summaryrefslogtreecommitdiff
Commit message (Expand)AuthorAgeFilesLines
* Now using a single parse table with multiple entry points, instead of multipleAdrian Thurston2009-02-1314-154/+220
* got graphviz output workingAdrian Thurston2009-02-086-304/+62
* Can now install and run colm from the installed location.Adrian Thurston2009-02-072-10/+22
* Colm logging and target logging are now turned on with -v and -l. No need toAdrian Thurston2009-02-078-85/+77
* In the colm-only code guard log statements with a runtime flag. In the runtimeAdrian Thurston2009-02-0711-272/+703
* Reorganized the src variables. All sources are now a combination of RUNTIME andAdrian Thurston2009-02-071-15/+19
* Can use qualifications in yield statements.Adrian Thurston2009-02-071-1/+8
* Need to pop the reference param qualifications after destroying the iterator ...Adrian Thurston2009-02-072-3/+7
* REF_FROM_QUAL_REF needs to fetch the field. After a call with reference argsAdrian Thurston2009-02-075-4/+71
* Started on producing refefences from qualified items. Working on callingAdrian Thurston2009-02-065-13/+144
* Fixed a broken reference to flags (from alg) that was wrapped in an ifdefAdrian Thurston2009-02-061-1/+1
* Modified the example in the thesis. Copied it here.Adrian Thurston2008-12-091-27/+71
* The make_tree and make_token functions now produce normal non-parse trees. TheAdrian Thurston2008-11-236-76/+65
* Removed the parse algorithm data from the Tree structure. Not sending ignoreAdrian Thurston2008-11-226-20/+34
* More on the distinction between parse trees and simple trees, though they areAdrian Thurston2008-11-223-53/+56
* Start differentiating between trees and parse trees.Adrian Thurston2008-11-227-12/+48
* Removed the Alg data structure from the program. Sorted out some issues withAdrian Thurston2008-11-214-48/+17
* Moved the fields from the Alg structure into Tree.Adrian Thurston2008-11-216-102/+95
* Start on the movement of algorithm data back to the tree structure. Use a flagAdrian Thurston2008-11-213-4/+19
* Added some comments for the AF_ flags. Changed AF_GENERATED to AF_TERM_DUP toAdrian Thurston2008-11-205-88/+111
* Warning and error fixes for compiling with g++ 4.3.2.Adrian Thurston2008-11-136-7/+8
* Be a bit more careful about when we save and restore the original lhs parseAdrian Thurston2008-11-126-30/+88
* Removed Alg::parsed.Adrian Thurston2008-11-122-6/+0
* In the commit code restore the tree from the tree pointer extracted from theAdrian Thurston2008-11-121-3/+3
* Don't need to clear the entire vm_stack.Adrian Thurston2008-11-121-2/+1
* Use the result of the RESTORE_LHS after running undo code. Assert the resultAdrian Thurston2008-11-123-12/+18
* Stop using parsed_downref and instead downref parsed in the rcode_downref underAdrian Thurston2008-11-124-103/+9
* Need a VM stack for parsing patterns. The VM stack is now used for more thanAdrian Thurston2008-11-113-5/+17
* In XML output print the ignore tokens ahead of the token and on the same level.Adrian Thurston2008-11-113-25/+76
* need to call the parse functionAdrian Thurston2008-11-111-11/+12
* When committing we need to take into account the causeReduce of the next inputAdrian Thurston2008-11-113-10/+17
* Need to consider causeReduce when backing up over rcode in the the commit fun...Adrian Thurston2008-11-101-11/+56
* Added a RESTORE_LHS reverse instruction that is added to the reverse code whenAdrian Thurston2008-11-103-34/+93
* The order of node visitation in the commit code was changed. It was previouslyAdrian Thurston2008-11-093-35/+35
* Only store the original tree in Alg::parsed if the tree was changed.Adrian Thurston2008-11-091-4/+10
* Name change: gbl -> rtd.Adrian Thurston2008-11-084-42/+42
* Now downreffing all 'parsed' trees on a full commit only when revert is not on. Adrian Thurston2008-11-083-43/+67
* A global commit with revertOn cannot clear the parsed fields from Alg. Shown byAdrian Thurston2008-11-083-2/+130
* User iterator calls now distinguish between _WV and _WC code.Adrian Thurston2008-11-085-24/+84
* Removed upref_uiter_args from IN_UITER_CREATE. Seems to be left over from aAdrian Thurston2008-11-071-12/+0
* Split the IN_PARSE instruction into _WV and _WC versions instead of using aAdrian Thurston2008-11-073-16/+44
* Reverse code is now transferred to the allReverseCode buffer after eachAdrian Thurston2008-11-077-93/+112
* function movement.Adrian Thurston2008-11-071-31/+32
* Added the _notoken language element. This ignore token will be sent when aAdrian Thurston2008-11-062-3/+12
* Stop using AF_IGNORE, a peek into lelInfo is the right way to determine if aAdrian Thurston2008-11-063-7/+8
* The compile functions should be const. We need to call them more than once toAdrian Thurston2008-11-063-120/+111
* The revertOn flag is now passed to parser. But there are problems ... compilingAdrian Thurston2008-11-065-11/+17
* more work on the specialization of frame code into codeWV and codeWCAdrian Thurston2008-11-058-73/+84
* Some fixes for _WV and _WC frames. Also some cleanup.Adrian Thurston2008-11-056-40/+37
* Fixed some inconsistent ordering of _WV and _WC variables.Adrian Thurston2008-11-053-30/+31