1 Changelog for the c-ares project
3 * December 9 2008 (Gisle Vanem)
5 Fixes for Win32 targets using the Watt-32 tcp/ip stack.
7 * Dec 4 2008 (Daniel Stenberg)
9 Gregor Jasny provided the patch that introduces ares_set_socket_callback(),
10 and I edited it to also get duped by ares_dup().
12 * Dec 3 2008 (Daniel Stenberg)
16 I made sure the public ares_config struct looks like before and yet it
17 supports the ROTATE option thanks to c-ares now storing the "optmask"
18 internally. Thus we should be ABI compatible with the past release(s)
19 now. My efforts mentioned below should not break backwards ABI compliance.
21 Here's how I suggest we proceed with the API:
23 ares_init() will be primary "channel creator" function.
25 ares_init_options() will continue to work exactly like now and before. For
26 starters, it will be the (only) way to set the existing options.
28 ares_save_options() will continue to work like today, but will ONLY save
29 options that you can set today (including ARES_OPT_ROTATE actually) but new
30 options that we add may not be saved with this.
34 ares_dup() that instead can make a new channel and clone the config used
35 from an existing channel. It will then clone all config options, including
36 future new things we add.
38 ares_set_*() style functions that set (new) config options. As a start we
39 simply add these for new functionality, but over time we can also introduce
40 them for existing "struct ares_options" so that we can eventually deprecate
41 the two ares_*_options() functions.
43 ares_get_*() style functions for extracting info from a channel handle that
44 should be used instead of ares_save_options().
46 * Nov 26 2008 (Yang Tse)
47 - Brad Spencer provided changes to allow buildconf to work on OS X.
49 - Gerald Combs fixed a bug in ares_parse_ptr_reply() which would cause a
50 buffer to shrink instead of expand if a reply contained 8 or more records.
52 * Nov 25 2008 (Yang Tse)
53 - In preparation for the upcomming IPv6 nameservers patch, the internal
54 ares_addr union is now changed into an internal struct which also holds
57 * Nov 19 2008 (Daniel Stenberg)
58 - Brad Spencer brought the new function ares_gethostbyname_file() which simply
59 resolves a host name from the given file, using the regular hosts syntax.
61 * Nov 1 2008 (Daniel Stenberg)
62 - Carlo Contavalli added support for the glibc "rotate" option, as documented
65 causes round robin selection of nameservers from among those listed. This
66 has the effect of spreading the query load among all listed servers, rather
67 than having all clients try the first listed server first every time.
69 You can enable it with ARES_OPT_ROTATE
71 * Oct 21 2008 (Yang Tse)
72 Charles Hardin added handling of EINPROGRESS for UDP connects.
74 * Oct 18 2008 (Daniel Stenberg)
75 Charles Hardin made adig support a regular numerical dotted IP address for the
78 * Oct 7 2008 (Yang Tse)
79 - Added --enable-optimize configure option to enable and disable compiler
80 optimizations to allow decoupled setting from --enable-debug.
82 * Oct 2 2008 (Yang Tse)
83 - Added --enable-warnings configure option to enable and disable strict
84 compiler warnings to allow decoupled setting from --enable-debug.
86 * Sep 17 2008 (Yang Tse)
87 - Code reorganization to allow internal/private use of "nameser.h" to any
88 system that lacks arpa/nameser.h or arpa/nameser_compat.h header files.
90 * Sep 16 2008 (Yang Tse)
91 - Code reorganization to allow internal/private use of ares_writev to any
92 system that lacks the writev function.
94 * Sep 15 2008 (Yang Tse)
95 - Code reorganization to allow internal/private use of ares_strcasecmp to any
96 system that lacks the strcasecmp function.
98 - Improve configure detection of some string functions.
100 * Sep 11 2008 (Yang Tse)
101 - Code reorganization to allow internal/private use of ares_strdup to any
102 system that lacks the strdup function.
104 Version 1.5.3 (Aug 29, 2008)
106 * Aug 25 2008 (Yang Tse)
107 - Improvement by Brad House:
109 This patch addresses an issue in which a response could be sent back to the
110 source port of a client from a different address than the request was made to.
