Fix vertex istream operator

[sloriot]

Reported on cgal-discuss.

[sloriot]

@cjamin @mglisse I updated the testsuite to test the IO, but I have a runtime error. Did I miss something?

[mglisse]

operator>> for points seems broken. It is in a wrong file, and expects to read something different from what operator<< prints.

[lrineau]

@mglisse Where is the implementation of operator>>. Could you please fix it?

[mglisse]

Where is the implementation of operator>>. Could you please fix it?

It is in a weird place (Triangulation/IO but it should be in NewKernel_d if it exists at all). No time right now, I am traveling.

[lrineau]

This PR is the only thing that "blocks" CGAL-4.11.2. I will postpone it to CGAL-4.12.1, or later, so that I can release CGAL-4.11.2 and close that release branch.

[lrineau]

@mglisse commented on Jun 21, 2018, 3:00 PM GMT+2:

Where is the implementation of operator>>. Could you please fix it?

It is in a weird place (Triangulation/IO but it should be in NewKernel_d if it exists at all). No time right now, I am traveling.

Marc, could you comment, now?

[mglisse]

Branch mglisse/NewKernel_d-istream-glisse ?
For a 3D point with coordinates 7, 8, 9, we output 3 7 8 9. The code in Triangulation was expecting to read just 7 8 9, I changed it to expect 3 7 8 9. For fixed dimension, we could use 7 8 9, but for dynamic dimension it was forcing Clement's code to read a whole line and expect that it contains exactly the whole point and nothing else.

[lrineau]

@sloriot @mglisse I let you discuss about the best way to fix the issue...

[sloriot]

@mglisse I cherry-picked your commits in this branch.
I have to do 2 more modifs so that it works. @cjamin could you check 4ed546e and 3ed89cb?

[lrineau]

There is an issue with the base branch of this PR:

sloriot wants to merge 1,920 commits into CGAL:releases/CGAL-4.12-branch

[sloriot]

@maxGimeno merged master some some reason. I removed that commit and we're back to 11 new commits.

[maxGimeno]

I merged master because I couldn't use the cgal root as a cgal dir. I guess the branch is older than the working header only mechanism, but I didn't think of it at the time.

[lrineau]

Let me know once the PR is ready to be integrated.

[cjamin]

Hi. Sorry I missed this discussion, but to be honest, I hardly remember anything about this, except the code was a bit messy and that I didn't spend much time on trying to clean it up. Thanks for the fixes.

[sloriot]

Successfully tested in 4.14-Ic-40.