| Age | Commit message (Collapse) | Author |
|---|
|
from deprecated
|
|
addition the values optimization is now always applied because it become very cheep
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
-dump-source. implemented -dump-opt
|
|
|
|
|
|
|
|
|
|
merge Flags and ModuleFlags.
|
|
elimination optimization in PMCFG
|
|
data types.
|
|
character encoding correctly ; for this end, collected coding functions in GF.Text.Coding
|
|
startcatonly CFG trasnformation. Removed output formats that are now easily done with --cfg: "regular", "nolr".
|
|
|
|
|
|
|
|
|
|
GF.Speech.PGFToCFG, but should probably move somewhere else.
|
|
|
|
generated Haskell modules.
|
|
* output a PGF grammar in prolog readable syntax
* variables in abstract syntax (hypotheses and lambda-abstractions)
are translated to unique logical variables
* PGF terms in concrete syntax are translated to more prolog-like terms
|
|
|
|
|
|
Because the PGF file gets all the flags from the GFO file automatically, we get the coding flag in the PGF file too. This is undesirable, since PGF is supposed to always by in UTF-8.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
i.e. after pretty-printing instead
|
|
|
