purr-data issueshttps://git.purrdata.net/jwilkes/purr-data/-/issues2020-07-30T18:01:43Zhttps://git.purrdata.net/jwilkes/purr-data/-/issues/660Vertical viewport size regression in Ico's branch (nw.js 0.24.4)2020-07-30T18:01:43ZAlbert GräfVertical viewport size regression in Ico's branch (nw.js 0.24.4)@ico, unfortunately, it seems that I have found another, less obvious, "old-nw.js" regression in your branch. Using nw.js 0.24.4 on Linux (Manjaro), the viewport of a fully visible canvas (i.e., without scrollbars, and with the entire ca...@ico, unfortunately, it seems that I have found another, less obvious, "old-nw.js" regression in your branch. Using nw.js 0.24.4 on Linux (Manjaro), the viewport of a fully visible canvas (i.e., without scrollbars, and with the entire canvas fitting well into the visible area) seems slightly too large (at least in the vertical direction). The observable result is that, even though the window has no scrollbars, you can wiggle the mousewheel and have the window scroll a little amount in the vertical direction. You can easily reproduce this using the "about" window with the current master (rev. dd632b2d294969e2740b6fad91ff7ba6bb26fb1f) built against nw.js 0.22.4. The following little screencast shows what I mean:
![vertical-viewport-wiggle](/uploads/7b60140b5826e0b8c7d1350f05d40b49/vertical-viewport-wiggle.gif)
I can make this go away by going back to rev. 11f6610d31c8e656e3e593701e83337368cdbd91 (before your branch got merged back in). The problem also goes away with the current master, when using nw.js 0.47.0-beta1 instead of our baseline 0.24.4. So it seems fairly obvious to me that it's in fact a regression in your branch with respect to nw.js 0.24.4.Ivica BukvicIvica Bukvichttps://git.purrdata.net/jwilkes/purr-data/-/issues/659Better touchscreen support2020-07-17T21:09:19ZAlbert GräfBetter touchscreen supportOur current OOTB experience on touch screens is not very good, to put it kindly. :) You can select menu entries and click on objects, that's about it. But if you're using a touchscreen, you also want to be able to operate GUI objects, dr...Our current OOTB experience on touch screens is not very good, to put it kindly. :) You can select menu entries and click on objects, that's about it. But if you're using a touchscreen, you also want to be able to operate GUI objects, drag around objects, draw a rubberband to select a range of objects, etc., which currently doesn't work at all with the default touch support that the OS and/or nw.js offer.
This probably isn't straightforward, because mouse event handling needs to be modified in our GUI code. But having proper touch support would certainly be a big improvement, if not a killer feature, especially when running Purr Data on convertibles, as well as in teaching on a digital whiteboard (I actually have one of those in our MIDI lab, which works without a hitch on Linux). So can we please have a go at this?
I have no idea how to do this myself, but over at my Purr Data Github mirror, spidercatnat has submitted a [pull request](https://github.com/agraef/purr-data/pull/15), still WIP, which already has the basics implemented. ATM, this still has some bugs and is based on a pretty old (around 2.10.0) revision, but it rebases easily on the current master, you can find that in my [copy of spidercatnat's branch](https://github.com/agraef/purr-data/tree/touch-support-master) on Github (of course, I can pull that over to Gitlab if anyone here wants to play around with it).
