windows build download #110
Comments
|
Sounds like it's time to make a Windows build! I will start putting the
installer into a Seq66 release. Will take a bit of time to make sure it
works. In the meantime, build instructions are provided in
nsis/build_release_package.bat; after installing Qt Creator correctly and
setting paths, it automatic.
On Sun, May 7, 2023 at 06:47 furrig ***@***.***> wrote:
Hi !
Last available windows build seems to be here :
https://github.com/ahlstromcj/sequencer64-packages/tree/master/seq66/0.98.0
It's 2 years old, is there a more recent availability?
(if not build it myself but I'm shy with QT ;) )
—
Reply to this email directly, view it on GitHub
<#110>, or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AB45LKHRLZL2VTJHQMZSQVLXE54ULANCNFSM6AAAAAAXYZCOJQ>
.
You are receiving this because you are subscribed to this thread.Message
ID: ***@***.***>
--
Chris Ahlstrom
|
|
Thanks a lot ! I tried to build but had no success (logs are attached)
PS : to be able to launch make with 64 bits, I had to modify build_release_package.bat when it says |
|
I got rid of the compiler errors pretty quickly. I worked on installing NSIS
in my Windows virtual machine and updating the scripts to make 64-bit builds
the default and also allow the option of automatically building the installer
after the build, under Windows, instead of copying the build to Linux to use
the Linux version of NSIS. Still working out some issues, including why the
intial first-startup files are not being written to
C:/Users/user/AppData/Local/seq66. Then have to figure out why the error with
output ports is cropping up again.
But hope to get all this done by the weekend.
|
|
Still working out this issue. Had to refactor the code for the configuration directories, found a bug in the Linux version of the internal portmidi engine that also affected Windows. The Windows MIDI Mapper issue somehow came back and I am working that out. Added support for using NSIS (if installed) in Windows in the build batch file. Improved the port naming a bit. Trying to handle present-but-unavailable MIDI ports better. Long story short, it'll be a little while yet. :-( |
|
Whew, lotta stuff tweaked. Got build_release_package.bat to build and the application to run (though there's a bogus error that pops up in the output port that grabs the Korg nanoKEY2. Low priority unless you find something similar, since my Launchpad Mini causes no message and the nanokey works for entering notes. Lotta surprises along the way. Tomorrow I will likely make 0.99.5 official and add the installer to the release. I'm only worrying about 64-bit Windows at the moment; but the installer still prompts to install to Program Files (x86); you change that to plain Program Files (Window 10) and it works. If you cannot wait, you can check out the "portfix" branch and see if you can break it. |
This file lists __major__ changes from version 0.99.5
- Issue #110 Windows: Fixed many errors, added installer to this release,
no more going to the sequencer64 repository. Can build NSIS installer
in Windows now.
- Fixed portmidi bugs in Linux and Windows, enhanced the device naming.
- Greatly enhanced the event editor tab and added more events that can
be view and modified.
- Made port-mapping the default. At first startup the map exactly matches
the existing ports; can be changed in the 'rc' file or Preferences dialog.
Used the Edit / Preferences / MIDI Clock / Make Maps button to
refresh the port setup when devices are added or removed.
- Eliminated "missing ctrl" message at first startup.
- Fixed port ID setting in midibus, and adding output flag for
ALSA MIDI info.
- Internal refactoring to regularize handling of the session/config
directory between Linux and Windows.
- Shows disabled/unavailable MIDI devices as grayed in various dropdowns.
- Rearranged the Seq66 man pages more sensibly.
Read the NEWS, README.md, and TODO files. Never-ending!
|
Okay, pushed 0.99.5 to master and created a new release. The new release has an exe installer attached to it. There are some minor issues that may crop up, but read data/readme.windows and the user manual. If any showstoppers, sound the alarm! I won't close this issue until you are satisfied. |
|
Hi ! Spent a couple of hours on it, below some issues I found
Patterns editor
|
|
Thanks for the reports. I will check them out. A couple comments....
The MIDI Clock tab problem is long standing, and also seems dependent on the Qt theme. By moving the mouse down further in the tab's top thumb, one can find an area where the thumb might be highlighted and the click will work. I need to goo-goo a bit to see if this problem has affected other apps.
Ctrl-Z Keystrokes generally require clicking in the piano roll first to give it focus. The most annoying instance is clicking the space bar to toggle playback when the focus is still on a drop-down; one has to click an empty button in the grid to let the space bar work. Could perhaps allow certain keystrokes to be forwarded to the piano rolls.
As for velocity changes, I assume you're doing that in the data pane (the one with the vertical lines for each note. It's working as "designed", but it sounds good to make it obey selections if any exist.
