Replaced eval { die } construct with a bool return value indicating
success or failure of an automatic arrangement of parts on the print bed.

Don't know exactly what is happening here, but throwing a "die" inside
a XS function and then catching it inside an eval {} block is suspcious.
This commit is contained in:
bubnikv 2016-11-04 15:03:51 +01:00
parent 3fc57ba8d8
commit 6217622865
7 changed files with 35 additions and 20 deletions

View file

@ -424,9 +424,12 @@ class ArrangeItemIndex {
ArrangeItem item;
ArrangeItemIndex(coordf_t _index, ArrangeItem _item) : index(_index), item(_item) {};
};
Pointfs
arrange(size_t total_parts, const Pointf &part_size, coordf_t dist, const BoundingBoxf* bb)
bool
arrange(size_t total_parts, const Pointf &part_size, coordf_t dist, const BoundingBoxf* bb, Pointfs &positions)
{
positions.clear();
Pointf part = part_size;
// use actual part size (the largest) plus separation distance (half on each side) in spacing algorithm
@ -446,7 +449,7 @@ arrange(size_t total_parts, const Pointf &part_size, coordf_t dist, const Boundi
size_t cellw = floor((area.x + dist) / part.x);
size_t cellh = floor((area.y + dist) / part.y);
if (total_parts > (cellw * cellh))
CONFESS(PRINTF_ZU " parts won't fit in your print area!\n", total_parts);
return false;
// total space used by cells
Pointf cells(cellw * part.x, cellh * part.y);
@ -527,7 +530,6 @@ arrange(size_t total_parts, const Pointf &part_size, coordf_t dist, const Boundi
}
}
// now we actually place objects into cells, positioned such that the left and bottom borders are at 0
Pointfs positions;
for (size_t i = 1; i <= total_parts; ++i) {
ArrangeItemIndex c = cellsorder.front();
cellsorder.erase(cellsorder.begin());
@ -544,7 +546,7 @@ arrange(size_t total_parts, const Pointf &part_size, coordf_t dist, const Boundi
}
}
return positions;
return true;
}
#endif