pgsql: Sync regex code with Tcl 8.6.4. - Mailing list pgsql-committers

From Tom Lane
Subject pgsql: Sync regex code with Tcl 8.6.4.
Date
Msg-id E1ZcIKi-0001RN-Gp@gemulon.postgresql.org
Whole thread Raw
List pgsql-committers
Sync regex code with Tcl 8.6.4.

Sync our regex code with upstream changes since last time we did this,
which was Tcl 8.5.11 (see commit 08fd6ff37f71485e2fc04bc6ce07d2a483c36702).

The only functional change here is to disbelieve that an octal escape is
three digits long if it would exceed \377.  That's a bug fix, but it's
a minor one and could change the interpretation of working regexes, so
don't back-patch.

In addition to that, s/INFINITY/DUPINF/ to eliminate the risk of collisions
with <math.h>'s macro, and s/LOCAL/NOPROP/ because that also seems like
an unnecessarily collision-prone macro name.

There were some other cosmetic changes in their copy that I did not adopt,
notably a rather half-hearted attempt at renaming some of the C functions
in a more verbose style.  (I'm not necessarily against the concept, but
renaming just a few functions in the package is not an improvement.)

Branch
------
master

Details
-------
http://git.postgresql.org/pg/commitdiff/b44d92b67b65a76f92448b5a282aae72820ac676

Modified Files
--------------
src/backend/regex/regc_lex.c |    6 ++++++
src/backend/regex/regcomp.c  |   14 +++++++-------
src/backend/regex/regexec.c  |    6 +++---
src/include/regex/regguts.h  |   13 +++++--------
4 files changed, 21 insertions(+), 18 deletions(-)


pgsql-committers by date:

Previous
From: Tom Lane
Date:
Subject: pgsql: Fix documentation of regular expression character-entry escapes.
Next
From: Robert Haas
Date:
Subject: pgsql: Determine whether it's safe to attempt a parallel plan for a que