I will start looking into the other issues soon... Thanks!
|
|
Concerning velocity changes on selected notes only, an example how Seq24 works : https://www.youtube.com/watch?v=i3Vpi3oxdqk&t=294s |
|
I built seq24 from source (a pain now because g++ thinks the seq24 mutex class
is ambiguous with std::mutex) it has the same feature as seq66: if no notes
are selected, the velocity change affects all notes; otherwise, it affects only
selected notes. So I think this is a non-issue. The behavior is documented
deep in the user manual (a needle in a haystack :-D).
|
|
Whoops, now I see what you mean. Didn't see that on my little laptop. Will
fixit.
|
|
Just an update. I think I've got all of the newer issues resolved except for the display of the tutorial and manual. I refactored the related code to be more robust, but it still acts the same. In the meantime, I fixed issues with redirecting the console messages to a log file. Trying to improve application icon support. Also finally got the pattern editor panes to respond uniformly to the arrow and page keys no matter with pane has focus. The piano roll is the "scroll master" which the keyboard, time, event, and data panes are "scroll slaves". The "hjkl" (vi motion keys) work. Somehow I can't get the wheel events (mouse scrolling) to totally work, it works only if the piano roll has focus. Currently am adding new Windows icons to the NSIS installer. I hope to release all this next weekend; would like to pound on it some more. Thanks for all your reports! They make for a better product! |
|
Hi all, have been trying to build this on my old win xp machine with no luck, still on seq24. Is it possible to release a 32 bit version of the installer? Otherwise i will study further on building it. thnx! |
|
I will test a 32-bit build and make an additional installer. Probably a
couple or three days from now. Currently hammering on the code in the
portfix branch. The build-release batch file explains what to do, you're
impatient and willing to monkey around with it.
On Sun, Jun 4, 2023 at 11:58 Moerkie ***@***.***> wrote:
Hi all, have been trying to build this on my old win xp machine with no
luck, still on seq24. Is it possible to release a 32 bit version of the
installer? Otherwise i will study further on building it. thnx!
—
Reply to this email directly, view it on GitHub
<#110 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AB45LKAHLVLGRZZSIKPYRH3XJSWD3ANCNFSM6AAAAAAXYZCOJQ>
.
You are receiving this because you commented.Message ID:
***@***.***>
--
Chris Ahlstrom
|
|
That’s great to hear. I’m inexperienced with coding, I started working
through the ‘qmake’ tutorial you wrote. It sure feels like I’m monkeying
more than actually building.
Will do my best to study on this material.
Op zo 4 jun. 2023 om 19:54 schreef C. Ahlstrom ***@***.***>
… I will test a 32-bit build and make an additional installer. Probably a
couple or three days from now. Currently hammering on the code in the
portfix branch. The build-release batch file explains what to do, you're
impatient and willing to monkey around with it.
On Sun, Jun 4, 2023 at 11:58 Moerkie ***@***.***> wrote:
> Hi all, have been trying to build this on my old win xp machine with no
> luck, still on seq24. Is it possible to release a 32 bit version of the
> installer? Otherwise i will study further on building it. thnx!
>
> —
> Reply to this email directly, view it on GitHub
> <#110 (comment)
>,
> or unsubscribe
> <
https://github.com/notifications/unsubscribe-auth/AB45LKAHLVLGRZZSIKPYRH3XJSWD3ANCNFSM6AAAAAAXYZCOJQ
>
> .
> You are receiving this because you commented.Message ID:
> ***@***.***>
>
--
Chris Ahlstrom
—
Reply to this email directly, view it on GitHub
<#110 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/BAKBMD2ZIEO3PNVN42MTXODXJTDWJANCNFSM6AAAAAAXYZCOJQ>
.
You are receiving this because you commented.Message ID:
***@***.***>
|
|
Just found a couple of compiler errors in the 32-bit build, so no need to even try to build it yourself :-( |
|
After trying to get a Win32 build on my 64-bit windows using mingw32, I am punting this issue to another release. I can make a 32-bit version of qpseq66.exe in Qt Creator, and it runs if I started it from Qt Creator. However, it does not run on its own, even after doing the Qt Creator deploy command. And using the build batch script yields errors: (1) WinMain() unresolved; (2) a whole bunch of Qt modules undefined. I tweaked the make to try to match what Qt Creator does when it makes, to no avail. The only other option I think is to get an old 32-bit version of Windows, installed Qt Creator on that, and try it. But I won't do that, as I have no obvious source for that old operating system. Sorry. If someone finds a way to do it, please let me know. For now, only the 64-bit version will be provided. :-( |
|
That’s a pitty but thank you so much for trying. Great to see that this
program is alive and kicking 👍🏼
Op ma 12 jun. 2023 om 14:00 schreef C. Ahlstrom ***@***.***>
… After trying to get a Win32 build on my 64-bit windows using mingw32, I am
punting this issue to another release. I can make a 32-bit version of
qpseq66.exe in Qt Creator, and it runs if I started it from Qt Creator.
