Bug#649274: Forming a new upstream for timidity (and reporting various issues with current deb pkg)
Attached is a debian/ for a new package for 2.14. Package-wise there aren't many changes: * Update patches * Use xaw3d instead of xaw I've tested the regular package and the extra interfaces package. A regular user can build this using the following procedure: * Download the 2.14 source. Rename file to timidity_2.14.0.orig.tar.bz2 * Extract the source. Extract the attached debian/ directory under the new source directory. Make sure the orig file is in the parent directory of the source directory. * Run "dpkg-buildpackage". -- Yair timidity_2.14.0-1.debian.tar.xz Description: application/xz
Bug#649274: Forming a new upstream for timidity (and reporting various issues with current deb pkg)
Timidity has had a new upstream by now (2.14.0). See http://sourceforge.net/mailarchive/forum.php?thread_name=201206290019.AA00513%40tamuki.linet.gr.jpforum_name=timidity- talk and http://timidity.cvs.sourceforge.net/viewvc/timidity/timidity/ChangeLog?revision=1.704view=markup for the changelogs. Yours, Yair K.
Bug#649274: Forming a new upstream for timidity (and reporting various issues with current deb pkg) (fwd)
Hi, On 06/11/2012 02:13 AM, Geoffrey Thomas wrote: That rounds up the wrong (or so I believe) fixes in the Debian pkg. As said I hope to do a new upstream release soon, amongst a lot of bugfixes this will also include IPV6 support for the relevant bits of timidity. Thanks, and thanks for identifying these issues. I see there's been a bit of activity a week or two ago in upstream (and upstream looks to have most the patches originally submitted with this bug report), so I'll see if I can just make a new orig tarball based on upstream git and drop our patches entirely. Correct, I've been working together with some of the upstream maintainers (who have resumed work on timidity after some mails from me) on getting a new timidity release ready. All the Debian fixes (minus the ones I found to not make sense) + all Fedora fixes + all Suse fixes + many more fixes are currently in upstream CVS, a new release is scheduled for the end of this month (June 2012), so the best cause of action is probably to just wait for that release and switch to it. Regards, Hans -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#649274: Forming a new upstream for timidity (and reporting various issues with current deb pkg)
tags 649274 + pending thanks On Sat, 19 Nov 2011, Hans de Goede wrote in part: So about the issues I've found while reviewing all the changes the Debian pkgs make to the upstream tarbal: --- timidity-2.13.2.orig/timidity/reverb.c +++ timidity-2.13.2/timidity/reverb.c @@ -1624,8 +1630,8 @@ static void do_ch_reverb_panning_delay(i buf[i] += r; buf[++i] += l; - if (++index0 == buf_size) {index0 = 0;} - if (++buf_index == buf_size) {buf_index = 0;} + if (index0++ == buf_size) {index0 = 0;} + if (buf_index++ == buf_size) {buf_index = 0;} } memset(reverb_effect_buffer, 0, sizeof(int32) * count); info-index[0] = index0; This changes pre-increment to post increment, *at the end of a loop*, causing index0 (and buf_index) to be ranged as: 0 = index0 = buf_index which should of course be: 0 = index0 buf_index So I believe this Debian patch to be wrong, and the original code to be correct. Having looked at the code, agreed. I'll remove this in the next upload. --- timidity-2.13.2.orig/timidity/playmidi.c +++ timidity-2.13.2/timidity/playmidi.c @@ -7980,7 +7980,7 @@ int play_event(MidiEvent *ev) channel[ch].temper_type = current_event-a; ctl_mode_event(CTLE_TEMPER_TYPE, 1, ch, channel[ch].temper_type if (temper_type_mute) { - if (temper_type_mute 1 current_event-a + if ((temper_type_mute 1 current_event-a) - ((current_event-a = 0x40) ? 0x3c : SET_CHANNELMASK(channel_mute, ch); ctl_mode_event(CTLE_MUTE, 1, ch, 1); @@ -8003,7 +8003,7 @@ int play_event(MidiEvent *ev) ctl_mode_event(CTLE_TEMPER_TYPE, 1, i, channel[i].tempe } if (temper_type_mute) { - if (temper_type_mute 1 current_event-a + if ((temper_type_mute 1 current_event-a) - ((current_event-a = 0x40) ? 0x3c : FILL_CHANNELMASK(channel_mute); for (i = 0; i MAX_CHANNELS; i++) I believe this patch is meant to fix a compiler warning, unfortunately the patch does more then that, it actually changes the meaning of the code. The original code calculates how much to shift the 1 by before masking by seeing if a = 0x40 and in that case shifting by (a - 0x3c), otherwise it shifts by just a (actually (a - 0), but that is the same). The precedence rules relevant here (for the original code) are, first do the - op, then the and then . The added () change the order in which this operations are done. Given that temper_type_mute is strictly a bitfield, doing a - operation on it makes no sense, and I believe the old code is correct. Note that my patchset has what I believe is a correct fix for the compiler warning in patch 0010-Fix-silence-various-compiler-warnings.patch Also agreed, and also will revert. This code looks pretty funny, but it makes slightly more sense in the context of the -Q manpage options: On the other hand, to put `t' character after -Q option or to use --temper-mute describes temperament mute. This mutes channels of specific temperament type n. For preset temperament, n can range 0 to 3. For user-defined temperament, n can range 4 to 7. So the target cases are a = 0 through 3, or a = 0x40 through 0x43. +++ timidity-2.13.2/timidity/instrum.c @@ -1070,7 +1071,7 @@ Instrument *load_instrument(int dr, int } /* panning */ if (ip != NULL bank-tone[prog].pan != -1) { - pan = ((int) bank-tone[prog].pan 0x7f) - 64; + pan = ((int) bank-tone[prog].pan 0x7f); /* - 64 */; for (i = 0; i ip-samples; i++) { panning = (int) ip-sample[i].panning + pan; panning = (panning 0) ? 