111 This is one form of a DNS cache poisoning attack.
113 The patch simply uses recvfrom() rather than recv() and validates that the
114 address returned from recvfrom() matches the address of the server we have
115 connected to. Only necessary on UDP sockets as they are connection-less, TCP
118 - Fix by George Neill:
119 Fixed compilation of acountry sample application failure on some systems.
121 * Aug 4 2008 (Daniel Stenberg)
122 - Fix by Tofu Linden:
125 * Users (usually, but not always) on 2-Wire routers and the Comcast service
126 and a wired connection to their router would find that the second and
127 subsequent DNS lookups from fresh processes using c-ares to resolve the same
128 address would cause the process to never see a reply (it keeps polling for
129 around 1m15s before giving up).
132 * On such a machine (and yeah, it took us a lot of QA to find the systems
133 that reproduce such a specific problem!), do 'ahost www.secondlife.com',
134 then do it again. The first process's lookup will work, subsequent lookups
135 will time-out and fail.
138 * init_id_key() was calling randomize_key() *before* it initialized
139 key->state, meaning that the randomness generated by randomize_key() is
140 immediately overwritten with deterministic values. (/dev/urandom was also
141 being read incorrectly in the c-ares version we were using, but this was
142 fixed in a later version.)
143 * This makes the stream of generated query-IDs from any new c-ares process
144 be an identical and predictable sequence of IDs.
145 * This makes the 2-Wire's default built-in DNS server detect these queries
146 as probable-duplicates and (erroneously) not respond at all.
149 * Aug 4 2008 (Yang Tse)
150 - Autoconf 2.62 has changed the behaviour of the AC_AIX macro which we use.
151 Prior versions of autoconf defined _ALL_SOURCE if _AIX was defined. 2.62
152 version of AC_AIX defines _ALL_SOURCE and other four preprocessor symbols
153 no matter if the system is AIX or not. To keep the traditional behaviour,
154 and an uniform one across autoconf versions AC_AIX is replaced with our
155 own internal macro CARES_CHECK_AIX_ALL_SOURCE.
157 * Aug 1 2008 (Yang Tse)
158 - Configure process now checks if the preprocessor _REENTRANT symbol is already
159 defined. If it isn't currently defined a set of checks are performed to test
160 if its definition is required to make visible to the compiler a set of *_r
161 functions. Finally, if _REENTRANT is already defined or needed it takes care
162 of making adjustments necessary to ensure that it is defined equally for the
163 configure process tests and generated config file.
165 * Jul 20 2008 (Yang Tse)
166 - When recvfrom prototype uses a void pointer for arguments 2, 5 or 6 this will
167 now cause the definition, as appropriate, of RECVFROM_TYPE_ARG2_IS_VOID,
168 RECVFROM_TYPE_ARG5_IS_VOID or RECVFROM_TYPE_ARG6_IS_VOID.
170 * Jul 17 2008 (Yang Tse)
171 - RECVFROM_TYPE_ARG2, RECVFROM_TYPE_ARG5 and RECVFROM_TYPE_ARG6 are now defined
172 to the data type pointed by its respective argument and not the pointer type.
174 * Jul 16 2008 (Yang Tse)
175 - Improved configure detection of number of arguments for getservbyport_r.
176 Detection is now based on compilation checks instead of linker ones.
178 - Configure process now checks availability of recvfrom() socket function and
179 finds out its return type and the types of its arguments. Added definitions
180 for non-configure systems config files, and introduced macro sreadfrom which
181 will be used on udp sockets as a recvfrom() wrapper in the future.
183 * Jul 15 2008 (Yang Tse)
184 - Introduce definition of _REENTRANT symbol in setup.h to improve library
185 usability. Previously the configure process only used the AC_SYS_LARGEFILE
186 macro for debug builds, now it is also used for non-debug ones enabling the
187 use of configure options --enable-largefile and --disable-largefile which
188 might be needed for library compatibility. Remove checking the size of
189 curl_off_t, it is no longer needed.
