Test 3 is out. This should resolve all the trouble we've had with test2 except that one tester on MacOS is still finding that Pd 0.55 loads way more slowly than 0.54 - I don't see that happening on my machine and I don't know how many machines will be affected... it might take a while to track this one down, but hopefully it won't affect many people. If you _do_ notice longer startup times on MacOS please let me (or the Pd list) know - with more data points it migth be easier for me to replicate the problem for myself.
as usual: http://msp.ucsd.edu/software.htm OR https://msp.ucsd.edu/software.htm
source on github: https://github.com/pure-data/pure-data
cheers
Miller
Pd-announce mailing list Pd-announce@lists.iem.at https://lists.puredata.info/listinfo/pd-announce
Cool, I also see we have a new table size for osc~ and the issue https://github.com/pure-data/pure-data/issues/105 is finally fixed, right?
Seems like something worth sharing in the release notes, can you provide a draft here so I can include it?
thanks
Em sex., 31 de mai. de 2024 às 14:10, Miller Puckette < mpuckette@cloud.ucsd.edu> escreveu:
Test 3 is out. This should resolve all the trouble we've had with test2 except that one tester on MacOS is still finding that Pd 0.55 loads way more slowly than 0.54 - I don't see that happening on my machine and I don't know how many machines will be affected... it might take a while to track this one down, but hopefully it won't affect many people. If you _do_ notice longer startup times on MacOS please let me (or the Pd list) know - with more data points it migth be easier for me to replicate the problem for myself.
as usual: http://msp.ucsd.edu/software.htm OR https://msp.ucsd.edu/software.htm
source on github: https://github.com/pure-data/pure-data
cheers
Miller
Pd-announce mailing list Pd-announce@lists.iem.at https://lists.puredata.info/listinfo/pd-announce _______________________________________________ Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> https://lists.puredata.info/listinfo/pd-list
Thanks for this new release! Unfortunately, I can confirm the extra startup delay for macOS... I tested on an M1 machine with 64GB of memory running macOS 14.5.
Also, I get the attached popup every time I open the app, even after clicking "allow." Maybe there's something in macOS preferences that will suppress it? After clicking "allow," it takes a full 8sec for the Pd window to appear. I've only timed this casually with my watch at this point, but after trying multiple times I can say that it's consistently around 8sec for me here. Pd-0.54-1 consistently springs to life in about 1sec for me.
I'm happy to run any specific test routines if I can be useful on this. William
On Fri, May 31, 2024 at 1:10 PM Miller Puckette mpuckette@cloud.ucsd.edu wrote:
Test 3 is out. This should resolve all the trouble we've had with test2 except that one tester on MacOS is still finding that Pd 0.55 loads way more slowly than 0.54 - I don't see that happening on my machine and I don't know how many machines will be affected... it might take a while to track this one down, but hopefully it won't affect many people. If you _do_ notice longer startup times on MacOS please let me (or the Pd list) know - with more data points it migth be easier for me to replicate the problem for myself.
as usual: http://msp.ucsd.edu/software.htm OR https://msp.ucsd.edu/software.htm
source on github: https://github.com/pure-data/pure-data
cheers
Miller
Pd-announce mailing list Pd-announce@lists.iem.at https://lists.puredata.info/listinfo/pd-announce _______________________________________________ Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> https://lists.puredata.info/listinfo/pd-list
Hmm, sounds like there are more than one problem...
Can you (and Ben too if he's willing) try these...
time ..../Pd-0.55-0test2c.app/Contents/Resources/bin/pd -send "pd quit"
time ..../Pd-0.55-0test2c.app/Contents/Resources/bin/pd -nosound -send "pd quit"
time ..../Pd-0.55-0test2c.app/Contents/Resources/bin/pd -noprefs -send "pd quit"
time ..../Pd-0.55-0test2c.app/Contents/Resources/bin/pd -nostdpath -send "pd quit"
I'm grasping at straws here :) Guessing that William's 8-second startup and new popup might be something external that's fighting with some Mac plist setting in Pd that might have changed.
thanks M
On 6/1/24 12:12 AM, William Brent wrote:
Thanks for this new release! Unfortunately, I can confirm the extra startup delay for macOS... I tested on an M1 machine with 64GB of memory running macOS 14.5.
