 Chromium Code Reviews
 Chromium Code Reviews Issue 412006:
  posix: upgrade libevent from 1.4.7 to 1.4.13  (Closed)
    
  
    Issue 412006:
  posix: upgrade libevent from 1.4.7 to 1.4.13  (Closed) 
  | OLD | NEW | 
|---|---|
| 1 Changes in 1.4.13-stable: | |
| 2 o If the kernel tells us that there are a negative number of bytes to read from a socket, do not believe it. Fixes bug 2841177; found by Alexander Pronchenkov . | |
| 3 o Do not allocate the maximum event queue and fd array for the epoll backend at startup. Instead, start out accepting 32 events at a time, and double the queu e's size when it seems that the OS is generating events faster than we're reques ting them. Saves up to 512K per epoll-based event_base. Resolves bug 2839240. | |
| 4 o Fix compilation on Android, which forgot to define fd_mask in its sys/select. h | |
| 5 o Do not drop data from evbuffer when out of memory; reported by Jacek Masiulan iec | |
| 6 o Rename our replacement compat/sys/_time.h header to avoid build a conflict on HPUX; reported by Kathryn Hogg. | |
| 7 o Build kqueue.c correctly on GNU/kFreeBSD platforms. Patch pulled upstream fro m Debian. | |
| 8 o Fix a problem with excessive memory allocation when using multiple event prio rities. | |
| 9 o When running set[ug]id, don't check the environment. Based on a patch from Op enBSD. | |
| 10 | |
| 11 | |
| 12 Changes in 1.4.12-stable: | |
| 13 o Try to contain degree of failure when running on a win32 version so heavily f irewalled that we can't fake a socketpair. | |
| 14 o Fix an obscure timing-dependent, allocator-dependent crash in the evdns code. | |
| 15 o Use __VA_ARGS__ syntax for varargs macros in event_rpcgen when compiler is no t GCC. | |
| 16 o Activate fd events in a pseudorandom order with O(N) backends, so that we don 't systematically favor low fds (select) or earlier-added fds (poll, win32). | |
| 17 o Fix another pair of fencepost bugs in epoll.c. [Patch from Adam Langley.] | |
| 
Mark Mentovai
2009/11/20 16:31:33
nice
 | |
| 18 o Do not break evdns connections to nameservers when our IP changes. | |
| 19 o Set truncated flag correctly in evdns server replies. | |
| 20 o Disable strict aliasing with GCC: our code is not compliant with it. | |
| 
Mark Mentovai
2009/11/20 16:31:33
Does this need any attention on our side?
 | |
| 21 | |
| 22 Changes in 1.4.11-stable: | |
| 23 o Fix a bug when removing a timeout from the heap. [Patch from Marko Kreen] | |
| 24 o Remove the limit on size of HTTP headers by removing static buffers. | |
| 25 o Fix a nasty dangling pointer bug in epoll.c that could occur after epoll_reca lc(). [Patch from Kevin Springborn] | |
| 26 o Distribute Win32-Code/event-config.h, not ./event-config.h | |
| 27 | |
| 28 Changes in 1.4.10-stable: | |
| 29 o clean up buffered http connection data on reset; reported by Brian O'Kelley | |
| 30 o bug fix and potential race condition in signal handling; from Alexander Drozd ov | |
| 31 o rename the Solaris event ports backend to evport | |
| 32 o support compilation on Haiku | |
| 33 o fix signal processing when a signal callback delivers a signal; from Alexande r Drozdov | |
| 34 o const-ify some arguments to evdns functions. | |
| 35 o off-by-one error in epoll_recalc; reported by Victor Goya | |
| 36 o include Doxyfile in tar ball; from Jeff Garzik | |
| 37 o correctly parse queries with encoded \r, \n or + characters | |
| 38 | |
| 39 Changes in 1.4.9-stable: | |
| 40 o event_add would not return error for some backends; from Dean McNamee | |
| 
Mark Mentovai
2009/11/20 16:31:33
nice
 | |
| 41 o Clear the timer cache on entering the event loop; reported by Victor Chang | |
| 42 o Only bind the socket on connect when a local address has been provided; repor ted by Alejo Sanchez | |
| 43 o Allow setting of local port for evhttp connections to support millions of con nections from a single system; from Richard Jones. | |
| 44 o Clear the timer cache when leaving the event loop; reported by Robin Haberkor n | |
| 45 o Fix a typo in setting the global event base; reported by lance. | |
| 46 o Fix a memory leak when reading multi-line headers | |
| 47 o Fix a memory leak by not running explicit close detection for server connecti ons | |
| 48 | |
| 49 Changes in 1.4.8-stable: | |
| 50 o Match the query in DNS replies to the query in the request; from Vsevolod Sta khov. | |
| 51 o Fix a merge problem in which name_from_addr returned pointers to the stack; f ound by Jiang Hong. | |
| 52 o Do not remove Accept-Encoding header | |
| 53 | |
| 1 Changes in 1.4.7-stable: | 54 Changes in 1.4.7-stable: | 
| 2 o Fix a bug where headers arriving in multiple packets were not parsed; fix fro m Jiang Hong; test by me. | 55 o Fix a bug where headers arriving in multiple packets were not parsed; fix fro m Jiang Hong; test by me. | 
| 3 | 56 | 
| 4 Changes in 1.4.6-stable: | 57 Changes in 1.4.6-stable: | 
| 5 o evutil.h now includes <stdarg.h> directly | 58 o evutil.h now includes <stdarg.h> directly | 
| 6 o switch all uses of [v]snprintf over to evutil | 59 o switch all uses of [v]snprintf over to evutil | 
| 7 o Correct handling of trailing headers in chunked replies; from Scott Lamb. | 60 o Correct handling of trailing headers in chunked replies; from Scott Lamb. | 
