oscarsaleta/p4

P4 (Polynomial Planar Phase Portraits) software for phase portrait computation and representation in the plane or other projections such as the Poincaré Sphere.

p4's Issues

Possible separatrice stability missclassification

When pressing Plot All Separatrices several times (to get longer integrations) separatrices sometimes are changed from stable to unstable in the plot (from blue to red or viceversa).

x' and y' fields are not completed when loading file even in 1 piece cases

When there are more than 1 vector fields, they only complete if there's a component that's equal in every piece, but in the case of 1 piece they should always complete so there's something wrong with the detection of common elements.

Plot All Separatrices not working properly on Linux

It's confirmed that "Plot All Separatrices" works differently in Linux and Windows

if output window is closed, it won't open again

When another evaluation is performed, output window doesn't open back.

Segfault if trying to compute a constant as a separating curve

Title. The original P5 can handle this case (plots a point, Maple detects it's a constant, and it works fine)

Piecewise config window sphere aspect ratio is not 1

Set the aspect ratio so the sphere is a sphere by default.

Crash on isoclines plot

Cannot read computation results and the program crashes with a segmentation fault

Segmentation fault when evaluating arbitrary curve twice

If Evaluate button is pressed a second time, P4 crashes with a segfault

Load VF not working

The x' and y' input fields don't fill when a VF is loaded. Need to check if this means that the VF is not properly read, but in the case of a single VF it should always be displayed.

EDIT: the VFs are not properly loaded. They are treated as (0,0), so the load function is bugged

If there are more than 1 VF, just display one of them (properly combined with the VF selection on the right side).

Segmentation fault when trying to delete last limit cycle

Example: Van der Pol with mu=1: (y,x+(1+x^2)*y. Evaluate, Plot and search the limit cycle (unique). Then press Delete Last LC, program crashes.

Shift+left mouse button to select separatrice causes program to crash

Segmentation fault after selecting an already plotted separatrice.

Info:

• Separating curve: x^2+2*y-1
• Vector field 1: (-y+x*y+y^2, x+y*x^2) outside of origin region
• Vector field 2: (x+y, x^2) inside of origin region

Plot all separatrices and then shift + left click to select one = SIGABRT.

P4 infinite loop in plot all separatrices

Evaluate any vector field, plot it and press Plot All Separatrices, the program will become unresponsive until force-killed.

Segmentation fault when trying to compute isoclines

Issue seems related to static function insert_isoclines_points from math_isoclines.cpp

Error: `errorOccurred` is not a member of `QProcess`

This is a compilation error raised in master and beyond.

The cause of the error is that in Qt versions > 5.5, the signal for an error in QProcess changed, from QProcess::error() to QProcess::errorOccurred(), and this change is not backwards compatible with versions <= 5.5.

Segmentation fault when adding a second separating curve

Added one curve, evaluated it, and then pressed Add button, causing P4 to crash.

Improve integration near separating curves

If an integration step crosses a separating curve, it should backtrace until the intersection point and recompute that part of the step using the new vector field equations. This will solve poorly computed separatrices near the separating curves, especially when a singularity is found near such curves.

If Qt is installed in a custom directory, compile_install_p4 doesn't find it

The script fails because it cannot detect Qt or qmake, even if the QMAKE variable is passed, because it then looks for Qt modules with pkg_config and doesn't find them.

Delete VF doesn't delete up to 1 VF

Maybe it does, but the dropdown menu shows 1 and 2, which would mean that even when deleting VF 2 it still remains there.

Problems when reading results for some VFs

Seems that the singularity/separatrice types are missread or misswritten and this causes a failure in readTables. Will have to test more with different VFs and different P4 versions to see where the problem lies

Segmentation fault when computing separating curves

Looks like the segfault comes from reading the results after computation

Selecting singular point not working

Shift+click on plot makes the animation of selecting the correct singular point, but the Separatrices window for plotting a separatrice from that singularity doesn't detect the selection

Segmentation fault when plotting arbitrary curve

Looks like the curve window pointer gets messed up after pressing the plot button, which causes P4 to crash when the status of the delete buttons is changed after computation of the plots.

Uncovered after solving #20

Select next separatrice can cause a crash

Steps: Shift + click to select a separatrice, then click the "select next" button until program crashes

Separating curves (for VF regions) are not plotted in Piecewise configuration window

Could be related to #18 and #20 , since it looks like the computation results are not read properly for the separating curves.

