29 |
or type of conflicts. |
| 30 |
|
| 31 |
|
| 32 |
|
| 33 |
|