The directory object now has error handling like the rest of the
dart:io. The added tests to some extend overlaps with some of the
tests in DirectpryTest.dart, but I found it better to have error tests
in a single place.
I removed a few type checks sitting in the C++ code for the native
functions as for file we currently don't have them. I have opened
issue 2305 regarding where to do type checks for dart:io native
functions.
R=ager@google.com
BUG=
TEST=
Review URL: https://chromiumcodereview.appspot.com//9773018
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@5866 260f80e4-7a28-3924-810f-c04153c831b5
When compiling all code, stop compiling after the first error.
Previously, we ignored compilation errors.
This was a feature wish so we can detect errors in libraries.
In fact, there are errors in the library, e.g. references
to non-existing top-level functions in utf16 code. And possibly
others. If this code gets checked in, any tests that depends on
--compile_all will break.
First problem:
'dart:utf': Error: line 607 pos 7: identifier 'is16BitCodeUnit' is not declared in this scope
if (is16BitCodeUnit()) {
^
If I hack that, the next one is:
'dart:utf': Error: line 613 pos 9: identifier 'utf16CodeUnitsToCodepoints' is not declared in this scope
utf16CodeUnitsToCodepoints(codeunits, 0, null, replacementCodepoint));
^
Didn't investigate any more than that...
Review URL: https://chromiumcodereview.appspot.com//9836064
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@5835 260f80e4-7a28-3924-810f-c04153c831b5
Now all both sync and async methods on File and RandomAccessFile
should have correct error handling.
Changed readByte and readByteSync to return -1 instead of
reporting an error when end of file is reached. It seems to make
the most sense.
The amount of boiler-plate code in both the native functions and
native port functions in file.cc is growing. Maybe we should
consider some refactoring, e.g. with a table of the expected
argument types for each native port functions together with known
position of the FILE* argument.
R=ager@google.com
BUG=
TEST=
Review URL: https://chromiumcodereview.appspot.com//9839053
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@5812 260f80e4-7a28-3924-810f-c04153c831b5
- Fixed the bigint instance allocation path when full snapshots are read.
- JSRegExp objects are only serialized/deserialized in a message snapshot and
hence do not need that special path. There is an assertion in JSRegExp to
ensure that these objects are not part of a full snapshot.
Review URL: https://chromiumcodereview.appspot.com//9835074
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@5802 260f80e4-7a28-3924-810f-c04153c831b5
The socket errors and HTTP parser errors are now handled by the
HTTP library and propagated to the related onError callback.
For HttpClient there is no onError on the HttpClient object
itself. Errors on active connections (including connection
errors) are reported on the HttpClientConnection. Errors on
connections sitting in the keep alive pool are handled by just
closing the connection and removing it from the pool. Errors on
the HttpClientConnection are also reported through the onError
callbacks on any active streams.
I am not particulair happy of the way errors are propagated to
the active streams but that can be refactored later.
Added more checks and tests to the HTTP parser.
R=ager@google.com, ajohnsen@google.com
BUG=
TEST=
Review URL: https://chromiumcodereview.appspot.com//9834008
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@5768 260f80e4-7a28-3924-810f-c04153c831b5
Previously, the byte array get and set routines for integers would not
correctly narrow an integer when the width of a value being stored was
greater than the element width. This change implements the behavior in
the Web IDL specification.
In addition, an implementation for uint64 get and set is now provided.
BUG=2042
Review URL: https://chromiumcodereview.appspot.com//9769013
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@5738 260f80e4-7a28-3924-810f-c04153c831b5
The message content detection and handling should now be more
correct and according to the specification (RFC 2616 section 4.3).
* Requests with no Content-Length and no Transfer-Encoding have no message body
* Responses with status code 1xx, 204 and 304 have no message body
* Responses which are responses to HEAD requests have no message body
Also added support for detecting chunk-extensions and ignoring these.
R=ajohnsen@google.com
BUG=
TEST=
Review URL: https://chromiumcodereview.appspot.com//9812005
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@5722 260f80e4-7a28-3924-810f-c04153c831b5
Change the streams onError callbacks to get an Exception argument on
all streams.