Piecewise separatrices/orbits aren't computed correctly depending on the region

Separatrices and orbits are computed for one of the regions and plotted in other regions instead of each region having its own dynamics.

iOS Monterey y P4

Instale iOS Monterey y P4 no funciona, necesito ayuda.

Simplify P4InputVF evaluate code

There are many methods that do very similar jobs in evaluating Maple scripts. They could be probably all merged into one or 2 methods that do slightly different jobs depending on parameters.

In Piecewise Configuration window, cannot re-evaluate curves

If Eval has been pressed before, the button won't reactivate even if the user changes input curves or adds new ones.

Crash on continue integration from separatrice

Conditions: degenerate point from (x+y,x^2-1) at infinity with separating curve x that intersects with that singularity at infinity. Try to integrate from it and program crashes.

Related to start_plot_de_sep() from math_desep.cpp, but will likely affect also saddle and semi-elementary singularities.

Error when computing Lyapunov constants

Maple output: Error (readlyapunovresults): invalid subscript selector. The lyapunov executable could not be computing properly or the results are not written/read correctly.

Plot doesn't work in a region near the center

Looks like P4 detects that region as virtual and doesn't plot separatrices inside. The orbits also get cut and don't continue outside of the region, so there is something that stops computation and plotting when the border is crossed.

Individual separatrice integration doesn't work

In the Plot Separatrice dialog, selecting a singularity and then pressing Start integrating sep does nothing.

Plot not working: cannot read computation results

Doesn't happen every time, but often. Example VF: (x+y,x^2).

Cannot compute arbitrary curves

The Maple error is:
! Error (optimizepolynomial2) "cannot determine if this expression is true or false: not rounded"

Plot separating curve doesn't plot the full curve

The curve x is only plotted in half plane (for y<0). No more info available for now

Keep output window scrolling down after first evaluation

Right now, after the first evaluation, the output window won't keep scrolling down with new output text, which makes the user have to scroll down manually and not be able to follow the output as it's written

Help window is embedded in the main window

Cannot close it since it would close the main window

Old Maple versions DAG error

Some Maple versions older than Maple 2015 print out an error like "unexpected DAG type", and the computation ends without an error print from P4 itself.

Tests will be made with previous Maple versions to see if there is something that makes the scripts incompatible with older Maple versions.

P4 hangs when looking for limit cycles

This issue affects Linux, Windows and Mac. The search for limit cycles keeps stuck at 0% and P4 becomes unresponsive

P4 crash on specific VF

The VF is simple: (x*y,x)

[Mac] Finite results window does not reopen

After being closed it does not reopen when clicking the button

Plot arbitrary curve not working

P4 evaluates the curve but results cannot be read for plotting after computation. Could be related to #18

P4 in Windows, error computing Lyapunov constants

! Error (readline) "file or directory does not exist"

It could have something to do with Maple not being able to read the results file because notation needs double backslashes \

Adapt plot window size to toolbar witdh so no buttons are hidden

Keeping aspect ratio of plot sphere should be automatic since it forces a 1:1 aspect ratio on creation.

If QDIR is set but QMAKE is not, compile_install_p4 will use default qmake instead of the one in QDIR

The checking should be done first for a custom QDIR that sets a QMAKE, then look for the default one if not set.

[Mac] No P4 icon in macOS dock

Would be nice to find a way to display the icon, but it is unlikely since installation is unix-like and not .app/.dmg

Division by zero error in Maple

Vector field:

[ (2*(1+x))*(36*x^2*y^2+685*x*y^3+1120*y^4+57462*x^2*y+5007*x*y^2+14125*y^3+172109*x^2+43763*x*y+62091*y^2+105013*x+89293*y),
-274*x*y^4-1120*y^5-2878*x*y^3-17074*y^4+13106*x*y^2-98078*y^3+125454*x*y-244374*y^2-302400*x-210026*y ]

Unexpected behavior of compile_install_p4 when P4 is installed in both root and home and not uninstalled during execution

The script will ask if you want to install P4 in home anyway, leading to an error if the user says yes.

Add VF selector to isoclines dialog

This could be good for computing isoclines only in one vector field