Hi,
I know @hpcoder is busy, and there's not much development going on on Scidavis at the moment, so I don't expect any new release soon with new features.
I've seen there's been some commits related to bug fixes since the last release (2.3.0). I'm personally interested on bug 400, related to exporting tables as ASCII, which is currently not working, but I imagine other people might be interested on other bug fixes.
I wonder if it's posible to release a new binary point release (2.3.1), to include these. I'm not a very experience user in compiling my own sofware, and I use SciDAVIs on Linux and Windows, so it'll be much easier for me to get updated binaries.
Thank you
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I'll have a go at doing a release shortly (it would be numbered 2.4.0
- 2.3.1 would entail no source code difference between it and 2.3.0 :)
Windows and Linux builds should be straight forward - MacOSX is still
a problem, however...
On Tue, Dec 29, 2020 at 12:15:17PM -0000, Miguel Agueda wrote:
Hi,
I know @hpcoder is busy, and there's not much development going on on Scidavis
at the moment, so I don't expect any new release soon with new features.
I've seen there's been some commits related to bug fixes since the last release
(2.3.0). I'm personally interested on bug 400, related to exporting tables as
ASCII, which is currently not working, but I imagine other people might be
interested on other bug fixes.
I wonder if it's posible to release a new binary point release (2.3.1), to
include these. I'm not a very experience user in compiling my own sofware, and
I use SciDAVIs on Linux and Windows, so it'll be much easier for me to get
updated binaries.
Great! I'm happy to hear that.
I did manage to compile it myself yesterday. I couldn't do it on Fedora that I use, because I got an error about lupdate not being avaliable, but I manage to compile on Ubuntu 18.04 without python and then copy it over Fedora and link scidavis to the needed libraries on the system.
It's kind of janky, and I'm not confident it'll work very well, but I can use it to export the tables.
2.3.1 would entail no source code difference between it and 2.3.0
Okay, I see what you mean. I wasn't sure about how it worked, and I thought maybe the fixes weren't substantial enough to justify a new release.
Thanks for the help
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Just a brief update on release plans. Chinoune Mehdi has been doing
some excellent work towards moving SciDAVis from the dreadful qmake
system to the much more capable cmake system, as well as building qwt5
and qwtplot3d as git submodules, which solves a lot of compatibility
issues.
Unfortunately, this has caused grief for the Windows build system we
use (MXE), and I'm still strying to work through the issues for that
before I can make a release. The MacOSX build is working well, and
python is working again on Mac, which is good news. Even the Linux
builds are working again, after a period where I could not build
SciDaVis 2.x on my OpenSUSE Tumbleweed system.
--
Dr Russell Standish Phone 0425 253119 (mobile)
Principal, High Performance Coders hpcoder@hpcoders.com.au http://www.hpcoders.com.au
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Thank you for the update. I'm glad to see so much progress and wider compatibility.
I imagine that the Windows build issues are the main blocker for a new release, if nothing else comes up. Given the new build system, it should also be easier to build on more linux distros.
Regarding doing builds from source on linux, is the procedure still the same, as described in github INSTALL.md ?
Thank you
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
On Sun, Feb 07, 2021 at 10:51:16AM -0000, Miguel Agueda wrote:
Thank you for the update. I'm glad to see so much progress and wider
compatibility.
I imagine that the Windows build issues are the main blocker for a new release,
if nothing else comes up. Given the new build system, it should also be easier
to build on more linux distros.
Regarding doing builds from source on linux, is the procedure still the same,
as described in github INSTALL.md ?
That page will need to be updated.
The most important difference is the use of cmake. The old qmake
system will probably still work for a while, but expect it to break
eventually.
Dependencies remain the same, however it is better to use git
submodule versions of both qwt and qwtplot3d, as these contain
important bugfixes.
So ensure that you have run "git submodule update --init --recursive"
in SciDAVis's git checkout.
Then make sure BUILD_QWT and BUILD_QWTPLOT3D is set to ON by running
ccmake or similar.
My current blocker for Windows builds is that the MXE install for
qttools does not provide a cmake file for QtDesigner, which is needed
to build QwtPlot3D. I'm stumped as to why not - maybe its a version
thing, MXE can sometimes be stuck on old versions. I was thinking of
falling back to the older method of building QwtPlot3D using qmake,
but now we have some important fixes in qwtplot3d, so I need to
rethink.
--
Dr Russell Standish Phone 0425 253119 (mobile)
Principal, High Performance Coders hpcoder@hpcoders.com.au http://www.hpcoders.com.au
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Brief update on the release for MacOSX. I was exploring the use of
CPack to create the "bundle" (.app firectory), and it does produce a
nicely installable .dmg, however dylib references are not
correct. There is a fix_bundle script, but it seems to die in
SciDAVis's case. I eventually figured out how to call my existing
mkMacDist script from CMake, however when doing this, the crash is
back. Something is askew with rewriting the references in the dylibs -
at least I now have another clue, which have been in short supply.
There is also the whole code signing and notarization malarky to deal
with to, however, there is a known procedure for doing that. It would
be nice to integrate it into cmake, however.