Also, I get the attached popup every time I open the app, even after clicking "allow." Maybe there's something in macOS preferences that will suppress it? After clicking "allow," it takes a full 8sec for the Pd window to appear. I've only timed this casually with my watch at this point, but after trying multiple times I can say that it's consistently around 8sec for me here. Pd-0.54-1 consistently springs to life in about 1sec for me.
I'm happy to run any specific test routines if I can be useful on this. William
On Fri, May 31, 2024 at 1:10 PM Miller Puckette mpuckette@cloud.ucsd.edu wrote:
Test 3 is out. This should resolve all the trouble we've had with test2 except that one tester on MacOS is still finding that Pd 0.55 loads way more slowly than 0.54 - I don't see that happening on my machine and I don't know how many machines will be affected... it might take a while to track this one down, but hopefully it won't affect many people. If you _do_ notice longer startup times on MacOS please let me (or the Pd list) know - with more data points it migth be easier for me to replicate the problem for myself. as usual: http://msp.ucsd.edu/software.htm OR https://msp.ucsd.edu/software.htm source on github: https://github.com/pure-data/pure-data <https://urldefense.com/v3/__https://github.com/pure-data/pure-data__;!!Mih3wA!GVjycnnXvxANGR5EvpOhP-gDSVl648XDg2uY4mOCl4Bx2j-OYstW0TnDQ7n7R0W2zf5OF7RaFNFSCSjeka4$> cheers Miller _______________________________________________ Pd-announce mailing list Pd-announce@lists.iem.at https://lists.puredata.info/listinfo/pd-announce <https://urldefense.com/v3/__https://lists.puredata.info/listinfo/pd-announce__;!!Mih3wA!GVjycnnXvxANGR5EvpOhP-gDSVl648XDg2uY4mOCl4Bx2j-OYstW0TnDQ7n7R0W2zf5OF7RaFNFSA-t2mYg$> _______________________________________________ Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> https://lists.puredata.info/listinfo/pd-list <https://urldefense.com/v3/__https://lists.puredata.info/listinfo/pd-list__;!!Mih3wA!GVjycnnXvxANGR5EvpOhP-gDSVl648XDg2uY4mOCl4Bx2j-OYstW0TnDQ7n7R0W2zf5OF7RaFNFSC4IW8vI$>-- William Brent
“Great minds flock together” Conflations: conversational idiom for the 21st century
www.conflations.com https://urldefense.com/v3/__http://www.conflations.com__;!!Mih3wA!GVjycnnXvxANGR5EvpOhP-gDSVl648XDg2uY4mOCl4Bx2j-OYstW0TnDQ7n7R0W2zf5OF7RaFNFShf8l5Rw$
Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> https://urldefense.com/v3/__https://lists.puredata.info/listinfo/pd-list__;!...
Can you (and Ben too if he's willing) try these...
