The benefit to this is fairly small, but it does mean a little less work needs
to be done in the build scripts for cross compiling. The C version wasn't
especially concise so it was not obviously better in any way.
Character classification is no longer affected by quirks of standard library implementation
Lookup table for own function was generated with Python thanks to Unicode Database module from its standard library
Explicitly set locale for POSIX targets was reverted to C
https://forum.zdoom.org/viewtopic.php?t=65641&start=18#p1115930
- Disable the changes from commit 96afce241d
because it appears to not recognize default rules that reduce as
reducing.
- When tracing a parser, flush after every line output in case an
assert is hit so you can actually see what it did up to that point.
- Every update rolled into one, because I'm pretty sure I missed some while
updating lemon.c (not counting today's commits), since it wasn't always
updated at the same time as lemon.c.
- In particular, I think this check-in from 2016-06-06 was very important to
us after commit 3d5867d29e (For the
Lemon-generated parser, add a new action type SHIFTREDUCE and use it to
further compress the parser tables and improve parser performance.):
* Fix lempar.c so that the shift-reduce optimization works for error
processing.
— Fix a bug in destructorprocessing of Lemon. That has no impact on the SQLite grammar. The bug was introduced by prior work to optimize the Lemon-generated parser used by SQLite. (user: drh)
— Lemon bug fix: Do not merge action routines unless their destructors are
also identical. Problem and suggested fix reported on the mailing list by
Kelvin Sherlock. (user: drh)
— Lemon bug fix: Do not merge action routines unless their destructors are
also identical. Problem and suggested fix reported on the mailing list by
Kelvin Sherlock. (user: drh)
- Enhance Lemon so that it reorders the reduce rules such that rules without actions occur at the end and so that the first rule is number 0. This reduces the size of the jump table on the reduce switch, and helps the parser to run faster. (user: drh)
- Enhance Lemon so that if reduce code contains a comment of the form "/*A-overwrites-X*/" then a LHS label A is allowed to overwrite the RHS label X. (user: drh)
- In Lemon, add the ability for the left-most RHS label to be the same as the LHS label, causing the LHS values to be written directly into the stack. (user: drh)
- Improve the Lemon parser template (lempar.c) so that it avoids unnecessary work when the grammer defines YYNOERRORRECOVERY (as SQLite does). Slightly smaller and faster code results. (user: drh)
- Avoid recursion in the yy_find_shift_action() routine of the Lemon-generated parser, so that routine can be inlined, for a size reduction and performance increase. (user: drh)
- Fix uses of ctype functions (ex: isspace()) on signed characters in test programs and in some obscure extensions. No changes to the core. (user: drh)
- For the Lemon-generated parser, add a new action type SHIFTREDUCE and use it to further compress the parser tables and improve parser performance. (user: drh)
- Enhance the "lemon" executable so that it ignores -f, -W, -O, and -I command-line options. This permits most of the same options that are passed to the compiler to also be harmlessly passed to lemon, and thus simplifies makefiles. (user: drh)
- Modify the %nonassoc directive in lemon so that it generates a run-time error rather than a parsing conflict. This changes is due to a bug report on the mailing list. SQLite does not use the %nonassoc directive in its grammar so this change does not affect SQLite. (user: drh)
- In LEMON, limit the size of the grammar file to 100MB. This ensures that the program will never experience integer overflow. To be doubly sure, use calloc() instead of malloc() when allocating arrays. (user: drh)