191 * Jul 3 2008 (Daniel Stenberg)
192 - Phil Blundell: If you ask ares_gethostbyname() to do an AF_INET6 lookup and
193 the target host has only A records, it automatically falls back to an
194 AF_INET lookup and gives you the A results. However, if the target host has
195 a CNAME record, this behaviour is defeated since the original query does
196 return some data even though ares_parse_aaa_reply() doesn't consider it
197 relevant. Here's a small patch to make it behave the same with and without
200 * Jul 2 2008 (Yang Tse)
201 - Fallback to gettimeofday when monotonic clock is unavailable at run-time.
203 * Jun 30 2008 (Daniel Stenberg)
205 - As was pointed out to me by Andreas Schuldei, the MAXHOSTNAMELEN define is
206 not posix or anything and thus c-ares failed to build on hurd (and possibly
207 elsewhere). The define was also somewhat artificially used in the windows
208 port. Now, I instead rewrote the use of gethostbyname to enlarge the host
209 name buffer in case of need and totally avoid the use of the MAXHOSTNAMELEN
210 define. I thus also removed the defien from the namser.h file where it was
211 once added for the windows build.
213 I also fixed init_by_defaults() function to not leak memory in case if
216 * Jun 9 2008 (Yang Tse)
218 - Make libcares.pc generated file for pkg-config include information relative
219 to the libraries needed for the static linking of c-ares.
221 * May 30 2008 (Yang Tse)
223 - Brad House fixed a missing header file inclusion in adig sample program.
225 Version 1.5.2 (May 29, 2008)
227 * May 13 2008 (Daniel Stenberg)
229 - Introducing millisecond resolution support for the timeout option. See
230 ares_init_options()'s ARES_OPT_TIMEOUTMS.
232 * May 9 2008 (Yang Tse)
234 - Use monotonic time source if available, for private function ares__tvnow()
236 * May 7 2008 (Daniel Stenberg)
238 - Sebastian made c-ares able to return all PTR-records when doing reverse
239 lookups. It is not common practice to have multiple PTR-Records for a single
240 IP, but its perfectly legal and some sites have those.
242 - Doug Goldstein provided a configure patch: updates autoconf 2.13 usage to
243 autoconf 2.57 usage (which is the version you have specified as the minimum
244 version). It's a minor change but it does clean up some warnings with newer
245 autoconf (specifically 2.62).
247 * May 5 2008 (Yang Tse)
249 - Improved parsing of resolver configuration files.
251 * April 4 2008 (Daniel Stenberg)
253 - Eino Tuominen improved the code when a file is used to seed the randomizer.
255 - Alexey Simak made adig support NAPTR records
257 - Alexey Simak fixed the VC dsp file by adding the missing source file
260 * December 11 2007 (Gisle Vanem)
262 - Added another sample application; acountry.c which converts an
263 IPv4-address(es) and/or host-name(s) to country-name and country-code.
264 This uses the service of the DNSBL at countries.nerd.dk.
266 * December 3 2007 (Daniel Stenberg)
268 - Brad Spencer fixed the configure script to assume that there's no
269 /dev/urandom when built cross-compiled as then the script cannot check for
272 - Erik Kline cleaned up ares_gethostbyaddr.c:next_lookup() somewhat
274 Version 1.5.1 (Nov 21, 2007)
276 * November 21 2007 (Daniel Stenberg)
278 - Robin Cornelius pointed out that ares_llist.h was missing in the release
281 Version 1.5.0 (Nov 21, 2007)
283 * October 2 2007 (Daniel Stenberg)
285 - ares_strerror() segfaulted if the input error number was out of the currently
288 - Yang Tse: Avoid a segfault when generating a DNS "Transaction ID" in
289 internal function init_id_key() under low memory conditions.
291 * September 28 2007 (Daniel Stenberg)
293 - Bumped version to 1.5.0 for next release and soname bumped to 2 due to ABI
294 and API changes in the progress callback (and possibly more coming up from
297 * September 28 2007 (Steinar H. Gunderson)
299 - Don't skip a server if it's the only one. (Bugfix from the Google tree.)