It would be great if someone who knows this stuff better than me could lend spidercatnat a helping hand, so that we can make this work. (I suspect that it's Ico's branch which gives trouble trying to merge spidercatnat's work with HEAD right now. But that's hopefully not a big deal. I'm willing to do the grunt integration work, with a bit of help from Ico if needed.)Albert GräfAlbert Gräfhttps://git.purrdata.net/jwilkes/purr-data/-/issues/654metro / delay help files outdated (tempo message missing)2020-07-05T12:53:56ZAlexandre Porresmetro / delay help files outdated (tempo message missing)Hi, as of 0.45, metro, delay and timer take a "tempo" message. These have been implemented in purr data but are still missing in the help file of metro and delay (timer has it). I'd suggest a revision of the docs to find/fix more issues....Hi, as of 0.45, metro, delay and timer take a "tempo" message. These have been implemented in purr data but are still missing in the help file of metro and delay (timer has it). I'd suggest a revision of the docs to find/fix more issues. Cheers.https://git.purrdata.net/jwilkes/purr-data/-/issues/645The bar graph array scalars inside a subpatch (topwindow) look like a flat li...2020-06-17T19:06:37ZIvica BukvicThe bar graph array scalars inside a subpatch (topwindow) look like a flat line out of which bars grow out of both up and down, rather than appearing from the bottom of the windowI will look into this eventually. I wonder though, if having Jonathan do this since the bargraph is his baby may be easier. Please also note the issue #643 that is also related to this bug but should be fixed globally for all plots in a ...I will look into this eventually. I wonder though, if having Jonathan do this since the bargraph is his baby may be easier. Please also note the issue #643 that is also related to this bug but should be fixed globally for all plots in a separate merge request.Jonathan WilkesJonathan Wilkeshttps://git.purrdata.net/jwilkes/purr-data/-/issues/644The bar graph array scalars inside a subpatch (topwindow) look like a flat li...2020-06-09T01:38:31ZIvica BukvicThe bar graph array scalars inside a subpatch (topwindow) look like a flat line out of which bars grow out of both up and down, rather than appearing from the bottom of the windowI will look into this eventually. I wonder though, if having Jonathan do this since the bargraph is his baby may be easier. Please also note the issue #643 that is also relevant to this bug but should be fixed globally for all plots.I will look into this eventually. I wonder though, if having Jonathan do this since the bargraph is his baby may be easier. Please also note the issue #643 that is also relevant to this bug but should be fixed globally for all plots.https://git.purrdata.net/jwilkes/purr-data/-/issues/643When a GOP subpatch with an array or scalars is open, the array/scalars do no...2020-08-25T03:31:33ZIvica BukvicWhen a GOP subpatch with an array or scalars is open, the array/scalars do not scale across the entire window sizeIvica BukvicIvica Bukvichttps://git.purrdata.net/jwilkes/purr-data/-/issues/640Improving scrollbar appearance2021-03-17T23:15:38ZIvica BukvicImproving scrollbar appearanceThe following css suggests an appearance that is more in line with the 1.x. I would further add the semitransparency to minimize occlusion of the content behind the scrollbars:
http://jsfiddle.net/6KprJ/1/The following css suggests an appearance that is more in line with the 1.x. I would further add the semitransparency to minimize occlusion of the content behind the scrollbars:
http://jsfiddle.net/6KprJ/1/https://git.purrdata.net/jwilkes/purr-data/-/issues/639Zexy not included as part of the default libraries2020-06-05T22:23:24ZIvica BukvicZexy not included as part of the default librariesBreaks pd-l2ork 1.x compatibility and the K12 modeBreaks pd-l2ork 1.x compatibility and the K12 modehttps://git.purrdata.net/jwilkes/purr-data/-/issues/623When having 2 instances of the same GOP abstraction presents several issues2020-07-06T14:23:36ZIvica BukvicWhen having 2 instances of the same GOP abstraction presents several issues1) selecting one highlights all objects inside it when it shouldn't (only the outer edges should be selected according to 1.x and vanilla behavior
2) selecting one creates selection around the other which also has visible text (even if ...1) selecting one highlights all objects inside it when it shouldn't (only the outer edges should be selected according to 1.x and vanilla behavior
2) selecting one creates selection around the other which also has visible text (even if it is disabled in the text) which messes up the bbox/scrollbar calculation
Windows 2.10https://git.purrdata.net/jwilkes/purr-data/-/issues/610Some objects when selected are highlighted in blue (pd vanilla), others in or...2020-06-10T04:50:18ZIvica BukvicSome objects when selected are highlighted in blue (pd vanilla), others in orange (pd-l2ork). This may need to be made consistent. Comments are also hard to spot when selected on darker backgrounds since their edges are not highlighted like other objects.Windows 2.10Windows 2.10https://git.purrdata.net/jwilkes/purr-data/-/issues/601Pressing arrows when object is selected should move object not scroll the canvas2020-06-10T04:51:13ZIvica BukvicPressing arrows when object is selected should move object not scroll the canvas2.10, windows. I imagine windows scrolling would be appropriate if no object is selected.2.10, windows. I imagine windows scrolling would be appropriate if no object is selected.https://git.purrdata.net/jwilkes/purr-data/-/issues/590array name should be inside the array to provide a more predictable bbox calc...2020-05-24T16:06:00ZIvica Bukvicarray name should be inside the array to provide a more predictable bbox calculation and keep consistency with 1.0 branchKeeping it where it is now limits array's utility for visualization purposes--its bbox detection is not accurate (at least not in vanilla) as it does not factor in the label. Perhaps this should be one of the additions to the -legacy fla...Keeping it where it is now limits array's utility for visualization purposes--its bbox detection is not accurate (at least not in vanilla) as it does not factor in the label. Perhaps this should be one of the additions to the -legacy flag to support both vanilla and new behavior, even though outside the bbox inconsistency it should not have any major impact on the compatibility/operability. Long-term-wise this may be also something that can be repositioned?https://git.purrdata.net/jwilkes/purr-data/-/issues/586list cat clear command not clearing the list2020-03-28T17:24:28ZIvica Bukviclist cat clear command not clearing the listTitle says it all. Use the help file to test.
