Update of bug #68570 (group groff):
Status: In Progress => Fixed
Open/Closed: Open => Closed
_______________________________________________________
Follow-up Comment #1:
commit 4ec63c0d1fcbd383df1602078c61e1f9eaba801d
Author: G. Branden Robinson <[email protected]>
Date: Sun Jul 26 20:47:24 2026 -0500
[tbl]: Regression-test Savannah #68570.
* src/preproc/tbl/tests/do-not-segv-on-span-only-table-description.sh:
Do it.
* src/preproc/tbl/tbl.am (tbl_TESTS): Run test.
Test fails at this commit.
commit 1bbb5ff2f10362979e1609ae5ad95ba2a9f57394
Author: G. Branden Robinson <[email protected]>
Date: Sun Jul 26 20:48:40 2026 -0500
[tbl]: Fix Savannah #68570.
* src/preproc/tbl/main.cpp (process_format): Fix invalid memory access
(SEGV) when a table description consists of multiple rows of
horizontal span column classifiers (only). When processing each row
descriptor, keep track--with new local Boolean `is_first_column` of
whether we're processing the first column descriptor in that row. If
we encounter a horizontal span column classifier in the first column,
throw an error diagnostic instead of setting `got_format` true and
assigning the "type" `t` to `FORMAT_SPAN`. Set the new Boolean false
after successfully interpreting one column descriptor in the row
descriptor. Reset the new Boolean to true when advancing to a new
row description.
Fixes <https://savannah.gnu.org/bugs/?68570>. Problem appears to date
back to groff's birth. Thanks to Elias Hasas for the report and an
economical reproducer.
ANNOUNCE: Acknowledge Elias.
_______________________________________________________
Reply to this item at:
<https://savannah.gnu.org/bugs/?68570>
_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/
signature.asc
Description: PGP signature
