otp.git/lib/compiler, branch OTP_R14B03

Prepare release

2011-05-24T11:16:24+00:00

Update copyright years

2011-05-20T14:11:43+00:00

OTP-9094: [httpc] Add support for upload body streaming (PUT and POST).

2011-05-16T07:46:10+00:00

	  Filipe David Manana

OTP-9114: [ftp] Added (type) spec for all exported functions.

OTP-9123: mod_esi:deliver/2 made to accept binary data.
	  Bernard Duggan

OTP-9124: [httpd] Prevent XSS in error pages.
	  Michael Santos

OTP-9131: [httpd] Wrong security property names used in documentation.
	  Garrett Smith

OTP-9157: [httpd] Improved error messages.
	  Ricardo Catalinas Jim�nez

OTP-9158: [httpd] Fix timeout message generated by mod_esi.
	  Bernard Duggan

OTP-9202: [httpd] Extended support for file descriptors.
	  Attila Rajmund Nohl

OTP-9230: The default ssl kind has now been changed to essl.

OTP-9246: [httpc] httpc manager crash because of a handler retry
	  race condition.

Merge branch 'bmk/inets/inet56_integration' into dev

r

2011-05-12T06:03:42+00:00

Use Erlang specs and types for documentation

bs_match_SUITE: Improve coverage of beam_bsm

2011-04-12T03:54:47+00:00

beam_bsm: Eliminate uncovered line in warning generation

2011-04-12T03:54:47+00:00

In warning_translate_label/2, gb_trees:lookup/2 is called
to translate from the entry label for a function to its name.
Since the gb_tree has an entry for all functions in the module,
there is no way that the lookup can fail unless there is a
serious bug.

Therefore, use gb_trees:get/2 so that an exception and an
internal compiler error will be generated if the lookup would
ever fail.

match_SUITE: Cover a clause in beam_dead:forward/4

2011-04-12T03:54:47+00:00

beam_dead: Remove uncovered special case handling of empty blocks

2011-04-12T03:54:46+00:00

There is never any empty blocks when beam_dead is invoked.
Even if there were, they will be removed a little bit later in
forward/4.

beam_dead: Remove uncovered clauses in binary matching optimization

2011-04-12T03:52:58+00:00

In the optimization of binary matching, it seems that two clauses
cannot never be reached. Removing the clauses is safe, since that
would only mean that an opportunity for an optimization is lost

beam_dead: Remove uncoverable case clause in update_value_dict/3

2011-04-12T03:52:57+00:00

Because the code generator (v3_codegen) would not include the
same value more than once in a select_val/3 instruction and
because a label can only be referenced by one select_val/3
instruction, there is no way that the correct value could already
be in the gb_tree. (Even if it could happen, this change is
safe because only opportunity for an optimization would be missed;
incorrect code would not be generated.)