sure! - here are my results for the new Pd-0.55-0test3a (or did you actually mean to check Pd-0.55-0test2c once more?) with 'time': /Applications/Pd-0.55-0test3a.app/Contents/Resources/bin/pd -send "pd quit" 0.03s user 0.02s system 2% cpu 2.132 total /Applications/Pd-0.55-0test3a.app/Contents/Resources/bin/pd -noverbose -send 0.02s user 0.03s system 2% cpu 2.202 total /Applications/Pd-0.55-0test3a.app/Contents/Resources/bin/pd -noverbose -send 0.01s user 0.02s system 2% cpu 1.532 total /Applications/Pd-0.55-0test3a.app/Contents/Resources/bin/pd -noverbose -send 0.02s user 0.02s system 2% cpu 2.132 total
... there's a bit of variation though - therefore i'm additionally measuring with repeated runs through hyperfine with a "warmup" (as learnt from iohannes' tests ... see https://lists.puredata.info/pipermail/pd-list/2024-05/133227.html and https://lists.puredata.info/pipermail/pd-list/2024-05/133242.html ):
hyperfine --warmup 3 -L params '-send "pd quit"','-nosound -send "pd quit"','-noprefs -send "pd quit"','-nostdpath -send "pd quit"' '/Applications/Pd-0.55-0test3a.app/Contents/Resources/bin/pd {params}'
Benchmark 1: /Applications/Pd-0.55-0test3a.app/Contents/Resources/bin/pd -send "pd quit" Time (mean ± σ): 1.545 s ± 0.007 s [User: 0.017 s, System: 0.015 s] Range (min … max): 1.530 s … 1.553 s 10 runs
Benchmark 2: /Applications/Pd-0.55-0test3a.app/Contents/Resources/bin/pd -nosound -send "pd quit" Time (mean ± σ): 1.545 s ± 0.008 s [User: 0.017 s, System: 0.015 s] Range (min … max): 1.531 s … 1.554 s 10 runs
Benchmark 3: /Applications/Pd-0.55-0test3a.app/Contents/Resources/bin/pd -noprefs -send "pd quit" Time (mean ± σ): 1.524 s ± 0.016 s [User: 0.010 s, System: 0.014 s] Range (min … max): 1.505 s … 1.559 s 10 runs
Benchmark 4: /Applications/Pd-0.55-0test3a.app/Contents/Resources/bin/pd -nostdpath -send "pd quit" Time (mean ± σ): 1.546 s ± 0.015 s [User: 0.017 s, System: 0.015 s] Range (min … max): 1.529 s … 1.581 s 10 runs
... so these results are quite similar (unless started with "-nogui")
i also got the dialog that william mentions though. could this be connected? are those test versions signed the same way as Pd-0.54-1? i gave them permission manually in the security settings to avoid the repeated warnings and didn't remember if i did the same for pd-0.54-1. but obviously, it's not necessary there (just reinstalled that one to check, too ... hoping i didn't miss any persisting settings there!).
cheers, ben
On Sat, Jun 1, 2024 at 7:33 AM Miller Puckette mpuckette@cloud.ucsd.edu wrote:
Hmm, sounds like there are more than one problem...
Can you (and Ben too if he's willing) try these...
time ..../Pd-0.55-0test2c.app/Contents/Resources/bin/pd -send "pd quit"
time ..../Pd-0.55-0test2c.app/Contents/Resources/bin/pd -nosound -send "pd quit"
time ..../Pd-0.55-0test2c.app/Contents/Resources/bin/pd -noprefs -send "pd quit"
time ..../Pd-0.55-0test2c.app/Contents/Resources/bin/pd -nostdpath -send "pd quit"
I'm grasping at straws here :) Guessing that William's 8-second startup and new popup might be something external that's fighting with some Mac plist setting in Pd that might have changed.
thanks M
On 6/1/24 12:12 AM, William Brent wrote:
Thanks for this new release! Unfortunately, I can confirm the extra startup delay for macOS... I tested on an M1 machine with 64GB of memory running macOS 14.5.
Also, I get the attached popup every time I open the app, even after clicking "allow." Maybe there's something in macOS preferences that will suppress it? After clicking "allow," it takes a full 8sec for the Pd window to appear. I've only timed this casually with my watch at this point, but after trying multiple times I can say that it's consistently around 8sec for me here. Pd-0.54-1 consistently springs to life in about 1sec for me.