However, it does not run on its own, even after doing the Qt Creator deploy
command. And using the build batch script yields errors: (1) WinMain()
unresolved; (2) a whole bunch of Qt modules undefined. I tweaked the make
to try to match what Qt Creator does when it makes, to no avail. The only
other option I think is to get an old 32-bit version of Windows, installed
Qt Creator on that, and try it. But I won't do that, as I have no obvious
source for that old operating system. Sorry. If someone finds a way to do
it, please let me know. For now, only the 64-bit version will be provided.
:-(
—
Reply to this email directly, view it on GitHub
<#110 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/BAKBMD5Y4LJ226C5TBPFE6TXK4AEHANCNFSM6AAAAAAXYZCOJQ>
.
You are receiving this because you commented.Message ID:
***@***.***>
|
|
I was planning to make a new relase last weekend, but had to re-wrestle with the Microsoft MIDI Mapper, which, contrary to reports, is still in Windows 10. But it does NOT have the Control Panel entry people talk about. Right now, the situation on my VM is that device 0 is the MIDI Mapper, which grabs device 1, the builti-in GS Wavetable Synth, then devices 2 on up are the hardware devices plugged into the system. Device 1 is rendered unavailable. To avoid a chronic error message at startup, the code detects this locking in Windows and avoids reporting on it. (The alternative is to use the new option to suppress startup messages.) But .... .... This leads to a question: Do you use a third party mapper like CoolSoft MIDI Mapper or Midi-Ox? If so, does it have the same issue as the Microsoft MIDI Mapper? Any things I need to account for? Thanks! |
|
I'm not sure I understood the issue you explained but :
|
|
Can you post the midi-input, input maps, clocks, and clocks maps here? I
wonder if the Microsoft midi mapper is interfering. The 0.99.6 release
(Saturday I hope) handles this situation better. Was a hella lotta work.
On Fri, Jun 30, 2023 at 06:35 furrig ***@***.***> wrote:
I'm not sure I understood the issue you explained but :
- I have this error message when I start seq66 : "Session error
portmidi host error: [7] 'The specified device is already in use. Wait
until it is free, and then try again.' Check MIDI Clock & MIDI Input tabs
for unavailable/missing ports. Try the 'Make Maps' button and restart
Seq66."
- I use loopMidi (https://www.tobias-erichsen.de/software/loopmidi.html)
and all of his virtual ports appear and are available in seq66 prefs on
both MIDI Clock & MIDI Input tabs
—
Reply to this email directly, view it on GitHub
<#110 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AB45LKHROEKEUVWDF3OWXZDXN2TWLANCNFSM6AAAAAAXYZCOJQ>
.
You are receiving this because you commented.Message ID:
***@***.***>
--
Chris Ahlstrom
|
|
Oh, you can find the setup in your AppData/Local/seq66/qpseq66.rc file.
On Fri, Jun 30, 2023 at 06:42 Chris Ahlstrom ***@***.***> wrote:
Can you post the midi-input, input maps, clocks, and clocks maps here? I
wonder if the Microsoft midi mapper is interfering. The 0.99.6 release
(Saturday I hope) handles this situation better. Was a hella lotta work.
On Fri, Jun 30, 2023 at 06:35 furrig ***@***.***> wrote:
> I'm not sure I understood the issue you explained but :
>
> - I have this error message when I start seq66 : "Session error
> portmidi host error: [7] 'The specified device is already in use. Wait
> until it is free, and then try again.' Check MIDI Clock & MIDI Input tabs
> for unavailable/missing ports. Try the 'Make Maps' button and restart
> Seq66."
> - I use loopMidi (
> https://www.tobias-erichsen.de/software/loopmidi.html) and all of his
> virtual ports appear and are available in seq66 prefs on both MIDI Clock &
> MIDI Input tabs
>
> —
> Reply to this email directly, view it on GitHub
> <#110 (comment)>,
> or unsubscribe
> <https://github.com/notifications/unsubscribe-auth/AB45LKHROEKEUVWDF3OWXZDXN2TWLANCNFSM6AAAAAAXYZCOJQ>
> .
> You are receiving this because you commented.Message ID:
> ***@***.***>
>
--
Chris Ahlstrom
--
Chris Ahlstrom
|
|
Here it is |
|
I finally got seq64 to build and installed loopMidi and played with it using the portfix branch version of qpseq66. That seemed to work. So one option is to wait for tomorrow or Sunday to see if 0.99.6 does the trick. One weird thing is I swear I was building qseq64 in a 32-bit version awhile back. But now it builds as a 64-bit version, and I had to fix a buncha UINT_PTR vs UINT and DWORD_PTR vs DWORD issues. |
This file lists __major__ changes from version 0.99.6
- Issue #3 follow-ons:
- Pattern editor panes stay in sync with the piano roll using
the hjkl, arrow, and page keys. Scroll wheel works in
the piano roll.