301 - Made the query callbacks receive the number of timeouts that happened during
302 the execution of a query, and updated documentation accordingly. (Patch from
305 - Support a few more socket options: ARES_OPT_SOCK_SNDBUF and
308 - Always register for TCP events even if there are no outstanding queries, as
309 the other side could always close the connection, which is a valid event
310 which should be responded to.
312 * September 22 2007 (Daniel Stenberg)
314 - Steinar H. Gunderson fixed: Correctly clear sockets from the fd_set on in
315 several functions (write_tcp_data, read_tcp_data, read_udp_packets) so that
316 if it fails and the socket is closed the following code doesn't try to use
319 - Steinar H. Gunderson modified c-ares to now also do to DNS retries even when
320 TCP is used since there are several edge cases where it still makes sense.
322 - Brad House provided a fix for ares_save_options():
324 Apparently I overlooked something with the ares_save_options() where it
325 would try to do a malloc(0) when no options of that type needed to be saved.
326 On most platforms, this was fine because malloc(0) doesn't actually return
327 NULL, but on AIX it does, so ares_save_options would return ARES_ENOMEM.
329 * July 14 2007 (Daniel Stenberg)
331 - Vlad Dinulescu fixed two outstanding valgrind reports:
333 1. In ares_query.c , in find_query_by_id we compare q->qid (which is a short
334 int variable) with qid, which is declared as an int variable. Moreover,
335 DNS_HEADER_SET_QID is used to set the value of qid, but DNS_HEADER_SET_QID
336 sets only the first two bytes of qid. I think that qid should be declared as
337 "unsigned short" in this function.
339 2. The same problem occurs in ares_process.c, process_answer() . query->qid
340 (an unsigned short integer variable) is compared with id, which is an
341 integer variable. Moreover, id is initialized from DNS_HEADER_QID which sets
342 only the first two bytes of id. I think that the id variable should be
343 declared as "unsigned short" in this function.
345 Even after declaring these variables as "unsigned short", the valgrind
346 errors are still there. Which brings us to the third problem.
348 3. The third problem is that Valgrind assumes that query->qid is not
349 initialised correctly. And it does that because query->qid is set from
350 DNS_HEADER_QID(qbuf); Valgrind says that qbuf has unitialised bytes. And
351 qbuf has uninitialised bytes because of channel->next_id . And next_id is
352 set by ares_init.c:ares__generate_new_id() . I found that putting short r=0
353 in this function (instead of short r) makes all Valgrind warnings go away.
354 I have studied ares__rc4() too, and this is the offending line:
356 buffer_ptr[counter] ^= state[xorIndex]; (ares_query.c:62)
358 This is what triggers Valgrind.. buffer_ptr is unitialised in this function,
359 and by applying ^= on it, it remains unitialised.
361 Version 1.4.0 (June 8, 2007)
363 * June 4 2007 (Daniel Stenberg)
365 - James Bursa reported a major memory problem when resolving multi-IP names
366 and I found and fixed the problem. It was added by Ashish Sharma's patch
369 When I then tried to verify multiple entries in /etc/hosts after my fix, I
370 got another segfault and decided this code was not ripe for inclusion and I
375 - Brad Spencer found and fixed three flaws in the code, found with the new
376 gcc 4.2.0 warning: -Waddress
378 - Brad House fixed VS2005 compiler warnings due to time_t being 64bit.
379 He also made recent Microsoft compilers use _strdup() instead of strdup().
381 - Brad House's man pages for ares_save_options() and ares_destroy_options()
384 - Ashish Sharma provided a patch for supporting multiple entries in the
385 /etc/hosts file. Patch edited for coding style and functionality by me
390 - Shmulik Regev brought cryptographically secure transaction IDs:
392 The c-ares library implementation uses a DNS "Transaction ID" field that is
393 seeded with a pseudo random number (based on gettimeofday) which is
394 incremented (++) between consecutive calls and is therefore rather
395 predictable. In general, predictability of DNS Transaction ID is a well
396 known security problem (e.g.
397 http://bak.spc.org/dms/archive/dns_id_attack.txt) and makes a c-ares based
398 implementation vulnerable to DNS poisoning. Credit goes to Amit Klein
399 (Trusteer) for identifying this problem.
