Skip to content
This repository has been archived by the owner on Mar 29, 2024. It is now read-only.

TULIP_INT_VERSION uses less digits in tulip5 than tulip4 #8

Closed
rgiot opened this issue Mar 31, 2017 · 1 comment
Closed

TULIP_INT_VERSION uses less digits in tulip5 than tulip4 #8

rgiot opened this issue Mar 31, 2017 · 1 comment

Comments

@rgiot
Copy link
Contributor

rgiot commented Mar 31, 2017

we cannot make conditional code for tulip4 or tulip5 because TULIP_INT_VERSION of tulip 5 is lower than TULIP_INT_VERSION of tulip 4

@rgiot
Copy link
Contributor Author

rgiot commented Mar 31, 2017

For example we have to do

#if TULIP_INT_VERSION == 501
    tlp::node n;
    forEach(n, g->getNodes()){
#else
    for(tlp::node n:g->nodes()) {
#endif

instead of something like

#if TULIP_INT_VERSION >= 5000
    tlp::node n;
    forEach(n, g->getNodes()){
#else
    for(tlp::node n:g->nodes()) {
#endif

bpinaud referenced this issue Jan 29, 2019
TravisCI: Migrate Linux builds from trusty to xenial
p-mary referenced this issue in anlambert/talipot Jan 3, 2020
QOpenGL module is marked as deprecated since a while now so it is time
to remove its use from the Talipot codebase and promote the use of
QOpenGL* classes directly integrated in the QtGui module.

The big difference between QOpenGL and QtOpenGL from Qt5 is that all
rendering is performed in framebuffer objects, there is no more direct
rendering in the underlying os windows with its own OpenGL context.

Talipot OpenGL rendering also follows that idiom, all renderings are performed
offscreen using a shared OpenGL context. This also means that there is no
more QGLWidget as viewport for QGraphicsView. Talipot OpenGL scene are
now converted to QImage in order to display them using the default Qt raster
rendering engine. This should fixes the numerous rendering glitches observed
on MacOS.

First thing observed after the migration is a consequent performance boost
in OpenGL rendering when using an Intel GPU on a Linux host machine (especially
when selecting elements, it is now 10 times faster on debian stable).
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant