| commit | 1584c53112162565f30adcfd51ab61451cecdd09 | |
| author | Kris Katterjohn <katterjohn@gmail.com> | |
| Thu, 4 Nov 2021 20:05:27 +0000 (4 16:05 -0400) | ||
| committer | Kris Katterjohn <katterjohn@gmail.com> | |
| Thu, 4 Nov 2021 20:05:27 +0000 (4 16:05 -0400) | ||
| tree | c7bd0c8f6ef07e523cf75114d2885a1a6eea9d5a | treesnapshot (tar.gz zip) |
| parent | 88c361e01d824f78d6d5198138775af7f0634423 | commitdiff |
MREAD-SYNERR: check for an EOF value when traversing *PARSE-WINDOW*
It makes sense to stop accumulating characters for an error message
when we find the EOF value *PARSE-STREAM-EOF* in *PARSE-WINDOW*.
This EOF value can show up in *PARSE-WINDOW* when loading a Maxima
file, for example.
I stumbled across this when I got a lisp error out of MREAD-SYNERR.
The EOF value (which is -1) wasn't checked for and this led to a
bogus CHAR= comparison.
No problems with the test suite or share test suite.
It makes sense to stop accumulating characters for an error message
when we find the EOF value *PARSE-STREAM-EOF* in *PARSE-WINDOW*.
This EOF value can show up in *PARSE-WINDOW* when loading a Maxima
file, for example.
I stumbled across this when I got a lisp error out of MREAD-SYNERR.
The EOF value (which is -1) wasn't checked for and this led to a
bogus CHAR= comparison.
No problems with the test suite or share test suite.
| src/nparse.lisp | diffblobblamehistory |