#86 closed defect (fixed)
Fix parse edit offset
| Reported by: | David Aspinall | Owned by: | alex heneveld |
|---|---|---|---|
| Priority: | blocker | Milestone: | |
| Component: | 1:pg-eclipse | Keywords: | |
| Cc: |
Description (last modified by )
Editing document sets the parse edit offset, the position from which parsing begins next time.
The code for adjusting this had faulty logic in earlier versions. It was fixed but seems to have broken again, perhaps with parser adjustments. This causes bad breakage because it gets difficult to edit text.
Change History (6)
comment:1 Changed 17 years ago by
| Description: | modified (diff) |
|---|---|
| Priority: | major → blocker |
comment:2 Changed 17 years ago by
| Resolution: | → fixed |
|---|---|
| Status: | new → closed |
fixed as part of 124. agree we want some tests, will be added as new trac task.
comment:3 Changed 17 years ago by
| Resolution: | fixed |
|---|---|
| Status: | closed → reopened |
not quite fixed, still breaks when symbols are in use. (possibly related, "undo" can put the curpos in a funny place.)
comment:4 Changed 17 years ago by
| Owner: | changed from David Aspinall to alex heneveld |
|---|---|
| Status: | reopened → new |
comment:5 Changed 17 years ago by
| Resolution: | → fixed |
|---|---|
| Status: | new → closed |
whitespace handling is inconsistent (observed earlier), this was causing errors (seems to be wherever symbols are used but haven't explored that) made code more forgiving in more places as to whether whitespace is in the parse or not; seems to fix things
also fixed some related off-by-one errors (sending "goto cursor" when cursor is positioned at the start of a line should _not_ send the command on that line), added comments to specify the meaning of arguments (whether it refers to the char at that position or the inter-char position just before it, was ambiguous in several places)
We need to add/identifiy a test case for this in ed.inf.proofgeneral.tests