I'm happy to run any specific test routines if I can be useful on this. William
On Fri, May 31, 2024 at 1:10 PM Miller Puckette mpuckette@cloud.ucsd.edu wrote:
Test 3 is out. This should resolve all the trouble we've had with test2 except that one tester on MacOS is still finding that Pd 0.55 loads way more slowly than 0.54 - I don't see that happening on my machine and I don't know how many machines will be affected... it might take a while to track this one down, but hopefully it won't affect many people. If you _do_ notice longer startup times on MacOS please let me (or the Pd list) know - with more data points it migth be easier for me to replicate the problem for myself. as usual: http://msp.ucsd.edu/software.htm OR https://msp.ucsd.edu/software.htm source on github: https://github.com/pure-data/pure-data <https://urldefense.com/v3/__https://github.com/pure-data/pure-data__;!!Mih3wA!GVjycnnXvxANGR5EvpOhP-gDSVl648XDg2uY4mOCl4Bx2j-OYstW0TnDQ7n7R0W2zf5OF7RaFNFSCSjeka4$> cheers Miller _______________________________________________ Pd-announce mailing list Pd-announce@lists.iem.at https://lists.puredata.info/listinfo/pd-announce <https://urldefense.com/v3/__https://lists.puredata.info/listinfo/pd-announce__;!!Mih3wA!GVjycnnXvxANGR5EvpOhP-gDSVl648XDg2uY4mOCl4Bx2j-OYstW0TnDQ7n7R0W2zf5OF7RaFNFSA-t2mYg$> _______________________________________________ Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> https://lists.puredata.info/listinfo/pd-list <https://urldefense.com/v3/__https://lists.puredata.info/listinfo/pd-list__;!!Mih3wA!GVjycnnXvxANGR5EvpOhP-gDSVl648XDg2uY4mOCl4Bx2j-OYstW0TnDQ7n7R0W2zf5OF7RaFNFSC4IW8vI$>-- William Brent
“Great minds flock together” Conflations: conversational idiom for the 21st century
www.conflations.com https://urldefense.com/v3/__http://www.conflations.com__;!!Mih3wA!GVjycnnXvxANGR5EvpOhP-gDSVl648XDg2uY4mOCl4Bx2j-OYstW0TnDQ7n7R0W2zf5OF7RaFNFShf8l5Rw$
Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> https://urldefense.com/v3/__https://lists.puredata.info/listinfo/pd-list__;!...
Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> https://lists.puredata.info/listinfo/pd-list
Am 1. Juni 2024 10:50:19 MESZ schrieb Benjamin Wesch benjamin.wesch@gmail.com:
i also got the dialog that william mentions though. could this be connected?
Hmm. I've never seen these.
are those test versions signed the same way as Pd-0.54-1?
I think so.
mfg.sfg.jfd IOhannes
Am 1. Juni 2024 10:56:42 MESZ schrieb "IOhannes m zmölnig" zmoelnig@iem.at:
Am 1. Juni 2024 10:50:19 MESZ schrieb Benjamin Wesch benjamin.wesch@gmail.com: installed
i also got the dialog that william mentions though. could this be connected?
Hmm. I've never seen these.
Most likely is a new feature of iOS/macOS 14. But then Pd-0.54 should be affected as well (unless, maybe, wildly guessing, Pd-0.54 was installed with an older version of macOS, and the os upgrade kept existing permissions)
mfg.sfg.jfd IOhannes
OK, next thing I want to see is what happens if I recompile Pd 0.54-1... I just created a branch to see if I can get the IEM compile chain to grind it out anew.
Hopping on an airplane soon but will be about for 1/2 hour or so I think.
cheers
M
On 6/1/24 11:06 AM, IOhannes m zmölnig wrote:
Am 1. Juni 2024 10:56:42 MESZ schrieb "IOhannes m zmölnig" zmoelnig@iem.at:
Am 1. Juni 2024 10:50:19 MESZ schrieb Benjamin Wesch benjamin.wesch@gmail.com: installed
i also got the dialog that william mentions though. could this be connected?
Hmm. I've never seen these.
Most likely is a new feature of iOS/macOS 14. But then Pd-0.54 should be affected as well (unless, maybe, wildly guessing, Pd-0.54 was installed with an older version of macOS, and the os upgrade kept existing permissions)
mfg.sfg.jfd IOhannes
Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> https://urldefense.com/v3/__https://lists.puredata.info/listinfo/pd-list__;!...
OK, next thing I want to see is what happens if I recompile Pd 0.54-1...
I checked with Pd-0.54-1-1-gb12dcbcb now (from https://git.iem.at/pd/pure-data/-/jobs/72843/artifacts/download?file_type=ar... ) and get the same (good) results as for the official Pd-0.54-1 download. I also don't get these additional "[...] would like to access data from other apps." dialogs. (only once asking to confirm: '“Pd-0.54-1-1-gb12dcbcb.app” is an app downloaded from the internet. Are you sure you want to open it?')