0 This is (I believe) a fix for Debian bug 536252: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=536252 See the original bug for the rational of the fix. I believe this rationale is wrong, as the code here is adding 2 pans together, so it is correct to change one from 0-127 to -64 - 63 to get the desired effect. Before looking at the Debian changes I spend an entire day tracking down what I believe is the real cause for Debian bug 536252, after a similar issue was reported in Fedora bug 710927: https://bugzilla.redhat.com/show_bug.cgi?id=710927 What is going on (and a workaround-ish way to fix it) can be found in 0009-sndfont-Work-around-soundfonts-with-missing-links-be.patch I agree that the Debian patch looks wrong and will revert this too. I've looked briefly at the RHBZ report and patch 0009, and
Bug#649274: Forming a new upstream for timidity (and reporting various issues with current deb pkg)
Hello! The timidity maintainer on sf.net is Urabe Shyouhei (the given name in the later, I believe). It's trivial to find his other e-mail addresses. I also found and contacted him on twitter: https://twitter.com/shyouhei -- Regards, Pavel Roskin -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#649274: Forming a new upstream for timidity (and reporting various issues with current deb pkg)
Hi, On 11/19/2011 09:59 PM, Geoffrey Thomas wrote: On Sat, 19 Nov 2011, Neil Williams wrote: CC'ing the only person to express some interest. Geoffrey, if you are no longer interested in timidity, despite signs of interest from a possible new upstream, please retitle #585039 as O: instead of ITA: Thanks for Ccing me. The real world had caught up with me a bit this semester, both in terms of taking away time for non-academic technical work and for writing music, but I'm graduating soon and done with interviews, so I have more time now (and have a laptop running testing) and do intend to maintain it. I'd given my sponsor a debdiff to adopt the package and fix the FTBFS. I think we'd just both forgotten about it -- I've just poked him about uploading it. Hans, thanks for the patches and review, and I'll try to take a look at them over the next few days. I'd definitely be interested in contributing to a revived upstream -- yes, timidity does need some love. :) That is great to hear! Let me know if you've any questions. Note that my patches are based on the current CVS code, not on the latest tarbal! I'm afraid that I'm not getting anywhere so far with my attempts to take over the current sf.net project, my mail to all the developers and the single admin it has now has bounced for at least the admin. I'll try to contact the sf.net staff, but AFAIK they no longer allow taking over projects without help from one of the former admins. Regards, Hans -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#649274: Forming a new upstream for timidity (and reporting various issues with current deb pkg)
On Sat, 19 Nov 2011 15:05:00 +0100 Hans de Goede hdego...@redhat.com wrote: This is a bit unusual bug-report I'm afraid normally I would've send this as an email to the Debian package maintainer of timidity, but it seems that timidity is currently orphaned in Debian :| There has been no interest from anyone wanting to take over Debian maintenance of timidity since the previous maintainer orphaned it. It now has a release critical bug because the current source fails to build. The bug has been open for 3 months and is sufficient reason to remove timidity from Debian today. As you've expressed some interest, I'm willing to not file the removal bug right now but that doesn't exclude someone else doing precisely that. Given the status of the package in Debian, it is more likely that timidity will be removed rather than updated. Maybe once there is a functioning upstream and a new upstream release, someone may reintroduce the package into Debian. CC'ing the only person to express some interest. Geoffrey, if you are no longer interested in timidity, despite signs of interest from a possible new upstream, please retitle #585039 as O: instead of ITA: Hans: have you any clues about the pulseaudio issues? Joost Yervante Damad comments in the bug report orphaning timidity: If you want to take over maintenance, be prepared to deal with obscure pulseaudio issues. #585039 I think it would also be a very good idea, Hans, if you put a short message on bug #585039 about your interest in a new, fixed, upstream release as this will be one of the places people will look before seeking removal of