On Tue, 2 Dec 2014, jdow wrote:
On 2014-12-02 12:15, Kevin A. McGrail wrote:
On 12/2/2014 3:10 PM, jdow wrote:
> On 2014-12-02 10:10, John Hardin wrote:
> > On Tue, 2 Dec 2014, Burnie wrote:
> >
> > > On 12/02/2014 03:12 AM, John Hardin wrote:
> > > > On Mon, 1 Dec 2014, Kevin A. McGrail wrote:
> > > >
> > > > > On 12/1/2014 8:03 PM, John Hardin wrote:
> > > > > > For now, the only issue that has ever arisen in years is the
> > > > > > 5010000
> > > > > test so I would stick with just that for now.
> > > >
> > > > Ok.
> > > >
> > > > https://issues.apache.org/SpamAssassin/show_bug.cgi?id=7107
> > >
> > > Just FYI: The nested if example in the patch/doc will still
> > > give a lint warning for perl < 5.10
> > >
> > > if can(Mail::SpamAssassin::Conf::perl_min_version_5010000)
> > > if version > 3.004001 && perl_version >= 5.018000
> > > body INVALID_RE_SYNTAX_IN_PERL_BEFORE_5_18 /(?[ \p{Thai} &
> > > \p{Digit} ])/
> > > endif
> > > endif
> > >
> > > Dec 2 03:53:48.550 [30251] warn: Argument "perl_version" isn't
> > > numeric in
> > > numeric ge (>=) at (eval 2521) line 2.
> >
> > ARGH!
> >
> > Well, I suppose we're back to hoping the distro maintainers accept the
> > perl_version patch for their LTR release versions of older SA
> > releases.
> >
> > > - IMHO, that single '+' character may be the single most annoying
> > > character in
> > > SA for years? :-\
> >
> > indeed.
>
> Does this show the error?
>
> if can(Mail::SpamAssassin::Conf::perl_min_version_5010000)
> && version > 3.004001 && perl_version >= 5.018000
> body INVALID_RE_SYNTAX_IN_PERL_BEFORE_5_18 /(?[ \p{Thai} &
> \p{Digit} ])/
> endif
>
> Perhaps the same trick can (almost) work again.
>
> {^_^}
There is no need for the other checks.
if can(Mail::SpamAssassin::Conf::perl_min_version_5010000) is enough since
it
doesn't exist until 3.4.1.
If you are locally using SA trunk and writing your own rules that require
certain perl_versions, you can use the if perl_version >= XYZ logic
without concern.
regards,
KAM
Perhaps test it just the same to see if the basic technique works? I suspect
it should. That way it may be a messy way to handle the problem without
falling into even nastier messes.
I suspect it will fail just the same. I think it's something related to
shortcut logic in evals in 5.8.x, but I couldn't find anything in the
various perl release notes that looked relevant to that.
--
John Hardin KA7OHZ http://www.impsec.org/~jhardin/
jhar...@impsec.org FALaholic #11174 pgpk -a jhar...@impsec.org
key: 0xB8732E79 -- 2D8C 34F4 6411 F507 136C AF76 D822 E6E6 B873 2E79
-----------------------------------------------------------------------
Running away is the coward's way out of a war;
appeasement is the coward's way into a war. -- Thorax
-----------------------------------------------------------------------
13 days until Bill of Rights day