But then Pd-0.54 should be affected as well (unless, maybe, wildly guessing, Pd-0.54 was installed with an older version of macOS, and the os upgrade kept existing permissions)
Just to be (kind of) sure that this can't be connected to any remains from the original Pd-0.54, I also installed an old version that quite certainly had never been on my system yet: 0.52-2 ... that also doesn't give me these warnings. A quick search shows that this dialog is actually a Sonoma "feature" - but it obviously seems weird that it's not triggered for previous Pd versions.
Cheers, Ben
On Sat, Jun 1, 2024 at 11:29 AM Miller Puckette mpuckette@cloud.ucsd.edu wrote:
OK, next thing I want to see is what happens if I recompile Pd 0.54-1... I just created a branch to see if I can get the IEM compile chain to grind it out anew.
Hopping on an airplane soon but will be about for 1/2 hour or so I think.
cheers
M
On 6/1/24 11:06 AM, IOhannes m zmölnig wrote:
Am 1. Juni 2024 10:56:42 MESZ schrieb "IOhannes m zmölnig" zmoelnig@iem.at:
Am 1. Juni 2024 10:50:19 MESZ schrieb Benjamin Wesch benjamin.wesch@gmail.com: installed
i also got the dialog that william mentions though. could this be connected?
Hmm. I've never seen these.
Most likely is a new feature of iOS/macOS 14. But then Pd-0.54 should be affected as well (unless, maybe, wildly guessing, Pd-0.54 was installed with an older version of macOS, and the os upgrade kept existing permissions)
mfg.sfg.jfd IOhannes
Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> https://urldefense.com/v3/__https://lists.puredata.info/listinfo/pd-list__;!...
Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> https://lists.puredata.info/listinfo/pd-list
Thanks - so it's clearly something about the new code... I wish I could get the slowdown to happen on my machine so I could monkey with this myself :)
Miller
On 6/1/24 1:14 PM, Benjamin Wesch wrote:
OK, next thing I want to see is what happens if I recompile Pd 0.54-1...
I checked with Pd-0.54-1-1-gb12dcbcb now (from https://git.iem.at/pd/pure-data/-/jobs/72843/artifacts/download?file_type=ar... ) and get the same (good) results as for the official Pd-0.54-1 download. I also don't get these additional "[...] would like to access data from other apps." dialogs. (only once asking to confirm: '“Pd-0.54-1-1-gb12dcbcb.app” is an app downloaded from the internet. Are you sure you want to open it?')
But then Pd-0.54 should be affected as well (unless, maybe, wildly guessing, Pd-0.54 was installed with an older version of macOS, and the os upgrade kept existing permissions)
Just to be (kind of) sure that this can't be connected to any remains from the original Pd-0.54, I also installed an old version that quite certainly had never been on my system yet: 0.52-2 ... that also doesn't give me these warnings. A quick search shows that this dialog is actually a Sonoma "feature" - but it obviously seems weird that it's not triggered for previous Pd versions.
Cheers, Ben
On Sat, Jun 1, 2024 at 11:29 AM Miller Puckette mpuckette@cloud.ucsd.edu wrote:
OK, next thing I want to see is what happens if I recompile Pd 0.54-1... I just created a branch to see if I can get the IEM compile chain to grind it out anew.
Hopping on an airplane soon but will be about for 1/2 hour or so I think.
cheers
M
On 6/1/24 11:06 AM, IOhannes m zmölnig wrote:
Am 1. Juni 2024 10:56:42 MESZ schrieb "IOhannes m zmölnig" zmoelnig@iem.at:
Am 1. Juni 2024 10:50:19 MESZ schrieb Benjamin Wesch benjamin.wesch@gmail.com: installed
i also got the dialog that william mentions though. could this be connected?
Hmm. I've never seen these.
