v0.1.7 20140403 * Introducing *gain* callback (preparing place for F::AF::S to move in). * *$uri* identifies transfer, but *$filename*. * Attempt to fix nix-but-linux FTBFSes (totally my fault). * Attempt to prevent losing diagnostics around non-empty *STDERR* (just in case). v0.1.6 20140318 * Transfer-is-in-progress check is moved to callback. What could be user-provided. Introducing *$ConfigData{tick}* and changing *$ConfigData{timeout}* handling. * Another attempt to fix got-''-expected-'{{{TERM}}}' build errors (guesswork). Another attempt to fix got-'1'-expected-'' build errors (being *@$log* filled with yet-to-be-processed message or empty isn't reliable anymore). * Adding unit with required manual configuration (slow inter-device copy). That's the very first torturing unit, please welcome. * Investigating weird compatibility problem (two units: t/0/EmMn8o.t and t/0/9raCtd.t). v0.1.5 20140226 * Trying to test minimally implemented File::AptFetch::Simple. * Attempt to fix got-''-expected-'{{{TERM}}}' build errors (guesswork). Attempt to fix got-'1'-expected-'' build errors (probably, in vein). Attempt to fix unexpected locale settings build errors. v0.1.4 20140209 F::AF::S pre-release. * *$ConfigData{apt_source}* parser has been rewriten: accepting new config entries, more strictures. * Incompatible change! Message parser has been rewritten: long promised changing hyphens to underscores, more strictures. * Pipe reader has been rewritten: runs on select(2), eval has been dropped; side effect -- all possible messages are read at once (not one-by-one, as before). Probable regression (may be there's none): what if a message is that big that it doesn't fit in buffer? * TS has been reloaded. Ridiculous "INIT { use_ok }" has been dropped from t/0.t. * Attempt to declare legal stuff correctly without dependecies boosting. (v0.1.5 -- nope.) v0.0.9 20101205 Mostly cosmetic change (one underscore added in _cache_configuration parsing re) in order to work with APT of squeeze. * Requiring perl of v.5.8.0 at minimum. * Playing with POD in diags has been plain stupid. * Misguided attempt to survive the die of "Interrupted system call" (btw, TS only, F::AF is still affected). * Misguided attempt to fix missing log of signaling tests. * Adding simplest test-unit that's run first and BAIL_OUTs for any reason. v0.0.8 20101125 Panic release -- methods have no will to report any progress. * *$ConfigData{timeout}* has been invented. Default timeout was bumped 12x times in order to recover from too silent APT-methods. * Hopefully fixing part of FAILs of cpan-testers. v0.0.7 20090511 * First release. v0.0.0 20090222 * Initialization.