| 8 o Support multi-line HTTP headers; based on a patch from Moshe Litvin | 61 o Support multi-line HTTP headers; based on a patch from Moshe Litvin | 
| 9 o Reject negative Content-Length headers; anonymous bug report | 62 o Reject negative Content-Length headers; anonymous bug report | 
| 10 o Detect CLOCK_MONOTONIC at runtime for evdns; anonymous bug report | 63 o Detect CLOCK_MONOTONIC at runtime for evdns; anonymous bug report | 
| (...skipping 30 matching lines...) Expand all Loading... | |
| 41 o use libevent's internal timercmp on all platforms, to avoid bugs on old platf orms where timercmp(a,b,<=) is buggy. | 94 o use libevent's internal timercmp on all platforms, to avoid bugs on old platf orms where timercmp(a,b,<=) is buggy. | 
| 42 o reduce system calls for getting current time by caching it. | 95 o reduce system calls for getting current time by caching it. | 
| 43 o fix evhttp_bind_socket() so that multiple sockets can be bound by the same ht tp server. | 96 o fix evhttp_bind_socket() so that multiple sockets can be bound by the same ht tp server. | 
| 44 o Build test directory correctly with CPPFLAGS set. | 97 o Build test directory correctly with CPPFLAGS set. | 
| 45 o Fix build under Visual C++ 2005. | 98 o Fix build under Visual C++ 2005. | 
| 46 o Expose evhttp_accept_socket() API. | 99 o Expose evhttp_accept_socket() API. | 
| 47 o Merge windows gettimeofday() replacement into a new evutil_gettimeofday() fun ction. | 100 o Merge windows gettimeofday() replacement into a new evutil_gettimeofday() fun ction. | 
| 48 o Fix autoconf script behavior on IRIX. | 101 o Fix autoconf script behavior on IRIX. | 
| 49 o Make sure winsock2.h include always comes before windows.h include. | 102 o Make sure winsock2.h include always comes before windows.h include. | 
| 50 | 103 | 
| 51 | |
| 52 Changes in 1.4.3-stable: | 104 Changes in 1.4.3-stable: | 
| 53 o include Content-Length in reply for HTTP/1.0 requests with keep-alive | 105 o include Content-Length in reply for HTTP/1.0 requests with keep-alive | 
| 54 o Patch from Tani Hosokawa: make some functions in http.c threadsafe. | 106 o Patch from Tani Hosokawa: make some functions in http.c threadsafe. | 
| 55 o Do not free the kqop file descriptor in other processes, also allow it to be 0; from Andrei Nigmatulin | 107 o Do not free the kqop file descriptor in other processes, also allow it to be 0; from Andrei Nigmatulin | 
| 56 o make event_rpcgen.py generate code include event-config.h; reported by Sam Ba nks. | 108 o make event_rpcgen.py generate code include event-config.h; reported by Sam Ba nks. | 
| 57 o make event methods static so that they are not exported; from Andrei Nigmatul in | 109 o make event methods static so that they are not exported; from Andrei Nigmatul in | 
| 58 o make RPC replies use application/octet-stream as mime type | 110 o make RPC replies use application/octet-stream as mime type | 
| 59 o do not delete uninitialized timeout event in evdns | 111 o do not delete uninitialized timeout event in evdns | 
| 60 | 112 | 
| 61 Changes in 1.4.2-rc: | 113 Changes in 1.4.2-rc: | 
| (...skipping 78 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... | |
| 140 o Make event_base_free() succeed even if there are pending non-internal events on a base. This may still leak memory and fds, but at least it no longer crashe s. | 192 o Make event_base_free() succeed even if there are pending non-internal events on a base. This may still leak memory and fds, but at least it no longer crashe s. | 
| 141 o Post-process the config.h file into a new, installed event-config.h file that we can install, and whose macros will be safe to include in header files. | 193 o Post-process the config.h file into a new, installed event-config.h file that we can install, and whose macros will be safe to include in header files. | 
| 142 o Remove the long-deprecated acconfig.h file. | 194 o Remove the long-deprecated acconfig.h file. | 
| 143 o Do not require #include <sys/types.h> before #include <event.h>. | 195 o Do not require #include <sys/types.h> before #include <event.h>. | 
| 144 o Add new evutil_timer* functions to wrap (or replace) the regular timeval mani pulation functions. | 196 o Add new evutil_timer* functions to wrap (or replace) the regular timeval mani pulation functions. | 
| 145 o Fix many build issues when using the Microsoft C compiler. | 197 o Fix many build issues when using the Microsoft C compiler. | 
| 146 o Remove a bash-ism in autogen.sh | 198 o Remove a bash-ism in autogen.sh | 
| 147 o When calling event_del on a signal, restore the signal handler's previous val ue rather than setting it to SIG_DFL. Patch from Christopher Layne. | 199 o When calling event_del on a signal, restore the signal handler's previous val ue rather than setting it to SIG_DFL. Patch from Christopher Layne. | 
| 148 o Make the logic for active events work better with internal events; patch from Christopher Layne. | 200 o Make the logic for active events work better with internal events; patch from Christopher Layne. | 
| 149 o We do not need to specially remove a timeout before calling event_del; patch from Christopher Layne. | 201 o We do not need to specially remove a timeout before calling event_del; patch from Christopher Layne. | 
| OLD | NEW |