Most likely is a new feature of iOS/macOS 14. But then Pd-0.54 should be affected as well (unless, maybe, wildly guessing, Pd-0.54 was installed with an older version of macOS, and the os upgrade kept existing permissions)
mfg.sfg.jfd IOhannes
Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> https://urldefense.com/v3/__https://lists.puredata.info/listinfo/pd-list__;!...
Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> https://lists.puredata.info/listinfo/pd-list
On 6/1/24 20:57, Miller Puckette wrote:
Thanks - so it's clearly something about the new code...
+1
I wish I could get the slowdown to happen on my machine so I could monkey with this myself :)
yep.
i might have missed it, but does anybody of the users who experience the slowdown also see this if they compile Pd themselves?
now that (I think) we've established that the problem occurs only on macOS 14 "Sonoma", I will try to reproduce it with a Sonoma VM when i'm back at the uni on monday.
once we've found a way to create and test build with faster turnaround times than via the CI, it should be straightforward to bisect the problem.
gfamrd IOhannes
On Sat, Jun 1, 2024 at 10:04 PM IOhannes m zmölnig zmoelnig@iem.at wrote:
i might have missed it, but does anybody of the users who experience the slowdown also see this if they compile Pd themselves?
i didn't mention it, but tested with a local build here: getting quite exactly the same results and even the warning dialog.
ran these tests with that build once more now: /Applications/Pd-0.55-0.app/Contents/Resources/bin/pd -noverbose -send 0.02s user 0.02s system 2% cpu 1.460 total /Applications/Pd-0.55-0.app/Contents/Resources/bin/pd -noverbose -nosound 0.02s user 0.02s system 2% cpu 1.453 total /Applications/Pd-0.55-0.app/Contents/Resources/bin/pd -noverbose -noprefs 0.01s user 0.02s system 2% cpu 1.418 total /Applications/Pd-0.55-0.app/Contents/Resources/bin/pd -noverbose -nostdpath 0.02s user 0.02s system 2% cpu 1.493 total
cheers, ben
On 6/1/24 22:26, Benjamin Wesch wrote:
On Sat, Jun 1, 2024 at 10:04 PM IOhannes m zmölnig zmoelnig@iem.at wrote:
i might have missed it, but does anybody of the users who experience the slowdown also see this if they compile Pd themselves?
i didn't mention it, but tested with a local build here: getting quite exactly the same results and even the warning dialog.
cool (that you are able to produce the problem with self-built binaries).
if you know how to use git bisect, youcould try to narrow down the
commit that introduced the slowdown.
gfmsd IOhannes
On Sat, Jun 1, 2024 at 11:27 PM IOhannes m zmölnig zmoelnig@iem.at wrote:
if you know how to use
git bisect, you could try to narrow down the commit that introduced the slowdown.
this is new to me, i admit - but hoping that i did it correctly, i now get: 2fe920f88abb814fc89719a617ce48d63a3eae3a is the first bad commit
cheers, ben
was curious now and also tested the current master without this single commit. warning dialogs are all gone and measurements look incredible:
Benchmark 1: Pd-0.55.0-maybelessbad.app/Contents/Resources/bin/pd -send "pd quit" Time (mean ± σ): 240.7 ms ± 5.9 ms [User: 17.3 ms, System: 16.2 ms] Range (min … max): 224.7 ms … 246.9 ms 12 runs
so the question now probably is what we're missing without the "bad commit"?