I seem to recall Jonathan doing some optimizations on the object a while ago being unsure why something was implemented the way it was. This may have been a side-effect of that change as the...Title says it all. Use the help file to test.
I seem to recall Jonathan doing some optimizations on the object a while ago being unsure why something was implemented the way it was. This may have been a side-effect of that change as the clear works fine in 1.0 branch.https://git.purrdata.net/jwilkes/purr-data/-/issues/585Purr Data is not for real-time MIDI and Audio projects!2020-05-04T15:41:26ZJoseph GastelaisPurr Data is not for real-time MIDI and Audio projects!Hello,
I know this issue title is a little bit provocative but it's the true reality I'm facing to. Am I the only one?
Thanks for releasing Purr Data v.2.10.1 (with Pd 0.48.0 & NW.js 0.24.4). For the record, Pd Vanilla is actually at v...Hello,
I know this issue title is a little bit provocative but it's the true reality I'm facing to. Am I the only one?
Thanks for releasing Purr Data v.2.10.1 (with Pd 0.48.0 & NW.js 0.24.4). For the record, Pd Vanilla is actually at v.0.50.2 and NW.js at v.0.44.5.
1- Despite the recent release of its last version, this open-source development software is not meeting yet its declaration of intention:
"Purr Data serves the same purpose than Pd (Pure Data); i.e. a graphical data-flow programming environment which is geared towards real-time interactive computer music and multimedia applications but offers a new and much improved graphical user interface and includes many 3rd party plug-ins."
2- After few months of development, I arrived to the (temporary) conclusion that unfortunately I cannot use Purr Data for any serious real-time interactive computer music and audio projects. The real-time problems were increasing as my project was getting bigger even when the sub-patchs/GOPs are in idle mode.
These major issues seem to be linked to the combination of several elements, at least:
- Purr Data cannot handle simply several CPU in parallel when using only one patch contending several sub-patchs/GOPs.
- Purr Data modern GUI is too much CPU-consuming, due to at least the usage of only one CPU.
- Purr Data is not able to properly handle the Mouse, even when just moving it (with no click, no drag, no scroll).
- Purr Data is an old turtle when using Copy/Paste or Duplicate or Moving just few tens of objects in EditMode.
Please, have a look at some of my previous posts regarding these major issues:
- #581 - EditMode / Reliability Issues
- #555 - Thread 1 "purr-data" received signal SIGSEGV
- #553 - Copy/Paste / Duplicate / Move Weird behavior
And see also for MIDI objects:
- #563 - Support for 'Cyclone' 0.4
- #554 - Cyclone v.0.2 issues with MIDI
- #355 - 9 Cyclone objects that still need to be ported to Purr Data
3- After few weeks of heavy bugs testing of my MIDI/Audio project and Purr Data, it is now clear to me that this development tool has a lot of major and ongoing design issues which are not fixed yet, if they could be fixed.
I'm giving hereafter 3 new, significant and reproducible at will examples (under Linux). It's the same very bad behaviour under GNU/Linux (Linux Mint 19.3) with/without JACK as well as under Windows 10 (v.1909) with/without JACK or ASIO -and- DSP = OFF. No error or warning mentioned in the Console.
Purr Data 2.10.1 / All GOPs = IDLE (except [pd dt]) => Moving my Mouse => see the very unexpected impact on CPU load:
a- with EditMode = OFF (Purr Data + Task Manager together)
![mdre+-0321_prd-em-off](/uploads/4a35193cedd7d3641c40db83e7bd6779/mdre+-0321_prd-em-off.jpg)
b- with EditMode = OFF (Purr Data + Htop alone)
![mdre+-0321_prd-htop](/uploads/d8f86d53c13b56b654ccee1e511dcd64/mdre+-0321_prd-htop.jpg)
c- with EditMode = ON (Purr Data + Task Manager together)
![mdre+-0321_prd-em-on](/uploads/9acd88e71265c97dd88bf16674115052/mdre+-0321_prd-em-on.jpg)
And then when I'm effectively running (with EditMode = Off) such or such GOP(s), the situation is getting worse and worse, leading very quickly to contentious erroneous MIDI values, audio glitches from short to long, random freezes of few seconds and finally totally freezing the whole project by just moving the mouse over the GOPs. This is an undocumented and hidden feature of Purr Data.
Note that I cannot test my project under Pd Vanilla as it is loading it but with too many errors, just displaying about half of the GOPs and these GOPs are not functional even for stand-alone ones, with the required externals well loaded.
4- An example of Warnings messages in the Console after having closed my 'MDR-E+' project (with just Open / No use of the GOPs / Close)
`[4] image: warning: no image data in cache to free
legacy tcl command at 201 of ../shared/hammer/file.c: hammereditor_close .93778a40 0
legacy tcl command at 201 of ../shared/hammer/file.c: hammereditor_close .9377a630 0
legacy tcl command at 201 of ../shared/hammer/file.c: hammereditor_close .9377c220 0
legacy tcl command at 201 of ../shared/hammer/file.c: hammereditor_close .9377de10 0
legacy tcl command at 201 of ../shared/hammer/file.c: hammereditor_close .9377e130 0
legacy tcl command at 201 of ../shared/hammer/file.c: hammereditor_close .9377e450 0
legacy tcl command at 201 of ../shared/hammer/file.c: hammereditor_close .9377e770 0
legacy tcl command at 201 of ../shared/hammer/file.c: hammereditor_close .9377ea90 0
legacy tcl command at 201 of ../shared/hammer/file.c: hammereditor_close .9377edb0 0
legacy tcl command at 201 of ../shared/hammer/file.c: hammereditor_close .93798e00 0
legacy tcl command at 201 of ../shared/hammer/file.c: hammereditor_close .9379a9f0 0
legacy tcl command at 201 of ../shared/hammer/file.c: hammereditor_close .9379c5e0 0
legacy tcl command at 201 of ../shared/hammer/file.c: hammereditor_close .9379e1d0 0
legacy tcl command at 201 of ../shared/hammer/file.c: hammereditor_close .9379fdc0 0
legacy tcl command at 201 of ../shared/hammer/file.c: hammereditor_close .937a19b0 0
legacy tcl command at 201 of ../shared/hammer/file.c: hammereditor_close .937a35a0 0
legacy tcl command at 201 of ../shared/hammer/file.c: hammereditor_close .93805260 0
legacy tcl command at 201 of ../shared/hammer/file.c: hammereditor_close .9381b9c0 0
legacy tcl command at 201 of ../shared/hammer/file.c: hammereditor_close .9383c470 0
error: x55839318a5d0: no such object`
5- Reading again the useful "Meet the Cat: A Quick Introduction to Purr Data" written by Albert Gräf didn't help me.
So, for the time being, I put with regret my 'MDR-E+' project developed with Purr Data in standby (confinement!) until this situation could be clarified... or not.
Don't get me wrong. Purr Data is a great piece of open-source software and many thanks to the team working on it. It's viable as long as you are using it only for small projects like 'Simple MIDI Sequencer', my 2nd project under Purr Data. But as soon as you start to use it for bigger projects (but still reasonable), like my 'Music Disk Record Emulator+' (an emulation of a 30 years old Yamaha MIDI hardware) its MIDI and Audio real-time limitations become more than obvious.