Cheers
--
Dr Russell Standish Phone 0425 253119 (mobile)
Principal, High Performance Coders hpcoder@hpcoders.com.au http://www.hpcoders.com.au
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I've tried to build Scidavis from source using cmake, following your guidelines to use qwt and qwtplot3d git submodules. Aside from not being able to build python scripting ( because I haven't figured out how to get the PyQt dependencies), everything builds perfectly on Fedora 33, and all the unittest run succesfully.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi,
I know @hpcoder is busy, and there's not much development going on on Scidavis at the moment, so I don't expect any new release soon with new features.
I've seen there's been some commits related to bug fixes since the last release (2.3.0). I'm personally interested on bug 400, related to exporting tables as ASCII, which is currently not working, but I imagine other people might be interested on other bug fixes.
I wonder if it's posible to release a new binary point release (2.3.1), to include these. I'm not a very experience user in compiling my own sofware, and I use SciDAVIs on Linux and Windows, so it'll be much easier for me to get updated binaries.
Thank you
I'll have a go at doing a release shortly (it would be numbered 2.4.0
- 2.3.1 would entail no source code difference between it and 2.3.0 :)
Windows and Linux builds should be straight forward - MacOSX is still
a problem, however...
On Tue, Dec 29, 2020 at 12:15:17PM -0000, Miguel Agueda wrote:
--
Dr Russell Standish Phone 0425 253119 (mobile)
Principal, High Performance Coders hpcoder@hpcoders.com.au
http://www.hpcoders.com.au
Great! I'm happy to hear that.
I did manage to compile it myself yesterday. I couldn't do it on Fedora that I use, because I got an error about lupdate not being avaliable, but I manage to compile on Ubuntu 18.04 without python and then copy it over Fedora and link scidavis to the needed libraries on the system.
It's kind of janky, and I'm not confident it'll work very well, but I can use it to export the tables.
Okay, I see what you mean. I wasn't sure about how it worked, and I thought maybe the fixes weren't substantial enough to justify a new release.
Thanks for the help
Just a brief update on release plans. Chinoune Mehdi has been doing
some excellent work towards moving SciDAVis from the dreadful qmake
system to the much more capable cmake system, as well as building qwt5
and qwtplot3d as git submodules, which solves a lot of compatibility
issues.
Unfortunately, this has caused grief for the Windows build system we
use (MXE), and I'm still strying to work through the issues for that
before I can make a release. The MacOSX build is working well, and
python is working again on Mac, which is good news. Even the Linux
builds are working again, after a period where I could not build
SciDaVis 2.x on my OpenSUSE Tumbleweed system.
--
Dr Russell Standish Phone 0425 253119 (mobile)
Principal, High Performance Coders hpcoder@hpcoders.com.au
http://www.hpcoders.com.au
Thank you for the update. I'm glad to see so much progress and wider compatibility.
I imagine that the Windows build issues are the main blocker for a new release, if nothing else comes up. Given the new build system, it should also be easier to build on more linux distros.
Regarding doing builds from source on linux, is the procedure still the same, as described in github INSTALL.md ?
Thank you
On Sun, Feb 07, 2021 at 10:51:16AM -0000, Miguel Agueda wrote:
That page will need to be updated.
The most important difference is the use of cmake. The old qmake
system will probably still work for a while, but expect it to break
eventually.
Dependencies remain the same, however it is better to use git
submodule versions of both qwt and qwtplot3d, as these contain
important bugfixes.
So ensure that you have run "git submodule update --init --recursive"
in SciDAVis's git checkout.
Then make sure BUILD_QWT and BUILD_QWTPLOT3D is set to ON by running
ccmake or similar.
My current blocker for Windows builds is that the MXE install for
qttools does not provide a cmake file for QtDesigner, which is needed
to build QwtPlot3D. I'm stumped as to why not - maybe its a version
thing, MXE can sometimes be stuck on old versions. I was thinking of
falling back to the older method of building QwtPlot3D using qmake,
but now we have some important fixes in qwtplot3d, so I need to
rethink.
--
Dr Russell Standish Phone 0425 253119 (mobile)
Principal, High Performance Coders hpcoder@hpcoders.com.au
http://www.hpcoders.com.au
Brief update on the release for MacOSX. I was exploring the use of
CPack to create the "bundle" (.app firectory), and it does produce a
nicely installable .dmg, however dylib references are not
correct. There is a fix_bundle script, but it seems to die in
SciDAVis's case. I eventually figured out how to call my existing
mkMacDist script from CMake, however when doing this, the crash is
back. Something is askew with rewriting the references in the dylibs -
at least I now have another clue, which have been in short supply.
There is also the whole code signing and notarization malarky to deal
with to, however, there is a known procedure for doing that. It would
be nice to integrate it into cmake, however.
Cheers
--
Dr Russell Standish Phone 0425 253119 (mobile)
Principal, High Performance Coders hpcoder@hpcoders.com.au
http://www.hpcoders.com.au
I've tried to build Scidavis from source using cmake, following your guidelines to use qwt and qwtplot3d git submodules. Aside from not being able to build python scripting ( because I haven't figured out how to get the PyQt dependencies), everything builds perfectly on Fedora 33, and all the unittest run succesfully.