On Sun, Jun 2, 2024 at 12:32 AM Benjamin Wesch benjamin.wesch@gmail.com wrote:
On Sat, Jun 1, 2024 at 11:27 PM IOhannes m zmölnig zmoelnig@iem.at wrote:
if you know how to use
git bisect, you could try to narrow down the commit that introduced the slowdown.this is new to me, i admit - but hoping that i did it correctly, i now get: 2fe920f88abb814fc89719a617ce48d63a3eae3a is the first bad commit
cheers, ben
Am 2. Juni 2024 02:19:46 MESZ schrieb Benjamin Wesch benjamin.wesch@gmail.com:
this is new to me, i admit - but hoping that i did it correctly, i now get: 2fe920f88abb814fc89719a617ce48d63a3eae3a is the first bad commit
Thanks a lot for finding this! For whatever reasons, an application is obviously not allowed to read its own metainfo without penalty. I might be able to submit a fix for this in the evening (now it's early morning here) If somebody is faster, then please do not just revert the bad commit, but instead just replace the dynamic reading of ::pd_guiprefs::domain in https://github.com/pure-data/pure-data/blob/020a71901/tcl/pd_guiprefs.tcl#L72-76 to
set ::pd_guiprefs::domain org.puredata.pd.pd-gui
(or probably just remove the lines 72-76, as this *should* be handled by fallbacks; I'm currently afk so my view is a bit narrow)
mfg.sfg.jfd IOhannes
On Sun, Jun 2, 2024 at 8:45 AM IOhannes m zmölnig zmoelnig@iem.at wrote:
[...] replace the dynamic reading of ::pd_guiprefs::domain
without really understanding the whole mechanism: isn't this already defined here then? https://github.com/pure-data/pure-data/blob/master/tcl/pd_guiprefs.tcl#L21
or probably just remove the lines 72-76
both options work for me.
cheers, ben
On 6/2/24 09:34, Benjamin Wesch wrote:
On Sun, Jun 2, 2024 at 8:45 AM IOhannes m zmölnig zmoelnig@iem.at wrote:
[...] replace the dynamic reading of ::pd_guiprefs::domain
without really understanding the whole mechanism: isn't this already defined here then? https://github.com/pure-data/pure-data/blob/master/tcl/pd_guiprefs.tcl#L21
yes. (as i said: there should be some fallback mechanisms in place).
or probably just remove the lines 72-76
both options work for me.
thanks for confirming. i've created a PR [2331] for the fix. CI-builds should be available soon.
On Sun, Jun 2, 2024 at 10:42 AM IOhannes m zmölnig zmoelnig@iem.at wrote:
CI-builds should be available soon.
thanks! for the sake of completeness, here are my (reformatted) hyperfine results for the CI build vs. last test version, including the quitter-plugin.tcl variant. happy that this seems to be solved!
| Version | Parameters | Mean [ms] | |--------------------------+-------------------------+----------------| | 0.55-0test3a | -send "pd quit" -nogui | 59.8 ± 0.6 | | 0.55-0test3a-1-g85556737 | -send "pd quit" -nogui | 59.7 ± 0.6 | | 0.55-0test3a | -send "pd quit" | 1428.8 ± 19.5 | | 0.55-0test3a-1-g85556737 | -send "pd quit" | 218.7 ± 4.1 | | 0.55-0test3a | -path quitter/ | 1579.8 ± 11.0 | | 0.55-0test3a-1-g85556737 | -path quitter/ | 380.9 ± 9.6 |
(all tests ran 3 warmups and include parameters "-noprefs -nostdpath")
cheers, ben
Thanks for tracking this down... git-bisect can be tricky and/or time consuming!
Miller
On 6/2/24 2:19 AM, Benjamin Wesch wrote:
was curious now and also tested the current master without this single commit. warning dialogs are all gone and measurements look incredible:
Benchmark 1: Pd-0.55.0-maybelessbad.app/Contents/Resources/bin/pd -send "pd quit" Time (mean ± σ): 240.7 ms ± 5.9 ms [User: 17.3 ms, System: 16.2 ms] Range (min … max): 224.7 ms … 246.9 ms 12 runs
so the question now probably is what we're missing without the "bad commit"?
On Sun, Jun 2, 2024 at 12:32 AM Benjamin Wesch benjamin.wesch@gmail.com wrote:
On Sat, Jun 1, 2024 at 11:27 PM IOhannes m zmölnig zmoelnig@iem.at wrote:
if you know how to use
git bisect, you could try to narrow down the commit that introduced the slowdown.this is new to me, i admit - but hoping that i did it correctly, i now get: 2fe920f88abb814fc89719a617ce48d63a3eae3a is the first bad commit
cheers, ben
Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> https://urldefense.com/v3/__https://lists.puredata.info/listinfo/pd-list__;!...
