All releases tagged Minor bugfixes
Release Notes: This version features portability changes in that it removes a blank line in Makefiles that confused some non-GNU make tools. It uses setvbuf with buffer size, and since not all systems support 0 for _IOLBF size, it now uses BUFSIZ. The binary RPM is now linked statically against GSL, fixing "undefined symbol: gsl_cdf_chisq_Q" errors on systems running libgsl < 1.4.
Release Notes: A parsing error, a segfault, and a defect in bogominitrain.pl's norepetition mode have been fixed. bogominitrain.pl's verbose output has been enhanced, the format for the -T option has been revised, and -T has been documented in the man page.
Release Notes: Some parsing errors were corrected, among other things ensuring interoperability with flex 2.5.4 and 2.5.31. A '-T' switch was added for terse mode. bogominitrain.pl was updated to version 1.3. Some revisions to the FAQ were made.
Release Notes: This release corrects an outdated acinclude.m4 file and contains a minor documentation update.
Release Notes: This release fixes a database lock contention issue discovered on Mac OS X, revises bogoutil's command line, and fixes AIX building using the xlc compiler.
Release Notes: Memory leaks have been plugged in this version. The -o option processing has been fixed. The FAQ has been reorganized and has information on Asian spam, the format of verbose output, and hints on using SpamAssassin to train bogofilter.
Release Notes: Some bugs were fixed with precedence for directory specifications, and processing of folded X-Bogosity header lines and the spam_subject_tag. Help messages have been cleaned up, and help messages and version info were added. The bogofilter-faq.html has been expanded.
Release Notes: fprintf() arguments have been cleaned up and cast properly. Message and MIME header checks have been cleaned up. The maintenance code has received improvements and the regression tests have been expanded (and will now try to use gdb to obtain a stack backtrace when a test fails.)
Release Notes: Locking now uses random sleep intervals in case of lock conflicts. Special characters have been removed from regression test data to improve portability. Unknown content types are no longer decoded, thus improving robustness.
Release Notes: This release adds several fixes and minor improvements.