timidity. That said, interest from upstream is not of itself going to stop removal from Debian. The reason I'm sending this mail is because one of the Fedora packages I (co)maintain is timidity. Recently we got a number of bugreports related to timidity, and one of the conclusions was that timidity needs some love. I sympathise, I've felt the same about other packages and gone into the cycle of getting the SourceForge project re-assigned, porting the code to current libraries and systems, only to find that the codebase really cannot sustain a second transition or some dependency simply becomes abandonware. The workload can gradually become unsustainable and sometimes it's simpler to just accept that the package has had too much bit rot already and it would be easier to drop it. I also went through all the changes in the Debian package, and were relevant have added those too. Note that I deliberately did not include a few of the changes from Debian, as I believe they are wrong! See below for details. As the potential new upstream, you are welcome to make that decision. It's better for Debian if there is just a new upstream release and then someone with sufficient interest (not me!) can look at what might need to be done to bring the Debian package up to date. Sadly, it is more likely that timidity will have to be removed and then, possibly, reintroduced if (and only if) someone reading this message gets sufficiently motivated to work on timidity in Debian. So now I've a nice and polished version of timidity, and given that the latest official release has been 6 years ago I think it would be good to do a new official release, hence I've contacted the current admin and developers of the sf.net timidity project, hopefully they will allow me to take over the sf.net project, I would have loved to work together with the Debian maintainer on forming a new upstream, but alas. When there is a new release available via SF or some other site, please update bugs #649274 and #585039. (Don't feel obliged to keep with SF but generally I've found them supportive when someone offers to adopt an abandoned project). I believe this patch is meant to fix a compiler warning, unfortunately the patch does more then that, it actually changes the meaning of the code. ... as long as the package now builds on current Debian unstable (make distcheck using gcc-4.6 with binutils-gold on any current GNU/Linux distro will be a good test)... Sorry I cannot comment on the patches themselves, I don't care about timidity - my only concern is that broken abandoned packages in Debian either find new maintainers or get removed. Before looking at the Debian changes I spend an entire day tracking down what I believe is the real cause for Debian bug 536252, after a similar issue was reported in Fedora bug 710927: https://bugzilla.redhat.com/show_bug.cgi?id=710927 Be assured that your effort is respected but unless someone steps up to maintain your work in Debian, timidity *will* be removed. Now that timidity is on my radar, I'll keep an eye on it. If nobody responds to #649274 or #585039 or fixes the fail to build bug #639196 then I will ask for removal of timidity from Debian unstable and testing in ~ 10 days. As said I hope to do a new upstream release soon, amongst a lot of bugfixes this will also include IPV6 support
Bug#649274: Forming a new upstream for timidity (and reporting various issues with current deb pkg)
Hi, On 11/19/2011 04:39 PM, Neil Williams wrote: On Sat, 19 Nov 2011 15:05:00 +0100 Hans de Goedehdego...@redhat.com wrote: This is a bit unusual bug-report I'm afraid normally I would've send this as an email to the Debian package maintainer of timidity, but it seems that timidity is currently orphaned in Debian :| There has been no interest from anyone wanting to take over Debian maintenance of timidity since the previous maintainer orphaned it. It now has a release critical bug because the current source fails to build. The bug has been open for 3 months and is sufficient reason to remove timidity from Debian today. As you've expressed some interest, I'm willing to not file the removal bug right now but that doesn't exclude someone else doing precisely that. In case this much was not clear already note that my interest does not extend to actually maintaining the Debian package myself. Given the status of the package in Debian, it is more likely that timidity will be removed rather than updated. Understood, I'm just trying to inform anyone who may pick up timidity for Debian about my work, so that they don't start from scratch. Maybe once there is a functioning upstream and a new upstream release, someone may reintroduce the package into Debian. CC'ing the only person to express some interest. Geoffrey, if you are no longer interested in timidity, despite signs of interest from a possible new upstream, please retitle #585039 as O: instead of ITA: Hans: have you any clues about the pulseaudio issues? Joost Yervante Damad comments in the bug report orphaning timidity: If you want to take over maintenance, be prepared to deal with obscure pulseaudio issues. #585039 I know of no pulseaudio issues, but note