And for me, Pure Data Vanilla is not a viable alternative due to its bad looking and obsolete GUI.
Best,https://git.purrdata.net/jwilkes/purr-data/-/issues/568JGU packages for Purr Data 2.10.02019-12-04T22:47:20ZJoseph GastelaisJGU packages for Purr Data 2.10.0Hello Albert,
I'm a little bit confused. I have just discovered that v.2.10.0 is available since 2 weeks but no notification received in my Package Manager under Linux Mint 19.2 for updating my v.2.9.0. I though something was broken on ...Hello Albert,
I'm a little bit confused. I have just discovered that v.2.10.0 is available since 2 weeks but no notification received in my Package Manager under Linux Mint 19.2 for updating my v.2.9.0. I though something was broken on my PC.
Trying to understand what is happening, I have discovered that JGU packages were moved for the Debian/Ubuntu package to repositories on SuSE's Open Build Service (OBS) from your PPA on Launchpad.net. I have surely missed something!? Am I alone?
Question: What is the right procedure for replacing PPA with OBS on my PC, just remove PPA and replace it with OBS. And also do I have before to uninstall Purr Data 2.9.0?
Thank you. Best, Josephhttps://git.purrdata.net/jwilkes/purr-data/-/issues/562[text sequence] with -t flag crashes purr-data2019-11-28T14:06:57ZH. James Harkins[text sequence] with -t flag crashes purr-dataNew patch, create `[text sequence name -t 120 permin]` or any tempo and units. Crash bang.
Works fine in Pd vanilla.New patch, create `[text sequence name -t 120 permin]` or any tempo and units. Crash bang.
Works fine in Pd vanilla.https://git.purrdata.net/jwilkes/purr-data/-/issues/560Save dialog sometimes hangs2019-11-13T01:53:46ZH. James HarkinsSave dialog sometimes hangsTwice in the last two or three days, it's gone like this:
1. Create a new patch and put some objects in.
2. Hit Ctrl-S to save.
3. File dialog comes up, but is completely unresponsive.
*Usually* the save dialog works fine. That it fail...Twice in the last two or three days, it's gone like this:
1. Create a new patch and put some objects in.
2. Hit Ctrl-S to save.
3. File dialog comes up, but is completely unresponsive.
*Usually* the save dialog works fine. That it failed once would have been a fluke. Twice is worrisome.
The first time this happened, I was able to use the XFCE task manager to right-click on the purr-data process and kill it.
The second time (just now) I could open the task manager by keyboard, but right-click was unresponsive there too (?). But I was able to switch to a different workspace and click on an open terminal window and do `killall purr-data`.
Purr-data has been quite stable for the last few months and only recently seemed to go sideways in this way. It might be due to an Ubuntu Studio update (?).
Ubuntu Studio 18.04, XFCE 4.12, Pd-l2ork 2.9.0 20190819.https://git.purrdata.net/jwilkes/purr-data/-/issues/544Windows/msys2 compilation problems2019-07-30T20:29:46ZAlbert GräfWindows/msys2 compilation problemsOn the latest msys2/mingw (fully updated) I'm getting this error:
~~~
cd /home/ag/purr-data/externals/OSCx && ./configure
configure: loading site script /mingw32/etc/config.site
checking build system type... i686-w64-mingw32
checking ho...On the latest msys2/mingw (fully updated) I'm getting this error:
~~~
cd /home/ag/purr-data/externals/OSCx && ./configure
configure: loading site script /mingw32/etc/config.site
checking build system type... i686-w64-mingw32
checking host system type... i686-w64-mingw32
checking for gcc... gcc
checking whether the C compiler works... no
configure: error: in `/home/ag/purr-data/externals/OSCx':
configure: error: C compiler cannot create executables
See `config.log' for more details
make[3]: *** [Makefile:1362: /home/ag/purr-data/externals/OSCx/Makefile] Error 77
make[3]: Leaving directory '/home/ag/purr-data/externals'
make[2]: *** [Makefile:209: externals_install] Error 2
make[2]: Leaving directory '/home/ag/purr-data/packages'
make[1]: *** [Makefile:69: prebuilt_install] Error 2
make[1]: Leaving directory '/home/ag/purr-data/packages/win32_inno'
~~~
config.log shows:
~~~
configure:2634: gcc -DPD -I/home/ag/purr-data/pd/src -Wall -W -Wno-unused-parameter -ggdb -I/home/ag/purr-data/Gem/src -mms-bitfields -DMSW -DNT -D'O_NONBLOCK=1' -D'srand48(n)=srand((n))' -D'drand48()=((double)rand()/RAND_MAX)' -D'bzero(p,n)=memset(p,0,n)' conftest.c >&5
<command-line>: error: macro names must be identifiers
<command-line>: error: macro names must be identifiers
<command-line>: error: macro names must be identifiers
<command-line>: error: macro names must be identifiers
configure:2638: $? = 1
configure:2676: result: no
~~~
This is 100% reproducible for me. I'm not sure why we don't see these in the Windows runner, maybe it uses an older version of the msys2 base system?
