This is an automated email from the git hooks/post-receive script. fsfs pushed a change to annotated tag upstream/1.2206 in repository libnet-imap-simple-perl.
at 375024e (tag) tagging b699b6da8dccf66407cf35ea6661aacfc555fcce (commit) replaces upstream/1.2205 tagged by Florian Schlichting on Tue Dec 1 21:36:29 2015 +0100 - Log ----------------------------------------------------------------- Upstream version 1.2206 Alan Griffiths (1): Cannot set retry value to zero in constructor (https://rt.cpan.org/Ticket/Display.html?id=68460) Alex Vandiver (4): Fix a use of _escape in void context Escape the mailbox name to APPEND Add a way to close SELECT'd or EXAMINE'd maiboxes Add a ->noop method, for sending NOOP commands ChinaXing(ιδΊζ) (1): fix bug in "find_ssl_defaults" Drav (1): Make the end parentheses match more explicit, debug when enabled with a better error message Florian Schlichting (1): Imported Upstream version 1.2206 Holger Glaess (1): add method separator JIK (1): https://rt.cpan.org/Ticket/Display.html?id=34941 Jason Woodard (1): sorting within range2ist breaks IMAP sorting via search Jason Woodward (7): Add the ability to sort via search extended body_field_param production and fixed the body_fields definition Added part numbers to the data from body_summary Extended get() to support fetching a message part added SimpleX pod tidy up body summary data into BodySummary and Body objects updated SimpleX pod Jonathan Kamens (1): JIK's patch attached to RT#33925 Jorge N. via RT (1): uidsearch+sort does not return uid's but msgno's MSTEVENS (1): PATCH FOR https://rt.cpan.org/Ticket/Display.html?id=50130 Michael Stevens (1): proposed diff that makes podchecker happier Paul Miller (486): the original 1.17 release my first unseen() patch my msg_flags and initial put patches corion's starttls patch warning about a huseless somethin hash, blah blah flags bug patch moving a few things around die rcs die I pulled these from pjf's IPC::System::Simple I stripped the pods off Simple.pm, now to strip the code from Simple.pod... stripped the code off the pod critic patches This is a combination of critic suggestions and the disabling of some Corion debugphilic code. fixed a few pod problems and word-wrapped a bit fixed introduced bug made the pod_coverae pass worked on the README a bit my todolist, devtodo woops, screwed up the manifests forgot to skip itself test is born unto me made some progress on the test file and instructed skips for the makefile more birthing of teh imap server test MANIFEST the certs this is about as close as I can get without building auth and model classes auth and model skeletons catch sigchild during startup fixed startup made Net::IMAP::Simple::SSL totally un-necessary, if it ever was documented the use_ssl flag and fixed up a bunch of POD stuff and mispellings made the use_ssl flag more robust without requring IO::Socket::SSL at configure time altered whitespace this was pretty confusing in the debug log made the debugger print less superfluous info have to decide to use the funny port before we select the port note changes version up fail tests unless select works. make 0 message 0, but true fail loudly, should we do this everywhere? you need to tell perlcritic to be critical... I didn't know that. after making the critic critical, some changes were required the logs show this put() is working, ... o rly? where did the mail go? little typo woops, should be require, not use\! working on getting the tests to go, found some minor bugs I think when this test does more stuff I'll release as 1.19 point out bug queue this is my little login function wow, the manifest was really jacked, wonder when I fat fingered that bugs hrm, I think I fixed the oldest bug on rt ... a little note to myself I borked ->last k, this stuff seems to be working right document the bug I created, and fixed skip moved the tests around a little manifest changes making the server start-up, run tests, server down at end make it easier to run one test at a time different mailbox this is telling dividing up the tests sucks moved things around a bit to verify something found a new way to prevent DRY in the tests I don't understand why this exit-value is non-zero, but I'm blaming threads. the final test might be something like this, ... there, now the test can start and stop itself, without using signals... new release plan _process_cmd -- should always return a count or undef, even in list context going through all the process_cmd calls ... Revert "_process_cmd -- should always return a count or undef, even in list context" make the imap server startup noisy-to-the-right EXAMINE seems easy enough to support... depends on what the servers do when you try to put or change the flags after the command EXAMINE command getting ready to test the new() 4-connection ticket on rt well, I don't think Simple can detect *ANY* disconnects at all, much less in new() make it easier to run the server for a while... some startup scripts made this startup and shutdown better testing a few things by hand made the connectalot script do more stuff you can't reliably tell if a socket is closed until you try to read from it ... it either blocks until data arrives or returns 0 finally, man, that was hard to figure out. closing 30229 started _04 critic, was critical mmm, all tests failed everywhere because Net::TCP isn't core, duh, IO::Socket::INET is though this doesn't work... I have some ideas why, but it's time to go skydiving the reselect fires now, but something is still wrong. might have found a bug in Net::IMAP::Server, must look into it further. I had broken t/07 when I added things in t/15 this should close out 45953... I'll prolly write more tests for it though note changes undeleted seems like a natural addition more flag tests bump version for next tickets working on 33189 a little working on 33189 a little 33189 info attachment oops make these more configurable of course, if I'm making it configurable, it should use the configs ws changed the way debugging works on these guys status precursor superfluous STATUS command 755 status app manifest the contribs this will become the status test the working box shouldn't be escaped like that added a current_mailbox test haha, bug I introduced a bug when I unescaped the {working_box}, but then thought about it and taught _escape() to not modify the given argument bug these numbers are't matching up, but they should moving the t7lib:: to t:: so pause won't index it move these to inc so they don't get index manifest file moves described latest patches temporary fix for Net::IMAP::Server::Mailbox::unseen() v1.18 [reported] ridiculous bug ridiculous bug #2 missed this in the manifest somehow make my t/17 actually do what I wanted it to do 3.5 hours ago critic hated my naked comma evaluation, I think it was as clear improved docs for seen and unseen add my 33189 notes to the dist temporarily make the debug logger a little smarter deal with errors on msg_flags better, and use msg_flags from seen and deleted expand that test slightly documented the new flag error code documented my thinking for closing 33189 require Net::IMAP::Server if we're under AUTOMATED_TESTING this aught to test correctly trying to get expunge to be a little more informative describe latest changes I'm nolonger chaining ->_reselect, so some of the tests needed fixing note changes test the scalar mode this should work fine except that there aren't that many messages ... ok, this tests almost all the delete stuff I can think of (not counting various permutations of failures and successes) note about delete working correctly. I misunderstood the rfc 6.4.3 example copy tests altered TODOs moved the tests around a bit... I can't tell if there's a copybug in Net::IMAP::Server or not... sequence sets work just fine documented RT#40203's concerns my message body precursor I changed the way get() works. I'm pretty much going to have to undo it tomorrow though -- it'll bork everyone's code comprimise between $message = $imap->get(1) string and arrayref (do both) printing 40 spaces will prolly fail more than just using console codes... prove that our _process_cmd is a failure catch the line even on misbehaving servers that don't CRLF note the application of the patch lose 33189 txts couple little pod tweaks 1.1900 explicitly include all modules needed for the tests note changes this is a bug, woops include static copy of net-imap-server-1.20 going to release this failed to actually use the inc/ copy of Net::IMAP::Server, coder == fail optionally disable the inc copy make sure to show what version of the imap server we loaded BP's 1.22 with patch 1.22 n-i-server included trying to deal with test failures from the cpan-testers... Class::Accessor missing, weird ignore this stuff (updates) my test script my dovecot-pipe-server-thingy -- OK -> PREAUTH fix for https://rt.cpan.org/Ticket/Display.html?id=48166 versionup really weak search command misunderstood how the search result comes back I think we can do better, noted goal implementation mention the need for tests for top(), but don't actually do work accidentally killed manifest, fixed can't really publish this without docs. :) we can get a lot less set errors if we just mark these lines we don't parse as boring fix for perl critic filling out search a bit more likely, a lot of these functions work... failed to quote quoted strings I misunderstood the date comments in the RFC3501 search specification added pod added pod these really need "spaces" this I had jacked this up pretty bad -- git bisect ftw this is a quick cut and paste from t/15_flags.t this isn't what's advertised in the pod got the search tests online, not very thorough, but better than nothing version bump forgot to manifest the new test this is actually necessary for testing, didn't know that I can't figure out why this test is failing ... dev-version up for testing don't dork up the tester reports I really just want to test 22 with my tester, separating these make this ridiculously verbose 1.1910_2 ... wtf is going on there is no bug, there is only zuul I think this is what he's asking about this is a google bug, not an net-imap-simple one test for ticket #55552 I somewhat recall changing _unescape, oddly enough, but I don't think I fixed any of the actual uses of it. :) version