https://bz.apache.org/SpamAssassin/show_bug.cgi?id=7398
Bug ID: 7398
Summary: Unusual Gotcha when parsing rules..unicode
Product: Spamassassin
Version: 3.3.2
Hardware: PC
OS: Linux
Status: NEW
Severity: minor
Priority: P2
Component: spamassassin
Assignee: [email protected]
Reporter: [email protected]
Target Milestone: Undefined
Since I am here, might as well post this, as it trapped one of our people
trying to write some rules..
Started with a VERY simple regex that he could not get to score..
But it wasn't the rule he was working on, it was another legacy rule someone
wrote a while back.. for example..
/your\b\parent/i
Of course you can see the problem with the regex, the person added an extra
forward slash.. However, it was the results of that error, that make it worth
reporting.. Typically, you would have expected that \p to have been treated as
a literal, however when SA attempted to parse the message, it took this as a
unicode character, which when it failed to parse, simply dropped out, but
didn't fail lint or crash at that point, and it would not score any further
rules..
Apr 1 14:57:27.789 [16788] warn: plugin: eval failed: Timeout::_run: Can't
find Unicode property definition "a" at
/var/lib/spamassassin/3.003002/updates.spamassasin.org/test_rule.cf
After that, no more rule checking or scoring..
Suggest that lint might pick that up... otherwise without carefully looking at
the debug output, messages can bypass SA because it doesn't complete the
parsing and rule checking of the message, but it does complete it's run with no
errors..
Just thought I would post this since I was in the Bugzilla .. A gotcha that
could trap others.. and something that lint checking might need to pick up on..
Of course, this is an older version of SA, and this maybe have been addressed
already, but I could not find a matching report.
--
You are receiving this mail because:
You are the assignee for the bug.