Anyway, I can get rid of this error simply by removing the definitions `-D'srand48(n)=srand((n))'` and `-D'drand48()=((double)rand()/RAND_MAX)' -D'bzero(p,n)=memset(p,0,n)'` from the `WINDOWS_HACKS` variable in externals/Makefile, i.e.:
~~~diff
diff --git a/externals/Makefile b/externals/Makefile
index a193a457..ef699e9c 100644
--- a/externals/Makefile
+++ b/externals/Makefile
@@ -73,8 +73,9 @@ ifeq ($(OS_NAME),linux)
endif
ifeq ($(OS_NAME),windows)
BUILDSRC_OS_NAME = win
- WINDOWS_HACKS = -D'O_NONBLOCK=1' -D'srand48(n)=srand((n))' \
- -D'drand48()=((double)rand()/RAND_MAX)' -D'bzero(p,n)=memset(p,0,n)'
+ WINDOWS_HACKS = -DO_NONBLOCK=1
+# -D'srand48(n)=srand((n))' \
+# -D'drand48()=((double)rand()/RAND_MAX)' -D'bzero(p,n)=memset(p,0,n)'
# These don't seem to be needed:
# -D'PROT_READ=1' \
# -D'MAP_PRIVATE=2' \
~~~
Should I submit this as a merge request so that we can see whether this breaks the Windows runner? It goes without saying that having to patch up the sources each time I try to build the Windows version is a bit inconvenient, and other users might run into this issue as well, thus the sooner we can get rid of these Windows-specific hacks the better.Jonathan WilkesJonathan Wilkeshttps://git.purrdata.net/jwilkes/purr-data/-/issues/542Building on macOS 10.14 Mojave2019-07-25T18:06:08Znerronsconiinedieu@gmail.comBuilding on macOS 10.14 MojaveXcode 10 doesn't build Purr Data for some reason. Go get Command Line Tools (macOS 10.13) for Xcode 9.4.1 on https://developer.apple.com/download/more/.
Seems like there are changes in the locations of the libraries so `configure` can't...Xcode 10 doesn't build Purr Data for some reason. Go get Command Line Tools (macOS 10.13) for Xcode 9.4.1 on https://developer.apple.com/download/more/.
Seems like there are changes in the locations of the libraries so `configure` can't find dependencies. Also rumors are that 10.14 stopped supporting anything with `-mmacosx-version-min` less than 10.9, but by using the old Command Line Tools things are working out just fine.
Wasted a lot of time on this, so I just wrote it down in case anyone upgrades on impulse like me...https://git.purrdata.net/jwilkes/purr-data/-/issues/537Supporting Ubuntu 18.042019-05-08T09:59:24ZPatrickSupporting Ubuntu 18.04Would it be possible to add support for Ubuntu 18.04?
The best audio distribution, in my humble opinion, is [Ubuntu Studio](https://ubuntustudio.org). Would be nice to try to follow the release cycle (same as Ubuntu I think).Would it be possible to add support for Ubuntu 18.04?
The best audio distribution, in my humble opinion, is [Ubuntu Studio](https://ubuntustudio.org). Would be nice to try to follow the release cycle (same as Ubuntu I think).