up for release to fix #55552 doug reed patch version up to 1912 note changes merge with a couple more alexmv "monkey" patches heh, he had a copule more; version up again shhhh use 1.27 note changes and version up this effectively disables the whole line, meh I suspect this is not how alexmv will do it version up make a formal server log pull in a7f5e24 from Net::IMAP::Server make SimpleX optional This is woodwardj's first draft. use alester's license section change the bugs message note that I may not have permission to change the license Merge branch 'master' into SimpleX confusingly, manifest and not manifest simplex provide a way to use simplex from inc/slurp make inc build smarter fix package name this is goign to be fun maybe ignore the new test file add 2010 trailing whitespace ws pull the pod into the blibs track this, don't track that tweaked the synopsis slightly. I found I had to go check the source to see how to invoke simplex, so I described that quick at the top oops, wasn't checking return value correctly for local builds, it's convenient to have some unlink and default powers skel a body summary test heh, gotta ->new() the right package though I didn't like the way the body_summary changes shape so dramatically between multipart mixed messages and single part messages SINGLE part message works great multipart test Merge branch 'SimpleX' version up(s) I find myself editing the wrong ones, this'll make vim complain content_type vs hasparts make make rebuild the simplex stuff automatically I find myself editing the wrong ones, this'll make vim complain make make rebuild the simplex stuff automatically fixed test Merge branch 'SimpleX' log changes new invocation notes I disabled more perl critic stuff rather than do some of the more irritating things it suggested holy smokes! hooray! modification of read only variable how? oh, that's right. This bug still persists -- pfft more debug options critic "bug" document and then strip the )\r\n irritating %d-%m-%Y -> %d %b %Y don't select every search, only when we need to it seems the - is FWS and that's what google expects... most people seem to use the -, still %b instead of %m though say the right things here I don't think last() is quite what I want... I'm playing with this idea provide uid information caveat the star isn't a wild card, it means biggest test uid stuff [and fix bug in Search filters] critic ; didn't even know this was a magic variable... I use $< for that this is private, mark it that way I wonder if this will make more bugs than it creates... I doubt it note it skelled my ideas I'm going to write a grammar for this senseless renaming bug regex abuse detected, but used correctly partially working ahh, there we go shhh Use the grammar please move fetch to sX officially and return something sane I think this is coming along, but it still doesn't work for some reason caps talk about why we failed... NIL? (test) ... fail. I had a lot of fail in this gammar, I think I fixed a lot of it ok, I think I fixed most things learn to count getting my doc started this return style is immediately useful and doesn't require post processing actually, it still wasn't quite what I thought. Fixed star is allowed man, this is just really hard to get right I clearly don't understand the scoring, because this was crashing out another part of the grammar well, if the score won't work, reorder these so the body thing matches first also, don't let atoms have [] in them ignore this stuff (updates) took changing the logs around a little to realize Net-IMAP-Server gives more than one response per message. Add reverse() to throw away alternate responses there, this tests well, I like it the x00 type is too wierd to bother people with don't require selecting first document my damage version up shh maybe more behavior's later logout method this might work if it didn't sagfault ... going to look for a P::RD official way to do this. There msut be lexer access or something like it. this is like one million times better sexy fix for the .{32766} RE limitation, in fact, this was better all along note changes copyuid, and prep for launch going to make fake sockets that are actually pipes (eg ssh host dovecot) seq() is the opposite of uid() list2range uid fetch test uidfetch test list2range (found bug already) show that we can't correclty parse the tag 'BODYSTRUCTURE' ws only match BODY when it's not the middle of some other word. cover naked method manifest conflict seemed easy to resolve notes I did this at work, it's probably the right start typo just checking this is worth doing I think this is going to be neat... change the test server port(s) uidsearch code, needs docs doc for uidserach cool, trivial test expand changelog note slightly pretty simple, but useful use the sequencer typo hrm, I was fixing something that wasn't initially broken... oh wells this space may make a difference on some imap servers... doubtful, but it could happen merge was awful... meh this clearly needs a lotta work, but it'll do this is the most awesome thing in the world fixed the err loop might as well support the Simple.pm version up for the pipe mostly fixed