| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
|
|
* maint-r13:
Remove copyright headers in vsn.mk files
Conflicts:
lib/appmon/vsn.mk
lib/erl_docgen/vsn.mk
lib/inets/vsn.mk
lib/kernel/vsn.mk
lib/reltool/vsn.mk
lib/ssl/vsn.mk
lib/stdlib/vsn.mk
lib/tools/vsn.mk
lib/tv/vsn.mk
lib/xmerl/vsn.mk
|
|
Copyright notices serve no useful purpose in vsn.mk files, and
only complicate scripts that automatically update version numbers.
|
|
* tm/fix-cover-html-escaping:
Fix cover's HTML-generation logic to escape ampersands in source code
OTP-8776
|
|
Large parts of the ethread library have been rewritten. The
ethread library is an Erlang runtime system internal, portable
thread library used by the runtime system itself.
Most notable improvement is a reader optimized rwlock
implementation which dramatically improve the performance of
read-lock/read-unlock operations on multi processor systems by
avoiding ping-ponging of the rwlock cache lines. The reader
optimized rwlock implementation is used by miscellaneous
rwlocks in the runtime system that are known to be read-locked
frequently, and can be enabled on ETS tables by passing the
`{read_concurrency, true}' option upon table creation. See the
documentation of `ets:new/2' for more information.
The ethread library can now also use the libatomic_ops library
for atomic memory accesses. This makes it possible for the
Erlang runtime system to utilize optimized atomic operations
on more platforms than before. Use the
`--with-libatomic_ops=PATH' configure command line argument
when specifying where the libatomic_ops installation is
located. The libatomic_ops library can be downloaded from:
http://www.hpl.hp.com/research/linux/atomic_ops/
The changed API of the ethread library has also caused
modifications in the Erlang runtime system. Preparations for
the to come "delayed deallocation" feature has also been done
since it depends on the ethread library.
Note: When building for x86, the ethread library will now use
instructions that first appeared on the pentium 4 processor. If
you want the runtime system to be compatible with older
processors (back to 486) you need to pass the
`--enable-ethread-pre-pentium4-compatibility' configure command
line argument when configuring the system.
|
|
When cover generates HTML files and embeds source code within them, it
does not escape ampersands in the source code. These bare ampersands
get misinterpreted by web browsers as HTML character-entity
references; as a result, source code containing ampersands renders
improperly. This small patch fixes the problem by causing bare
ampersands to be escaped into as "&" character-entity references.
|
|
|
|
* egil/eprof:
Fix eprof to handle error cases
Dialyzer cleanup for eprof
Fix fprof_SUITE to use new eprof API
Switch pattern API for eprof profile with mfa
Update eprof documentation
Update eprof_SUITE with more basic tests
Rename sorting choices for eprof
Teach eprof to align text output
Add dump functionality to eprof
Add log functionality to eprof
Update eprof tests to reflect new eprof
Add start timestamp to eprof profiling
Teach eprof to use the new breakpoints
OTP-8706 egil/eprof
eprof has been reimplemented with support in the Erlang virtual machine and
is now both faster (i.e. slows down the code being measured less) and
scales much better. In measurements we saw speed-ups compared to the old
eprof ranging from 6 times (for sequential code that only uses one
scheduler/core) up to 84 times (for parallel code that uses 8 cores).
|
|
|
|
|
|
|
|
|
|
|
|
The eprof documentation is updated to reflect the new eprof application
and some API changes and convenience functions.
|
|
|
|
- Added {sort, mfa} option
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Follow the GNU recommendation of not introducing runtime
dependencies on the Common Lisp package in ELisp packages
that will be distributed for widespread use.
In practice this means it's ok to use cl macros but not
functions.
|
|
This determines the most recent invocation of EUnit and runs it
again -- whether one test or all tests in a module were run (and
if all a module's tests were run, whether cover compilation was
involved). This can be especially handy when one test is being
run repeatedly, eliminating the need to position the cursor in
the desired test just to run it.
Default keybinding: C-c C-e l
|
|
If the variable 'erlang-eunit-autosave' is non-nil, buffers will be
automatically saved just before running tests -- the "Do you want
to save?" prompt will be inhibited. This can be useful, reducing the
save-compile-load-test cycle to just one keychord.
|
|
Add a number of features to the emacs mode which make it easier to
work with eunit test cases while keeping track of code coverage.
* C-cC-ec:
Compile the module under test for code coverage analysis, run tests
and show the results of the coverage analysis in a new buffer.
* C-c-C-ev:
Compile the current module for code coverage analysis.
