labath added a comment.
Ok, I've put it into AdbClient for now. We can revisit it if we see a second
user somewhere...
Repository:
rL LLVM
http://reviews.llvm.org/D19533
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
http://lists.llvm
clayborg added a comment.
I would prefer a fix in the ADB client only if I had to pick. I don't think
anyone else will use this functionality and prefer to the keep the API as
simple as possible for Connection subclasses.
Repository:
rL LLVM
http://reviews.llvm.org/D19533
___
labath added a comment.
This is going to be messier than I expected. I would need to introduce the same
repeat-until-the-buffer-is-full loop into ConnectionGenericFileWindows, which I
don't think is good for code reuse. I see two possibilities:
- go back to the version with a separate function
This revision was automatically updated to reflect the committed changes.
Closed by commit rL268380: Add a read_full_buffer argument to
ConnectionFileDescriptor::Read (authored by labath).
Changed prior to commit:
http://reviews.llvm.org/D19533?vs=55397&id=55986#toc
Repository:
rL LLVM
http
labath added inline comments.
Comment at: source/Plugins/Platform/Android/AdbClient.cpp:37
@@ -36,3 +36,3 @@
-const uint32_t kReadTimeout = 100; // 1 second
+const uint32_t kReadTimeout = 400; // 4 seconds
const char * kOKAY = "OKAY";
ovyalov wrote:
> I
ovyalov added a comment.
LGTM
Comment at: source/Plugins/Platform/Android/AdbClient.cpp:37
@@ -36,3 +36,3 @@
-const uint32_t kReadTimeout = 100; // 1 second
+const uint32_t kReadTimeout = 400; // 4 seconds
const char * kOKAY = "OKAY";
It might be usef
labath updated this revision to Diff 55397.
labath added a comment.
Added proper handling of the timeout after retries, and moved the functionality
into
Connection::Read. I have called the parameter read_full_buffer, as I think that
better reflects
its intended use. I did not want to add a defau
clayborg requested changes to this revision.
clayborg added a comment.
This revision now requires changes to proceed.
The "timeout_usec" parameter to Connection::ReadAll() doesn't seem like it is
doing what is intended. It uses this timeout for every read call which means
the Connection::ReadAll
labath created this revision.
labath added reviewers: clayborg, zturner.
labath added a subscriber: lldb-commits.
Herald added subscribers: danalbert, tberghammer.
AdbClient was attempting to handle the case where the socket input arrived in
pieces, but it was
failing to handle the case where the