this... things aren't so good a little though just trying to figure out where we're freezing up (it's in the READLINE) this all works fine, except that it doesn't lawl, blocking forever even better note changes added a test document ps critic skip all of the functions in the pipe socket test and fix for https://rt.cpan.org/Ticket/Display.html?id=60537 this is where we snag the callback use the callback, if givin testing the readline callbacks document readline callbacks [1::2]:3 ipv6 patch bug bug possible fix for https://rt.cpan.org/Ticket/Display.html?id=62486 maybe like this.. note changes this is why my tests don't work β http://rt.cpan.org/Ticket/Attachment/828926/429411/ the patch wouldn't apply for me, I made this instead version up for release don't manifest this I don't want to maintain these anymore subtle doc buc noticed by aka HM 2k another subtle doc buc noticed by aka HM 2k another subtle doc buc noticed by aka HM 2k another subtle doc buc noticed by aka HM 2k construct test for https://rt.cpan.org/Ticket/Display.html?id=63282 I'm not sure that test would have functioned correctly this appears to fix https://rt.cpan.org/Ticket/Display.html?id=63282 version up https://rt.cpan.org/Ticket/Display.html?id=65526 http://goo.gl/FUQPn removed the die() comical little typo that 22 version up MYMETA is news to me version up for nosort patch yeah, these make rules screw up dmake (my strawberry seems to use nmake, whatever) and aren't really even used ... version up skel test that can hopefully reproduce https://rt.cpan.org/Ticket/Display.html?id=73431 yeah, shows the error nicely version up for fix crap, this works fine this could potentially fix the problem, but I still don't definitely know what it is/was ahh, here we go. This is the error. It is spurious ok, this almost ceraintly deals with https://rt.cpan.org/Ticket/Display.html?id=73431; but I introduced a bug that space wasn't whitespace, it was message body oops; if they don't have some of the test modules, this will produce three failures the real version number doesn't have that one Merge pull request #1 from kakadoo/master Merge remote-tracking branch 'github/master' version up document sep method version up this works... I feel like I should do better note changes and version up this is my horrible idea to make sure tests run in order... The one second wait per test is the best part note this should fix most of the test failures note add support for specific ssl versions note change version up docoment new setting huh, https://rt.cpan.org/Public/Bug/Display.html?id=78539 is true minor password bug affecting people who have backslashes and quotes in their passwords add IO::Socket to test requires disabled for 07 don't make racket about this not willing to maintain this shit anymore these too, gone don't need any of that shit anymore either. :p some out, some in fixed one test... 85 more to go nothing works at all, hurray there we go ugh k k couple more dovecot not done with this one... might reverse all this shit though there, this should help release a dev built fuck it ta this breaks all tests, naturally, but should probably still be added this breaks all local-server tests, and therefore fits nicely in this branch I need to handle this in some rational way ssl patches various merge problems. need to actually fix the tests now... various t/15 puts do not give the expected unseen and recent same but better imo made a lot of progress there docoment the new ssl... considering a warning about certificates when not checking with the CA cert database found another bug recent doesn't work on dovecot like I expected. If *I* put in my *current* session then the message isn't \recent like it is in Net::IMAP::Server. Not too shocking. We just remove all recent tests and move on there, another one bites the dust dovecot recent removed, test fixed another one down hrm, let's just make some mailboxes there we go finish rebuilding tests for new actual server environment cleaning up some stuff note explain hidden tests minor version change for no real reason, supz cpan bug new debug function yeah, ok: https://rt.cpan.org/Public/Bug/Display.html?id=89195 note attempt to fix https://rt.cpan.org/Public/Bug/Display.html?id=89296 ... doesn't actually address problem he's having in the ticket. further testing version up for spurious bug? awaiting info I fail to see how this wasn't a problem before fixed Merge pull request #3 from dravster/patch-1 partial reversion to reverse introduced bugs disabled errors Merge pull request #4 from hoelzro/master version up bug Rob Hoelz (1): Allow port to be set via params to new gregor herrmann (1): Spelling glitch holborn (1): the holborn patch for the count-type response from cyrus ----------------------------------------------------------------------- No new revisions were added by this update. -- Alioth's /usr/local/bin/git-commit-notice on /srv/git.debian.org/git/pkg-perl/packages/libnet-imap-simple-perl.git _______________________________________________ Pkg-perl-cvs-commits mailing list Pkg-perl-cvs-commits@lists.alioth.debian.org http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-perl-cvs-commits