Release Checklist
These are all for 1.4.0 release. There has been a 1.3.4 Windows/Linux beta release, and we are planning a 1.3.5 release on May 7, 2008. There should be a 1.3.6 release at the end of that month before GSoC students start coding. After that there will be frequent internal Beta releases during GSoC, but a final public 1.3.x as a 1.4.0 release candidate will not occur until after GSoC.
To take ownership of an issue, add your initials before it like so:
- [DM] for Dominic Mazzoni
- [LL] for Leland Lucius
- [JC] for James Crook
- etc...
This is a companion list to the Documentation Checklist.
|
Wording, where we track suggested changes to the exact words that appear on a dialogue
Pending website changes where we post suggested changes to our audacity.sourceforge.net and audacityteam.org sites (except the Wiki).
Unless indicated by discussion against particular items, these should be regarded as needing attention. When processed, please move them to the "Done" section of those pages.
- Patches are added and tracked on the Sourceforge Patch Tracker .
- There is also a page here where we discuss ideas for developing both this Wiki and the Manual Wiki, and monitor their progress. This includes security issues, bug fixes and enhanced functionality through installing additional extensions.
Contents
Essential fixes
Fixes are the top category, since it is difficult to estimate how long bugs will take to fix. See a discussion of what constitutes essentials on the discussion page.
Any items marked $ are believed fixed, awaiting testing.
Essential changes
These are other changes to source code that are deemed essential. We will do all of these for final 1.3.x release candidate or 1.4.0 release as noted below.
- Help/Manual issues. These are needed for when we release the final release candidate Beta:
- Script to extract html help from the Manual Wiki.
- Installer to be modified to also install Html help. Consider giving Windows installer an option to install help for all users, not just current user.
- Address documentation issues, including website updates.
- For 1.4, "completing the Manual". Needed before we can release any versions of 1.4.
- For 1.4, finalise the default additional plug-ins provided with Audacity and make them available on all platforms. No special action is needed on Linux for Nyquist plug-ins, as any *.ny plug-ins from the plug-ins dir are installed, except those deleted when the tarball is created using scripts/maketarball.sh. This should be the contents of the Audacity Plug-Ins folder after a fresh installation on Windows or Mac:
- beat.ny
- clicktrack.ny
- crossfadein.ny
- crossfadeout.ny
- delay.ny
- equalabel.ny
- GVerb.dll ++
- Hard Limiter.dll ++
- highpass.ny
- lowpass.ny
- pluck.ny
- rissetdrum.ny
- sc4.dll ++
- SilenceMarker.ny
- tremolo.ny
- vocalremover.ny
- vocoder.ny
- (Under discussion: should we add a chorus plug-in? We have a Nyquist flanger/chorus plug-in available for review, or should we port the simpler multivoice_chorus_1201.dll from the Steve Harris LADSPA set to Win and Mac?)
- Note:
- ++ These are not in source code but VJ and LL normally include them in Windows and Mac release builds
Essential fixes: Platform Specific
Mac
- Fix: Mac CoreAudio crash when input device's sample rate is 44100 and the output device's sample rate is 48000.
- $ Problem with inability to launch Audacity on non-admin accounts on Network Home ( see here) and on managed accounts - see here .
- BP: The problem with the 1.3.3 build is that it was built with permissions of rwxr-x--x. Therefore, anyone not in the group staff (non-admin accounts) would not have the ability to read the file (ie: could not launch it). CVS HEAD sets permissions correctly, and should solve this. Does it solve the Network Home issue? The link above suggests CHMOD 755 does not solve it?
- Test MJS's fix for: Fix crash on restart when noisegate.nrp is present (the nrp only seems to be written when in Clean Speech mode). Crashing on launch replicable in 1.3.3 on Windows and Linux when CleanSpeech enabled and .nrp present. Possibly related crashes reported on Mac when starting Noise Removal (1.3.2 on Windows crashes in this way - not on launch - when in CleanSpeech mode with nrp file present). See here for more details.
- MJS's fix seems to have worked to stop crashing on Windows and Linux. Still needs testing in a new OS X build. NRP file has its own directory in the user's folder for application data - need to make sure it actually gets written there. NRP file still only written for CleanSpeech mode at present
- The old issue of envelope points not restoring properly when reopening never quite seems to go away on Macs -see here . This report says the bug was found in 1.3.3 on a Mac OSX 10.4.10 on a PowerBook G4, July 15, 2007. MJS made some fixes for Windows. Perhaps they help here too?
- BP: Unable to reproduce on 1.3.3 - 10.5.2; Head - 10.5.2; 1.3.3 - 10.4.11, Intel iMac on 28 Mar 08.
- BP: Reproduced on 1.3.3 - 10.5.1, PPC iBook G4 on 28 Mar 08.
Mac *and* Linux
- [LL] Fix: Random bits of silence during recording (non-Jack version of this problem). [Mac only?]
- Progress: Udo tried smaller blocksize in linux/jack version of this problem and can now fault after 4 mins. Perhaps small blocksize will help us track down the problem on non-Jack set ups too. However, perhaps Udo's progress is only relevant to using Jack and no help to Mac?