that in Fedora, the default timidity config will first try to use pulseaudio through libao's pulse support, before trying alsa (which will end up using pulse through an alsa plugin) I know that I'm the one who made that change, it could very well be that I did that because when timidity uses alsa directly it does so in a way which is not compatible with pulse. Note that we've had no issues with timidity which are pulse related for a long time know, so I would not worry about this. I think it would also be a very good idea, Hans, if you put a short message on bug #585039 about your interest in a new, fixed, upstream release as this will be one of the places people will look before seeking removal of timidity. That said, interest from upstream is not of itself going to stop removal from Debian. The reason I'm sending this mail is because one of the Fedora packages I (co)maintain is timidity. Recently we got a number of bugreports related to timidity, and one of the conclusions was that timidity needs some love. I sympathise, I've felt the same about other packages and gone into the cycle of getting the SourceForge project re-assigned, porting the code to current libraries and systems, only to find that the codebase really cannot sustain a second transition or some dependency simply becomes abandonware. The workload can gradually become unsustainable and sometimes it's simpler to just accept that the package has had too much bit rot already and it would be easier to drop it. I hear you loud and clear (similar experiences on my side), but in the case of timidity I think keeping it alive is important, because it is one of the few software wavetable midi synths we have (and as such has unfortunately been forked into SDL_mixer, libtimidity and too many others). I also went through all the changes in the Debian package, and were relevant have added those too. Note that I deliberately did not include a few of the changes from Debian, as I believe they are wrong! See below for details. As the potential new upstream, you are welcome to make that decision. It's better for Debian if there is just a new upstream release and then someone with sufficient interest (not me!) can look at what might need to be done to bring the Debian package up to date. Ok, note though that I'm not all too familiar with timidity internals myself, hence my attempts to explain my decisions, so that others can inspect them if they want to. Sadly, it is more likely that timidity will have to be removed and then, possibly, reintroduced if (and only if) someone reading this message gets sufficiently motivated to work on timidity in Debian. Understood. So now I've a nice and polished version of timidity, and given that the latest official release has been 6 years ago I think it would be good to do a new official release, hence I've contacted the current admin and developers of the sf.net timidity project, hopefully they will allow me to take over the sf.net project, I would have loved to work together with the Debian maintainer on forming a new upstream, but alas. When there is a new release available via SF or some other site, please update bugs #649274 and #585039. (Don't feel obliged to keep with
Bug#649274: Forming a new upstream for timidity (and reporting various issues with current deb pkg)
If timidity doesn't get IPv6 support soon ? I don't see IPv6 as important. It's a major maintenance burden (a hack to firewall, configure, ...), IPSs use it to dominate ISP sales, and so far no one claims to have a final spec on it. when is IPv4 over IPv6 a sin? are bsd sockets a sin (they are old)? and html (pdf/flash is better)? personally i don't like hearing i added code but broke 1,000 other packages. just commenting on improvements that don't state they bump major version but do worse anyhow. not that you all did it :) us all i'd guess. Thanks, John Hendrickson p.s. (sorry i can usually fix code but cannot say I'm situated to maintain at this time) -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#649274: Forming a new upstream for timidity (and reporting various issues with current deb pkg)
On Sat, 19 Nov 2011, Neil Williams wrote: CC'ing the only person to express some interest. Geoffrey, if you are no longer interested in timidity, despite signs of interest from a possible new upstream, please retitle #585039 as O: instead of ITA: Thanks for Ccing me. The real world had caught up with me a bit this semester, both in terms of taking away time for non-academic technical work and for writing music, but I'm graduating soon and done with interviews, so I have more time now (and have a laptop running testing) and do intend to maintain it. I'd given my sponsor a debdiff to adopt the package and fix the FTBFS. I think we'd just both forgotten about it -- I've just poked him about uploading it. Hans, thanks for the patches and review, and I'll try to take a look at them over the next few days. I'd definitely be interested in contributing to a revived upstream -- yes, timidity does need some love. :) -- Geoffrey Thomas http://ldpreload.com geo...@ldpreload.com -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org