Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

merge master #2

Merged
merged 248 commits into from
Nov 11, 2014
Merged

merge master #2

merged 248 commits into from
Nov 11, 2014

Conversation

dhoegh
Copy link
Owner

@dhoegh dhoegh commented Nov 11, 2014

No description provided.

wavexx and others added 30 commits July 15, 2014 14:49
- Support the syntax `[user@]host[:port][*count] [bind_addr]` in the
  machinefile.
- Initial code just re-writes the list fed to addprocs() by repeating the host
  as specified, so there is no user-visible change.
Char now supports a limited set of "integer like" behavior.

* comparisons with integers
* Char - Char = Int
* Char + Int = Char

update docs and add a NEWS.md entry
Make Char no longer a subtype of Integer
I added a note that the updating-operators can change the type.  This is implicit in the current manual, but it nonetheless threw me for a loop the first time it happened to me (this is different, e.g., from C++), so I thought it should be made explicit.
floor(2^31-1*runif(1)) was always returning the value 2147483647. It is now floor(runif(1, min=0, max=2^31-1))
clarify docs for the is_valid_* functions
generalised and improved cat implementation
Both APIs are not very compatible, so let's use fill_array_* functions, which
are faster but require MersenneTwister to have an additional array field.
The state of the global RNG was previously handled by libdSFMT.
This commit allows the global RNG to benefit speed improvements resulting
from the use of fill_array_* functions (cf. previous commit).

However, scalar calls to rand() (i.e. producing one value instead of
filling an array) using the global RNG are now slower, as well as those
filling non-Float64 arrays (less than twice as slow though).
Wilfred and others added 28 commits November 11, 2014 00:27
This gives us more accurate highlighting, and makes julia-in-char more
accurate. Previously, it could be confused by:

                edit_insert(s, ';|')
            end
        end,
        '?' => function (s,o...)

because ; is legal before a character, and we weren't enforcing a
maximum length of character literals.
We can't go back before (point-min), and Emacs errors if we try.
These are about to become part of fairly low-level infrastructure,
so it's best to run them early.
Consistent varargs specialization in stagedfunctions (fix #8944)
julia-mode.el: Smarter indentation
fix `@doc`  for abstract types
export transpose! and ctranspose!
Improve docs for tempname() and tempdir()
dhoegh added a commit that referenced this pull request Nov 11, 2014
@dhoegh dhoegh merged commit f33da1f into dhoegh:master Nov 11, 2014
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.