Linux
- Reproducible problem reported by Michael Schwendt on Fedora 8 test rpms of Audacity CVS from 11 January 2008. "Open..." a sufficiently long audio file, start playback. then while playback is still active, "Open..." a second file of any size. Audacity GUI is locked for a long time. A long series of messages on the console is printed until playback of the first file is nearly done. Then everything returns to normal.
Expression 'ret' failed in 'src/hostapi/alsa/pa_linux_alsa.c', line: 1034
Expression 'AlsaOpen( hostApi, parameters, streamDir, &pcm )' failed in
'src/hostapi/alsa/pa_linux_alsa.c', line: 1066- This could be demoted to a lower category if other Linux users can't reproduce it.
- Jim Cline finds the same error message on console when starting audacity built in debian testing, 31 Jan. 2008. Causes same delay in GUI whenever trying to do anything. I don't know if it interferes with functions since I am just starting to use audacity, and don't know whether its failure is due to this or some configuration problem.
- Reproducible (but not entirely consistent) crashes reported by Darren on Ubuntu Gutsy 7.10 with 2.5 GB projects when using Undo command. Darren further reports that adding a Label Track in any projects causes Edit > Paste to flicker (the code does currently enable it), then closing the Label Track causes a crash every time irrespective of the actions taken before. None of these crashes are reproducible by Gale on Windows XP, but retained on "Watching Brief" just in case (it could affect Mac as well?)
- Fix: Recovery action if project rate not supported by card. OK on Windows.
- Support for OSS not functional (crashes).
- If we are releasing the tarball, as it looks like we will at least as beta, how do we address this? Do we need better release notes, something in the build system or what? We don't want zillions of repeat reports of a known limitation, or Linux users wasting their time trying to make Audacity do something it won't.
- OSS issue/crash reported by SIEGERSTEIN: 'failed in 'src/hostapi/oss/pa_unix_oss.c', line: 1206'
- Address Mihail's issue: PA19 uses oss; without load/compile oss support in alsa (snd-pcm-oss), audacity crashes on init PA19.
Priority Aim to
These are items which were in the aim-to list which are being prioritised. None of them are show stoppers which can hold up the release.
Marked for release noting
The following items have all been marked for release-notes. Of these we should try to clear the ones marked "!" which are deemed to be the most important ones to clear if time and resources are available.
Any items marked with a "$" are believed cleared, awaiting confirmation.
- Export Multiple fails with no export or warning if an empty label is encountered.
- ! The preference "Hold recorded data in memory until recording is stopped" is misnamed and in fact caches all recorded and imported audio data (and modified audio data when opening saved projects) for the duration of the session. It may cause a crash either when making long recordings or opening large files or projects. It should be reworded and consideration given to a separate preference for caching recorded and other audio data. This will let users have the benefits of caching when recording without the more dangerous and possibly less generally useful caching of imported and changed project data. GA suggests the new preference for only caching recordings could discard the cache after recording stops since the data is already written to disk at that point.
- When changing language in Preferences, "Audio Track" at the top of the Track Drop-down menu and the toolbars tooltips (except Selection Bar) don't change language until after restart, or until a forced reset.
- Pressing Play (but not spacebar) in a second project when another is already playing stops playback of the first project (is a regression on 1.2.6)
- IE or Firefox not brought to top (on some Windows machines) when help system invokes external browser.
- ! Fix: When you close all your tracks and File > Save or Save Project As.., no warning is given. Save Project As.. with no tracks open is legitimate if there have never have been any changes, so it's suggested the warning should be given only if there is an Undo history. Promoted from standard "aim to", as currently there is no protection in place for the user inadvertently saving an empty project.
- ! Ensure all edit menu items (not just cut and copy) are consistently enabled and work consistently. See here .
- JC: Believed fixed. However is it really what we want? E.g click split-new twice, and we get four tracks. Is it not less confusing to leave it greyed out?
- GA: If you turn off select-all-on-none, select an area and "Split New" twice, you also get four tracks, so I think it is what we want. Now, if you have a track selected with no time area selected, nothing on the clipboard and click the Edit menu, all the edit menu is enabled except Paste. That's fine, if someone hits some item that has no meaning when all is selected like Split, Join or Trim, they just get the whole track selected which is arguably more useful than nothing happening, and we don't grey those items when all is selected with select-all-on-none off anyway.
However ,if we click outside the track(s) so that none are selected, we lose ability to click on all of these (including ones that have no purpose with select all):
-Split Cut
-Trim
-Delete
-Split Delete
-Silence
-Join
-Disjoin
-Duplicate
-Find Zero Crossings
yet we could still Cut, Copy and Split (FWIW) and Split New. I think that will be very confusing as it's a halfway house between Select-all-on-none and not. So I think what I would like to clear these two items is for all Edit menu items to be available (except Paste if there is nothing on the clipboard) even if no tracks are selected.
PS There is also File > Export Selection to consider which behaves like some of the Edit menu items above i.e. it is greyed out when clicking outside the tracks. You might I suppose make a case just for this and Selection Save being greyed out when no audio is selected, even with select-all-on-none "on". But it would be more consistent if they simply were enabled whenever a selection is valid, like Edit > Copy.