And here are my results running those tests:
/Applications/Pd-0.55-0test3a.app/Contents/Resources/bin/pd -send "pd quit" 0.02s user 0.03s system 1% cpu 2.987 total
time /Applications/Pd-0.55-0test3a.app/Contents/Resources/bin/pd -nosound -send "pd quit" 0.02s user 0.03s system 1% cpu 2.976 total
time /Applications/Pd-0.55-0test3a.app/Contents/Resources/bin/pd -noprefs -send "pd quit" 0.02s user 0.03s system 1% cpu 3.150 total
time /Applications/Pd-0.55-0test3a.app/Contents/Resources/bin/pd -nostdpath -send "pd quit" 0.02s user 0.03s system 1% cpu 3.073 total
The same tests with 54-1 give me cpu times of about 0.35sec. And (probably unrelated), with 55-0, I get "error: error parsing startup arguments" when running each of these despite the arguments appearing to have their effect. With 54-1 I don't get the error post.
On Sat, Jun 1, 2024 at 1:33 AM Miller Puckette mpuckette@cloud.ucsd.edu wrote:
Hmm, sounds like there are more than one problem...
Can you (and Ben too if he's willing) try these...
time ..../Pd-0.55-0test2c.app/Contents/Resources/bin/pd -send "pd quit"
time ..../Pd-0.55-0test2c.app/Contents/Resources/bin/pd -nosound -send "pd quit"
time ..../Pd-0.55-0test2c.app/Contents/Resources/bin/pd -noprefs -send "pd quit"
time ..../Pd-0.55-0test2c.app/Contents/Resources/bin/pd -nostdpath -send "pd quit"
I'm grasping at straws here :) Guessing that William's 8-second startup and new popup might be something external that's fighting with some Mac plist setting in Pd that might have changed.
thanks M
On 6/1/24 12:12 AM, William Brent wrote:
Thanks for this new release! Unfortunately, I can confirm the extra startup delay for macOS... I tested on an M1 machine with 64GB of memory running macOS 14.5.
Also, I get the attached popup every time I open the app, even after clicking "allow." Maybe there's something in macOS preferences that will suppress it? After clicking "allow," it takes a full 8sec for the Pd window to appear. I've only timed this casually with my watch at this point, but after trying multiple times I can say that it's consistently around 8sec for me here. Pd-0.54-1 consistently springs to life in about 1sec for me.
I'm happy to run any specific test routines if I can be useful on this. William
On Fri, May 31, 2024 at 1:10 PM Miller Puckette mpuckette@cloud.ucsd.edu wrote:
Test 3 is out. This should resolve all the trouble we've had with test2 except that one tester on MacOS is still finding that Pd 0.55 loads way more slowly than 0.54 - I don't see that happening on my machine and I don't know how many machines will be affected... it might take a while to track this one down, but hopefully it won't affect many people. If you _do_ notice longer startup times on MacOS please let me (or the Pd list) know - with more data points it migth be easier for me to replicate the problem for myself. as usual: http://msp.ucsd.edu/software.htm OR https://msp.ucsd.edu/software.htm source on github: https://github.com/pure-data/pure-data <https://urldefense.com/v3/__https://github.com/pure-data/pure-data__;!!Mih3w...
cheers Miller _______________________________________________ Pd-announce mailing list Pd-announce@lists.iem.at https://lists.puredata.info/listinfo/pd-announce <https://urldefense.com/v3/__https://lists.puredata.info/listinfo/pd-announce...
_______________________________________________ Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> https://lists.puredata.info/listinfo/pd-list <https://urldefense.com/v3/__https://lists.puredata.info/listinfo/pd-list__;!...
-- William Brent
“Great minds flock together” Conflations: conversational idiom for the 21st century
www.conflations.com <
https://urldefense.com/v3/__http://www.conflations.com__;!!Mih3wA!GVjycnnXvx...
Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management ->
https://urldefense.com/v3/__https://lists.puredata.info/listinfo/pd-list__;!...
Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> https://lists.puredata.info/listinfo/pd-list
-
Alexandre Torres Porres -
Benjamin Wesch -
IOhannes m zmölnig -
Miller Puckette -
William Brent