- Issue #110 follow-ons:
- Addition of Start menu entries for Windows.
- Fixed access to the tutorial and manual.
- data/readme & doc/tutorial files fixed for NSIS installer.
- Fixed the saving of modified tempo changes.
- Fixed changing note velocities in the pattern editor data
pane. Improved velocity-change undo.
- Fixed error preventing changing the "background" pattern.
- Remaining issue: Building 32-bit (Windows XP) version on
64-bit Windows.
- Issue #111:
- Added support for editing, storing, and displaying time
signatures in the pattern and event editors.
- The first time-signature in a pattern becomes the main
time signature of the pattern.
- The data pane shows a time-signature as a simple fraction.
- Changing the time signature if at time 0 is automatic.
- Time signatures at later times are logged by setting the
current time with a click in the top half of the time line,
changing the beats and beat width, then clicking a
time-sig log button.
- Non-MIDI-standard beat-widths are supported as a
Seq66-specific "event".
- Fixed event filtering in the event (qstriggereditor) pane.
- As time-signatures change, Seq66 adjusts the piano roll,
time line (with measure counts), and event pane vertical
lines.
- Port-mapping prompts about port issues and allows an immediate
remap-and-restart.
- 'o' keystroke in seqroll toggles recording ('r' already
taken).
- Added a "quiet" option to not show startup message prompts.
- A log-file is now created by default & kept under a megabyte.
- Added the pattern port number to the Song Summary output.
- A large number of fixes of unrelated issues.
Read the NEWS, README.md, and TODO files. Working our assoff!
// vim: sw=4 ts=4 wm=15 et ft=sh
|
Latest release 0.99.6 in place in master branch for you to whack at! User manual updated to match. |
|
Thanks for the release :) Did some quick tests. |
|
What part of the user interface did you use? I can click or edit the BPM
spin box and the new BPM gets saved. (This on Linux, haven't yet tried it
in Windows).
I also have some time signature leftovers to fix and hope to get the next
release out with a shorter turnaround time.
…--
Chris Ahlstrom
|
|
I tried with both clicking and editing |
|
It turns out the problem of changing the tune BPM is due to a check for a file
name being active. If one starts Seq66 with no song, changing the BPM does not
cause the file modification asterisk to a appear. If one loads a song and then
changes the BPM, the modify flag appears and saving the file saves the new BPM.
So the work around is to load a song or create a pattern when Seq66 starts up
empty.
I'm not sure if I want to eliminate the file-check. Opinion? Thanks!
|
|
I don't have the same behaviour. |
|
I have duplicated the error in Windows and it is weird with a beard! The BPM is written as a c_bpmtag SeqSpec, but a breakpoint on the parse for that item is hit only on Linux. So either the writing of it or the reading is broken on Windows. But this code is not OS-dependent. So a debugging slog is underway. :-( |
|
So the bug was in the midifile::skip() function called with a negative number to move backward in the MIDI data. In the Windows build (using mingw32) the small negative number was being treated as positive, so that the skip was ignored and the data desired was not obtained. Added a "backup()" function; now calls backup(2) rather than skip(-2). I'm going to look into some other TODO's before making a new release. If you're in a hurry, clone the project, go to the seq66/nsis directory in a DOS window, and run the .BAT build script. Then can run the installer or unpack the 7z (7-Zip) file to see if the "portable pack" runs. |
This file lists __major__ changes in version 0.99.7
- Issue #110 follow-ons:
- Fixed saving tempo (BPM) in Windows when changed from main
window. Caused by mixing a long and size_t; messed up in
Windows builds.
- Issue #111 follow-ons:
- Fixed initial time-signature drawing in data pane.
- Fixed errors in inserting a time-signature.
- Added a pulse (tick) calculator to iterate through
time-signatures.
- Fixed an important port-translation bug in output port-mapping.
- Revamped the Playlist tab, as it was confusing and very buggy.
- Added auto-play and auto-advance to play-lists.
- Issue #102: Added Windows key-mapping to fix processing
"native virtual" keys, such as the arrow keys.
Read the NEWS, README.md, and TODO files.
|
Since we're up to version 0.99.8, closing this one. Reopen if necessary! |
Hi !
Last available windows build seems to be here : https://github.com/ahlstromcj/sequencer64-packages/tree/master/seq66/0.98.0
It's 2 years old, is there a more recent availability?
(if not build it myself but I'm shy with QT ;) )
The text was updated successfully, but these errors were encountered: