A even more sharper knife!

They did it! Slic3r as a new version in the wild with the coolest features you can find, take look:

This is a very important update, grab it now and enjoy! I want to say thanks to the wonderful community that has been helping with feedback, early testing and nice ideas. A lot of people is involved in our open "quality assurance" process! 

New features:

• graphical interface for composing multiple objects in one single print (with tools for rotation, scaling etc.)
• new --merge switch to compose multiple object with auto-positioning from command line
• new option to randomize starting points across layers
• automatic detection of additional required perimeters for avoiding haps in domed/sloping objects (the perimeters settings means nowminimum perimeters as Slic3r could add more when needed)
• sequential printing: print a complete object, then move onto next one, with automatic collision detection
• ability to export STL files of composed plates, as well as a Splitcommand to turn multi-object files into individual objects
• we can read OBJ files now

Improvements:

• very large memory savings and speed boosts, allowing to process high-resolutions files without getting out of memory; also, the number of threads is customizable
• very improved surface quality due to new smoothing algorithms
• the GUI doesn't block while slicing and a Cancel button was added to stop the process
• hole perimeters are extruded in reverse order (from the outermost to the innermost) to get better overhangs
• a slight compensation is applied to avoid small hole shrinkage
• support material now uses honeycomb pattern (much work is still needed on support material)
• retract before changing tool for dual extrusion

Changes:

• SVG colors were inverted to better support DLP printing (next versions will carry more settings for SVG export)

Bugfixes:

• fixed some GUI memory leaks
• fixed some fatal errors
• fixed a bug causing high memory consumption when infilling every 2 or more layers
• some holes where filled
• some nearly-thin walls were discarded
• removed tiny dots/blobs that were generated sometimes
• bottom layer speed ratio wasn't taken into account when estimating layer time
• omit any G92 E0 when in relative mode