Added actual error handling to the socket streams.
As it is difficult to determine where an error should be reported I
decided to call all the registered onError callbacks, that is on the
socket itself and on both input and output stream.
Did some additional cleanup of the naming of OnError argument in some tests.
R=ager@google.com
BUG=
TEST=
Review URL: https://chromiumcodereview.appspot.com//9812007
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@5717 260f80e4-7a28-3924-810f-c04153c831b5
There are currently no tests of the error handling, as it is hard to
create consistent tests. I have done some manual tests locally:
* Lowering the number of file descriptors available (ulimit -n) and hitting that limit
* Running two dart programs communicating and terminating one (using Ctrl-C)
* Running two dart programs communicating on two machines and pulling out the network cable.
If there is an error we always call the onError callback and never the onClosed callback.
On Windows there is the issue that both closing the connection correctly and terminating one end gives the same error (ERROR_NETNAME_DELETED) so both are reported as connection close. Pulling out the network cable gives a different (real) error on Windows though.
On Mac OS it turned out the for kqueue EV_EOF is also used to indicate errors with the error code set in the fflags field. EV_ERROR is only reported if there is an internal error in kevent processing (see http://developer.apple.com/library/mac/#documentation/Darwin/Reference/ManPages/man2/kqueue.2.html).
R=ager@google.com
BUG=
TEST=
Review URL: https://chromiumcodereview.appspot.com//9720045
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@5709 260f80e4-7a28-3924-810f-c04153c831b5
Turn the depth-first traversal into a pass to discover the graph's basic
block structure. Record basic-block predecessors in the block entry
instructions. Also record the last instruction in a block in the block
entry instruction (giving constant-time access to the block's successors).
Also, record the depth-first spanning tree of the traversal to be used for
dominator computation.
R=srdjan@google.com
BUG=
TEST=
Review URL: https://chromiumcodereview.appspot.com//9730003
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@5655 260f80e4-7a28-3924-810f-c04153c831b5
An empty regular expression had caused split to loop infinitely. Now,
the split method behaves like its namesake in ECMAScript and splits a
string along character boundaries.
The split, replaceFirst, and replaceAll methods only supported RegExp
and String patterns. Now, these methods are implemented on top of the
allMatches method and work for all subinterfaces of Pattern.
BUG=2171
Review URL: https://chromiumcodereview.appspot.com//9689089
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@5653 260f80e4-7a28-3924-810f-c04153c831b5
Changed the onError callback for sockets to take an exception argument.
First area to report errors for is name resolution.
Changed the IP address resolution from being synchronous with the
creation of the socket to being asynchronous thsough a lookup
service running on a native port. Currently this lookup service
is static on the socket class as we need to call it from a
factory method. We should probably add a separate name lookup
service at some point.
R=ager@google.com
BUG=
TEST=
Review URL: https://chromiumcodereview.appspot.com//9699017
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@5640 260f80e4-7a28-3924-810f-c04153c831b5
Local functions and function literals are unknown to the VM
until the enclosing function is compiled. This change introduces
a list of known closures per class, and a function to look up
the innermost closure at a given token index.
The debugger now checks whether there are newly discovered
closures when a function containing a pending breakpoint is
compiled. If necessary, the breakpoint is set to the inner
function.
Review URL: https://chromiumcodereview.appspot.com//9716004
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@5599 260f80e4-7a28-3924-810f-c04153c831b5
The Lengauer-Tarjan dominator/dominance frontier algorithm needs to
visit the blocks in reverse preorder (in the first pass) and preorder
(in the second). Compute that at the same time as postorder (e.g.,
using the same spanning tree).
Take this chance to make the flow graph visitors always operate on a
forward block ordering.
R=srdjan@google.com
BUG=
TEST=
Review URL: https://chromiumcodereview.appspot.com//9719003
git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@5598 260f80e4-7a28-3924-810f-c04153c831b5
srdjan@google.com