GitLab

extrude.

any scale factor < 0.0001 is now set to 0.0001 (0 had trouble with
interacting with other shapes).

values for x and y (default 1).  It makes the block really long.  It does
seem to work the way openscad's does.

updated logic blocks to do internal type checking
made text() block openscad compatible
fixed capitalization of "minus" in difference block
fixed a couple of field bugs in IE11
fixed mutator drag bug (again?)
added redraw blocks button (for fixing graphical glitches)
made undo/redo use icons instead of words
increased project name limit to 28 chars from 24

should return the convex hull of that shape.  Fixed the parser so that now
it does.

string, boolean, and number input, and use the toString() method on the
input so that you can actually print out numbers, and "true" and "false"
from the result of an equation or logical expression.  Fun?  I think so.
Had to modify the parser, the block definitions, the logic block
definitions (which used 1 and 0 instead of true and false) to do it.

Now, all loop variables as typed as numbers and all procedure arguments
are typed as "null", meaning they'll fit anywhere.  It isn't perfect, but
it is a start.

block associated with them).  Fixed.

needs more testing, but for now this seems done.

parser) - if a string is surrounded by double quotes, the length is two
less than it should be.

I then decided that I needed to type the variables (now that we have
number, boolean, and string variables).  It is mostly done and working
now.  I now need to make changing a variable name a real change to pick up
those too - right now the variable can change names without triggering a
type change (which it needs if it used to be an instance of a variable
with a different type, and now is either a new variable or linked to a
variable with a different type).  But I want to check this version in
first.

sides block "NAN" from variable fix
variables are blocked from attaching to loops and if statements because
that is what our parser can handle
code examples under the Help->Examples menu
version changed to 1.1.2
Procedure definitions are not counted as "shapes" for type checking
Procedure definitions are not disabled when using "disable all blocks"
3D text block
2D text block renamed, tooltip added
chain hull loop block restored (new shinier hull!)
loops missing fields now trigger the "missing fields" error and
highlighting
Openscad code easily exportable via the Project menu
Fixed error messages with cylinder block zero and negative fields
Updated logo with our shiny new BlocksCAD logo

restored original CSG setters (by setting _x,_y,_z values directly in the
hull where needed).
re-added chain hull back into the toolbox.

reindexFacesandVertices) to make hulling extra speedy.  It is working
great!

than the previous implementation (probably mostly because I cull the
repeated points out before performing the hull).
There are a couple things I should clean up in the code, but this is
already usable.

extrude, rotate_extrude, and fishyBlocks.

Revert "Merge branch 'development' of https://github.com/EinsteinsWorkshop/BlocksCAD into development"

This reverts commit 6b8acac9, reversing
changes made to 081327ec.

This reverts commit a542c8c5.

This reverts commit ee6d9919.

Added Examples to Help Menu.  Currently populated with one example
(torus).

Works offline.  Not yet tested with backend, and it neeeeeeds testing.
Thus, do not use yet.

en.js.

Conflicts:
	blockscad/blockscad.js

fixing blockscad version after merging field_lock

unequal values the next change re-sets the values equal to each other
again.  Putting in new blocks is used as the new value to lock to.
It doesn't work perfectly with undo, but this is the best I can do with
the way the Blockly event system is architected.  The worst behavior is
when you change radius 1 and 2 in an alternating pattern - you lose half
the states to the locking system overwriting what was supposed to be the
new value.  Otherwise, changes to 1 or 2 will work fine except you get an
extra undo state in there - you have to click twice to go back.

for your cone (is that useful?) seems mostly functional.  Needs testing.
I deliberately left the code long (there are several edge cases that can
be combined) in case testing reveals that the edge cases actually do need
different code.

block ratios.  I think I can ditch it but want a saved copy in case I'm
wrong about not needing it.