[Bug regex/1302] improve regex bitset performance (don't assume 32 bits)

[Bug regex/1302] improve regex bitset performance (don't assume 32 bits)

[cvs-commit at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=1302

--- Comment #4 from cvs-commit at gcc dot gnu.org <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Paul Eggert <eggert@sourceware.org>:

https://sourceware.org/git/gitweb.cgi?p=glibc.git;h=2cc478ed1be82711a6cac15aae683530b2e6732b

commit 2cc478ed1be82711a6cac15aae683530b2e6732b
Author: Paul Eggert <eggert@cs.ucla.edu>
Date:   Tue Aug 4 23:45:27 2020 -0700

    Copy regex_internal.h from Gnulib

    Sync this file from Gnulib, thus incorporating the following
    fix for a bug with regexps with 16 or more subexpressions:
    * posix/regex_internal.h (struct re_backref_cache_entry):
    Use bitset_word_t as the type of eps_reachable_subexps_map,
    instead of unsigned short int.  This fixes a bug I introduced
    to glibc in 2005-09-28T17:33:18Z!drepper@redhat.com (glibc commit
    2c05d33f90861d074dc12808dafbde30f487b1a0, BZ #1302).
    Remove unused member 'unused'.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug regex/1302] New: improve regex bitset performance (don't assume 32 bits)

[eggert at gnu dot org]

The regex bitset code currently uses an array of unsigned int words to
represent a bitset, and assumes these words are exactly 32 bits wide.
On modern 64-bit hosts, using a 64-bit word would require
approximately half the instructions for the bitset-related code.
Also, it's better not to assume 32-bits everywhere.  I'll attach a
patch.

-- 
           Summary: improve regex bitset performance (don't assume 32 bits)
           Product: glibc
           Version: 2.3.5
            Status: NEW
          Severity: normal
          Priority: P2
         Component: regex
        AssignedTo: gotom at debian dot or dot jp
        ReportedBy: eggert at gnu dot org
                CC: glibc-bugs-regex at sources dot redhat dot com,glibc-
                    bugs at sources dot redhat dot com
 BugsThisDependsOn: 1278,1285


http://sources.redhat.com/bugzilla/show_bug.cgi?id=1302

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug regex/1302] improve regex bitset performance (don't assume 32 bits)

[eggert at gnu dot org]

------- Additional Comments From eggert at gnu dot org  2005-09-06 07:33 -------
Created an attachment (id=650)
 --> (http://sources.redhat.com/bugzilla/attachment.cgi?id=650&action=view)
bitset performance / portability improvements


-- 


http://sources.redhat.com/bugzilla/show_bug.cgi?id=1302

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug regex/1302] improve regex bitset performance (don't assume 32 bits)

[eggert at gnu dot org]

------- Additional Comments From eggert at gnu dot org  2005-09-06 17:48 -------
Created an attachment (id=651)
 --> (http://sourceware.org/bugzilla/attachment.cgi?id=651&action=view)
bitset performance / portability improvements (revised)

Derek Price reported that the previous patch
elicits a "suggest parentheses around + or - in
operand of &" with gcc -Wall.  The revised patch
adds a pair of parentheses to avoid this warning.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
 Attachment #650 is|0                           |1
           obsolete|                            |


http://sourceware.org/bugzilla/show_bug.cgi?id=1302

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug regex/1302] improve regex bitset performance (don't assume 32 bits)

[drepper at redhat dot com]

-- 
Bug 1302 depends on bug 1278, which changed state.

Bug 1278 Summary: regex undefined behavior with shifting past word length
http://sourceware.org/bugzilla/show_bug.cgi?id=1278

           What    |Old Value                   |New Value
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

http://sourceware.org/bugzilla/show_bug.cgi?id=1302

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug regex/1302] improve regex bitset performance (don't assume 32 bits)

[drepper at redhat dot com]

------- Additional Comments From drepper at redhat dot com  2005-09-28 17:33 -------
I've checked in a significantly changed version of this patch.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED


http://sourceware.org/bugzilla/show_bug.cgi?id=1302

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug regex/1302] improve regex bitset performance (don't assume 32 bits)

[drepper at redhat dot com]

-- 
Bug 1302 depends on bug 1285, which changed state.

Bug 1285 Summary: regex code should use 'bool' (plus some bug fixes)
http://sourceware.org/bugzilla/show_bug.cgi?id=1285

           What    |Old Value                   |New Value
----------------------------------------------------------------------------
             Status|WAITING                     |RESOLVED
         Resolution|                            |WONTFIX

http://sourceware.org/bugzilla/show_bug.cgi?id=1302

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.