* C-cC-ea:
Show the results of the coverage analysis in a new buffer.
|
|
Previously there were only two options: either the EUnit test file was
placed directly within the test directory or within the same directory
as the source file. Now a list of candidate locations are supported.
This way more than two alternative locations are supported as well as
arbitrarily deep directory structures (i.e. test/eunit/x_tests.erl).
The default behaviour is still the same, but it's possible to specify
locations by setting the following two variables:
erlang-eunit-src-candidate-dirs
erlang-eunit-test-candidate-dirs
|
|
* dgud/emacs-catch-improvements:
Improved indentation of old catch.
Added more type highlighting and fixed record indentation with types.
|
|
An example that didn't work previously.
case Foo of
{ok, X} ->
ok;
_ ->
catch file:close(FD)
end.
|
|
Type highlighting reported by Jay Nelson
non_neg_integer() will highlight purple but pos_integer() does not.
Closing record indentation problem reported by Maxim Treskin:
-record(state, {
sequence_number = 1 :: integer()
}).
|
|
|
|
|
|
Hopefully this covers at least some of the common cases and makes the
flymake support more usable as is. The purpose of including the ebin
directory is to support things like behaviours and parse transforms.
|
|
|
|
|
|
When given the option {builtins,true} Xref now adds calls to operators.
|
|
* mh/escript-erlang-mode:
Associate files using interpreter "escript" with Erlang mode
|
|
* cb/emacs-eunit-run-current-test:
Add Emacs EUnit utility feature: run current test function.
|
|
* mh/fix-record-indentation:
Fix indentation of records with line breaks inside lists
|
|
* mh/inferior-erlang-cmd-uniq:
inferior-erlang: specify command to run, uniquify buffer names
|
|
* mh/erlang-mode-imenu-arity:
Let imenu distinguish functions by arity
|
|
Create and assign the erlang-mode keymap to erlang-mode-map once
erlang.el is loaded, not the first time it is used. This way of doing
it follows the Emacs Lisp convention described at
http://www.gnu.org/software/emacs/manual/html_node/elisp/Tips-for-Defining.html
and also makes it easier for users to customize the keymap.
Remove the now unneeded functions erlang-keymap-init and
erlang-mode-commands.
Also move the definition of inferior-erlang-use-cmm, so it is
available when defining erlang-mode-map.
|
|
Make the function names used by M-x imenu contain the function arity
as well as the name. This makes it possible to jump to all functions
of the same name, as opposed to only the one that is defined first.
For example, in the following module there is now a way to jump
directly to foo/1:
-module(foo).
foo() ->
ok.
bar() ->
error.
foo(A) ->
{ok, A}.
|
|
Currently make:files will fail with the atom 'error' and
no message when the input (.erl) file is unreadable or the
output (.beam) file is unwritable. This differs from erlc
which will print a useful error message, or when possible,
remove the unwritable output file and continue successfully.
This change removes the unnecessary checks on the files when
make:files is called and allows the error checking to be
done in compile:file, where the error messages are produced.
It does not affect the return value.
In particular this resolves the mysterious problem of make:files
failing but erlc succeeding, caused by an unwritable (usually
root owned) beam file in an otherwise writable build directory.
|
|
Trying to indent this piece of code by hitting TAB on each line:
foo() ->
[#foo{
foo = foo}].
used to cause an error. Fix by ignoring errors when trying to skip
backwards from a record.
Also add this test case to test.erl.orig and test.erl.indented.
|
|
Add a new feature to the command inferior-erlang: when called with a
prefix argument, prompt for the command to execute to start the Erlang
shell. When possible (i.e., in Emacs 23 and later), allow tab
completion as in a normal shell.
Also create the shell buffer in such a way that, if
uniquify-buffer-name-style is not nil and several Erlang shell buffers
are started in different directories, the names of those buffers
reflect their working directories instead of being *erlang*,
*erlang*<2>, etc.
|
|
A file whose first line contains "#!/usr/bin/env escript" should be
opened in erlang-mode, even if its name doesn't end in ".erl". Add an
entry to interpreter-mode-alist to that effect.
|
|
This determines the current test function under point (cursor), whether the active buffer
contains a test module (foo_tests) or the module itself (foo), then compiles the source
module (and test module if present) and runs that test function.
Since this added a third testing feature, there was some natural duplication that needed
to be factored out. The notion of 'compile and run tests' was decoupled from exactly
which tests should be run. The definitions of default keybindings were converted to a list
for clarity and conciseness.
|
