User:Rajb245: Difference between revisions
Jump to navigation
Jump to search
| Line 52: | Line 52: | ||
# [http://www.nsnam.org/bugzilla/show_bug.cgi?id=250 250 Tcp breaks if you set the delackcount > 2] -- Assigned to Craig for application | # [http://www.nsnam.org/bugzilla/show_bug.cgi?id=250 250 Tcp breaks if you set the delackcount > 2] -- Assigned to Craig for application | ||
# [http://www.nsnam.org/bugzilla/show_bug.cgi?id=365 365 TcpSocketImpl::Recv returns 0 without setting m_errno] -- Untested patch proposed | # [http://www.nsnam.org/bugzilla/show_bug.cgi?id=365 365 TcpSocketImpl::Recv returns 0 without setting m_errno] -- Untested patch proposed | ||
chat about it went like this:<br> | ##chat about it went like this:<br> | ||
[15:22] <rajb> so to be clear, if there is nothing to read (but we haven't reached EOF), there is a null pointer returned<br> | ##[15:22] <rajb> so to be clear, if there is nothing to read (but we haven't reached EOF), there is a null pointer returned<br> | ||
[15:22] <tomh_> ok, then raj, do you want to make a patch?<br> | ##[15:22] <tomh_> ok, then raj, do you want to make a patch?<br> | ||
[15:22] <rajb> but if it is EOF, we get a non-null pointer to an empty packet?<br> | ##[15:22] <rajb> but if it is EOF, we get a non-null pointer to an empty packet?<br> | ||
[15:22] <rajb> is this correct?<br> | ##[15:22] <rajb> is this correct?<br> | ||
[15:22] <rajb> if so, i can prepare the patch<br> | ##[15:22] <rajb> if so, i can prepare the patch<br> | ||
# [http://www.nsnam.org/bugzilla/show_bug.cgi?id=358 <s>358 calling listen on non-closed TCP sockets sends resets, should fail and do nothing</s>] -- Fixed | # [http://www.nsnam.org/bugzilla/show_bug.cgi?id=358 <s>358 calling listen on non-closed TCP sockets sends resets, should fail and do nothing</s>] -- Fixed | ||
# [http://www.nsnam.org/bugzilla/show_bug.cgi?id=411 411 OnOffApplication::OnTime does not work] -- Working on patch | # [http://www.nsnam.org/bugzilla/show_bug.cgi?id=411 411 OnOffApplication::OnTime does not work] -- Working on patch | ||
Revision as of 16:23, 3 December 2008
Main Page - Roadmap - Summer Projects - Project Ideas - Developer FAQ - Tools - Related Projects
HOWTOs - Installation - Troubleshooting - User FAQ - Samples - Models - Education - Contributed Code - Papers
About Raj
Raj Bhattacharjea is a full time developer of ns-3 in the employ of the Georgia Institute of Technology.
Completed Tasks
July 2008
- tcp-star-server example added
- tcp-errors example checked in. Subsequently removed.
- Cleaned up some TCP doxygen.
- Evensky test case added as a regression test and announced to list. Subsequently removed.
- Make PacketSink multitasking for TCP (closed bug 244)
- Fix TcpSocketImpl rx buffer logic (closed bug 248)
August 2008
- Add two unit TCP unit tests, one for a transfer, one for finite buffers.
- Make TypeId(const char * name) explicit (closed bug 281)
- Use scheduler to call up the stack in SimpleNetDevice (closed bug 287)
- Refactor OnOff so Start doesn't immediate close the socket (closed bug 259)
- Attended SIGCOMM, participated in giving the ns3 demo
September 2008
- Fix TCP closedown FINs (closed bug 242)
- Actually implemented TCP finite rx buffer (closed 239)
- Extensively tested ns-3 in prep for the 3.2 release
- filed bug 323 waf --valgrind doesn't check for valgrind first
- filed bug 337 opt build fails in address.cc on Ubuntu 8.10
- filed bug 343 utils/bench-packets is broken and asserts
- filed bug 349 packet unit test fails on optimized build
- filed bug 358 calling listen on non-closed TCP sockets sends resets, sh...
- filed bug 359 *SocketImpl::Bind returns something other than -1
- Helped Mathieu debug GTNetS TCP port for ns-3-simu use for his demo. End result was deciding that nsc is more mature at this point, but working through the issues identified some bugs, and got me a working simu setup to use for testing purposes.
October 2008
- Debugged the new server code commit mailhook, Tom fixed the exim conf and ns-commits list issues, so now it works.
- Resolved the hg-ssh script issue for code server main repository access control; problem was one of Mercurial python API changes ("import commands" to "import dispatch")
- Scheduled day long downtime on October 15th to reset mysql password, backup and move code, bugzilla, and wiki onto new machine. Code move was successful, but ECE didn't do DNS change, so migration will have to be repeated. Bugzilla move was successful, new wiki bombed.
- Uncoordinated DNS change from IT later that week forced a sloppy migration of code.nsnam.org, but the new server is up and operational.
- ns3 tutorial for WNS2 was completed; copy of the document is here. Revision control version here (slightly stale)
- Moved discussion of RandomVariable changes to the developers list
- Bugs 311 and 340 closed
- Migration of all services completed
- Made RC1 and RC2 of ns-3.2.1
November 2008
- Made RC3 of ns-3.2.1
- craigdo applied my patch for bug 359
Current Tasks, in rough order of priority
Address My Open Bugs for ns3.3
- 250 Tcp breaks if you set the delackcount > 2 -- Assigned to Craig for application
- 365 TcpSocketImpl::Recv returns 0 without setting m_errno -- Untested patch proposed
- chat about it went like this:
- [15:22] <rajb> so to be clear, if there is nothing to read (but we haven't reached EOF), there is a null pointer returned
- [15:22] <tomh_> ok, then raj, do you want to make a patch?
- [15:22] <rajb> but if it is EOF, we get a non-null pointer to an empty packet?
- [15:22] <rajb> is this correct?
- [15:22] <rajb> if so, i can prepare the patch
- chat about it went like this:
358 calling listen on non-closed TCP sockets sends resets, should fail and do nothing-- Fixed- 411 OnOffApplication::OnTime does not work -- Working on patch
- 392 Example star.cc confuses TCP -- Testing patch
410 Doxygen is not picking up class OnOffHelper-- Fixed- 341 Get unexpected dropped packets when using SetSendCallback with heavy traffic
- 423 TCP copy constructor wrong -- P5 bug, wontfix for 3.3
Random Tasks, lower priority
- Cleanup TcpSocketImpl::Recv code (its ugly and inefficient)
- Address bug 250
- Use ns-3-simu to debug TCP
- process microhttpd<->simulator "native" TCP<->tap net device<->real linux<->real firefox
- replacing simulator "native" TCP from above with NSC works, but above setup doesn't, indicating bugs
- Help Michelle with random variable API changes
- have Bugzilla auto update when a hg changeset references a tracked bug [1]