- ! Fix: Old projects open incorrectly - reported by Monty. With CVS Head, sample project created in 1.1.0 now correctly identifying the real orphans (there are 10), but the waveform is still opening as blank. Needs more investigation. GA tried a 1.1.3 Project with envelope edits in 180907 build and it opened fine. There are no envelope points, but there aren't any after saving/reopening the Project in 1.1.3 either.
- Beep on completing long process is not working on Windows XP and 98 for GA and DJS respectively. If not fixed, problem should be release noted.
- JC: Beep is using the wxWindows function wxBell(), which on some systems uses an internal speaker and does nothing on others. wxBell() seems only to call the system noise from pre-soundcard days, not one of the system WAV files. This system noise is often reported to be inaudible when called by wxBell(). Changing to a user selected sound or other way of producing sound is probably not sensible for 1.4 so will be release noted.
- ! Fix: Make all platforms consistently force Project Rate to the rate of imported file irrespective whether that rate is supported or not. As of 1.3.5.rc builds, this works on Windows but not for Michael on Linux. When done, add a Preference to always have Project Rate remain as per its Default Sample Rate Preference setting irrespective of the file being imported (this is currently "not aiming for 1.4.0").
- ! Tag import / export is near complete for MP3 but otherwise incomplete, and in places non-orthogonal. See description on Known Issues. This is not now language-dependent, and no issues which were regressions on 1.2.6 remain. Some known specifics: 1) Audacity "Year" MP3 tags not seen by Windows Media Player or Winamp, and "Year" OGG tags not seen by Amarok on Linux, or VLC on Windows or Linux. 2) No FLAC tags are seen by VLC on Windows or Linux. 3) WAV fields only seen patchily by other apps. as their support is patchy, but Genre tag consistently not seen though it is a valid WAV info tag.
- GA: Are we using ID3 V2.4 or 2.3? With the two main Windows players not currently seeing the Year tag, we may have a big reason for users not to upgrade from 1.2.6. Can we use v2.2 where WMP and Winamp see the year tag?
- Update: Leland I believe thinks we can't hack further back than 2.3. The problem seems to be that Windows Explorer/Media Player and Winamp currently expect a frame called "TYER" to indicate the year of the recording. ID3v2.3 seems to recommend passing year as the "TDRC" frame in preference, which is what lame_enc.dll does, resulting in the Winamp/WMP problem. But if you use LAME.exe 3.97 for command line export and force ID3v2.3 with "--id3v2-only", "TYER" is used and WMP and Winamp have no problem. So is it possible to use "TYER", and is this less bad than the current situation?
Not currently marked for release noting
It's suggested items marked as "!" are more important or produce the best return for effort expended.
- Memory leak in Effect::CountWaveTracks that could cause problems.
- ! Tab order of controls in SelectionBar is incorrect
- Firefox does not interpret "#" character in inner text URLs e.g. the link to "Super" (http://www.erightsoft.com/SUPER.html#Dnload ) does not go to anchor.
- Links from built-in text to wiki to be locally over-ridable in the same way that links from built-in text to external html files can be. This way we can have a local copy of the relevant Wiki pages, if we want to.
- ! When selecting a region over a label and cutting or copying, ensure label text, not the system clipboard content, is pasted into the text of the label (this is a regression on 1.2.6, and the 1.4 manual would need amendment if we can no longer cut/copy and paste multiple labels except by means of the Label Editor). System clipboard seems to get pasted because creating a label enables Paste in the edit menu (though this does not consistently happen, and Darren's observation of Paste flickering on Linux might be tied in with this).
- ! When selecting a region over a label that denotes a selection area and cutting or copying, ensure the pasted label is of the selection region, not a cursor point (this is a regression on 1.2.6)
- Generate effects do not fit in window, whereas opening and importing audio does (regression on 1.2.6, and presumed undesirable for most users. Serious nuisance if you want to create very short tones for analysis).
- Dependencies Dialogue: If you click on "Copy Selected Audio Into Project" in error, the copying cannot be cancelled from the progress dialogue.
- This can probably be left for after 1.4 unless it's an easy fix
- Re-enable file writing within Nyquist so that Nyquist plug-ins can remember user settings, but be careful about the permissions with file writing, and whether and when we allow over-writing of existing files. Needs someone with detailed LISP knowledge to look at it. Assumption is that we need to re-enable code intentionally disabled, not write new code.
- JC: xlisp functions such as xwrchar, xprint and xopen :OUTPUT seem to be already present and correct. Can't find disabled code. DS hypothesises it is code missing in 'the bridge', whatever that is.
- When writing a file as a result of a Nyquist prompt, file either is not written or is written to a location that is not asked for (the same one that Audacity was using to write noisegate.nrp to before it was fixed). See here for details.
Not aiming to do for 1.4.0
This has moved to release checklist not aiming for 1.4 and now includes former "Aim tos" from this page. Many of these "Aim tos" are relatively small items but could nonetheless make a worthwhile difference for the time involved. Please do not forget them.
If you claim ownership to items in "Not Aiming", please cut them from that page and park them here: