https://bugs.exim.org/show_bug.cgi?id=2761
Bug ID: 2761
Summary: DFA partial matching does not yield same results as
non-partial
Product: PCRE
Version: 10.36 (PCRE2)
Hardware: x86-64
OS: Linux
Status: NEW
Severity: bug
Priority: medium
Component: Code
Assignee: Philip.Hazel@???
Reporter: stevenjshuck@???
CC: pcre-dev@???
Created attachment 1390
-->
https://bugs.exim.org/attachment.cgi?id=1390&action=edit
dfa.c
I am attempting to match the pattern
<[^<]*>
to the subject
< foo < bar >
in two different ways:
1. a single contiguous subject "< foo < bar >"
2. the subject chunked into two strings, "< foo " and "< bar >", detecting
partial match and restarting per the docs.
Please find attached my test case. I compiled and ran it these ways:
$ gcc -DPARTIAL=PCRE2_PARTIAL_HARD dfa.c -o dfa -lpcre2-8 && ./dfa
$ gcc -DPARTIAL=PCRE2_PARTIAL_SOFT dfa.c -o dfa -lpcre2-8 && ./dfa
This was the output either way:
Pattern: "<[^<]*>"
Subject: "< foo < bar >"
Result: 1
First two ovector values: 6, 13
Subject: "< foo "
Result: -2
First two ovector values: 0, 6
Restarting...
Subject: "< bar >"
Result: -1
First two ovector values: 0, 6
For the purposes of this ticket, I don't care so much what the results are,
just that they be consistent between the chunked and non-chunked subjects.
--
You are receiving this mail because:
You are on the CC list for the bug.
