Fixed.
It turns out that bug 7302 exists and I misinterpreted a response to the
immediate case as an actual fix for the underlying flaw. I have now
committed a real fix.
On 20 Dec 2018, at 16:14, Kevin A. McGrail wrote:
+1
--
Kevin A. McGrail
VP Fundraising, Apache Software Foundation
Chair Emeritus Apache SpamAssassin Project
https://www.linkedin.com/in/kmcgrail - 703.798.0171
On Thu, Dec 20, 2018 at 4:04 PM Bill Cole <
sa-bugz-20080...@billmail.scconsult.com> wrote:
On 20 Dec 2018, at 15:53, Bill Cole wrote:
On 20 Dec 2018, at 14:25, Bill Cole wrote:
This caused a failure in the run_nightly script which apparently
doesn't prevent a broken update from being built and distributed.
I'm looking at run_nightly...
It's a problem in the 'make' step: a lint failure emits a nice
obvious
error message:
ERROR: LINT FAILED, suppressing output: rules/72_active.cf
But then the make succeeds. Seems like a poor choice...
Still working...
It's a build/mkrules problem:
if (lint_rule_text($text) != 0) {
warn "\nERROR: LINT FAILED, suppressing output: $file\n\n";
# don't suppress entirely, otherwise 'make
distcheck'/'disttest'
# will fail since the MANIFEST-listed output files will be
# empty.
# delete $output_file_text->{$file};
$output_file_text->{$file} = '';
}
Seems like a poor choice. IMHO make dist* *SHOULD* fail if the rules
can't pass lint.