401 The patch I wrote changes the implementation to use a more secure way of
402 generating unique IDs. It starts by obtaining a key with reasonable entropy
403 which is used with an RC4 stream to generate the cryptographically secure
406 Note that the key generation code (in ares_init:randomize_key) has two
407 versions, the Windows specific one uses a cryptographically safe function
408 provided (but undocumented :) by the operating system (described at
409 http://blogs.msdn.com/michael_howard/archive/2005/01/14/353379.aspx). The
410 default implementation is a bit naive and uses the standard 'rand'
411 function. Surely a better way to generate random keys exists for other
414 The patch can be tested by using the adig utility and using the '-s' option.
416 - Brad House added ares_save_options() and ares_destroy_options() that can be
417 used to keep options for later re-usal when ares_init_options() is used.
419 Problem: Calling ares_init() for each lookup can be unnecessarily resource
420 intensive. On windows, it must LoadLibrary() or search the registry
421 on each call to ares_init(). On unix, it must read and parse
422 multiple files to obtain the necessary configuration information. In
423 a single-threaded environment, it would make sense to only
424 ares_init() once, but in a heavily multi-threaded environment, it is
425 undesirable to ares_init() and ares_destroy() for each thread created
428 Solution: Create ares_save_options() and ares_destroy_options() functions to
429 retrieve and free options obtained from an initialized channel. The
430 options populated can be used to pass back into ares_init_options(),
431 it should populate all needed fields and not retrieve any information
432 from the system. Probably wise to destroy the cache every minute or
433 so to prevent the data from becoming stale.
435 - Daniel S added ares_process_fd() to allow applications to ask for processing
436 on specific sockets and thus avoiding select() and associated
437 functions/macros. This function will be used by upcoming libcurl releases
438 for this very reason. It also made me export the ares_socket_t type in the
439 public ares.h header file, since ares_process_fd() uses that type for two of
444 - Ravi Pratap fixed a flaw in the init_by_resolv_conf() function for windows
445 that could cause it to return a bad return code.
449 - Yang Tse: Provide ares_getopt() command-line parser function as a source
450 code helper function, not belonging to the actual c-ares library.
454 - Vlad Dinulescu added ares_parse_ns_reply().
458 - Yang Tse: Fix failure to get the search sequence of /etc/hosts and
459 DNS from /etc/nsswitch.conf, /etc/host.conf or /etc/svc.conf when
460 /etc/resolv.conf did not exist or was unable to read it.
464 - Install ares_dns.h too
466 - Michael Wallner fixed this problem: When I set domains in the options
467 struct, and there are domain/search entries in /etc/resolv.conf, the domains
468 of the options struct will be overridden.
472 - Yang Tse removed a couple of potential zero size memory allocations.
474 - Andreas Rieke fixed the line endings in the areslib.dsp file that I (Daniel)
475 broke in the 1.3.2 release. We should switch to a system where that file is
476 auto-generated. We could rip some code for that from curl...
478 Version 1.3.2 (November 3, 2006)
482 - Prevent ares_getsock() to overflow if more than 16 sockets are used.
486 - Guilherme Balena Versiani: I noted a strange BUG in Win32 port
487 (ares_init.c/get_iphlpapi_dns_info() function): when I disable the network
488 by hand or disconnect the network cable in Windows 2000 or Windows XP, my
489 application gets 127.0.0.1 as the only name server. The problem comes from
490 'GetNetworkParams' function, that returns the empty string "" as the only
491 name server in that case. Moreover, the Windows implementation of
492 inet_addr() returns INADDR_LOOPBACK instead of INADDR_NONE.
498 o made ares_version.h use extern "C" for c++ compilers
499 o fixed compiler warnings in ares_getnameinfo.c
500 o fixed a buffer position init for TCP reads
504 - Ravi Pratap fixed ares_getsock() to actually return the proper bitmap and
507 Version 1.3.1 (June 24, 2006)
511 - Gisle Vanem added getopt() to the ahost program. Currently accepts
512 only [-t {a|aaaa}] to specify address family in ares_gethostbyname().
516 - (wahern) Removed "big endian" DNS section and RR data integer parser
517 macros from ares_dns.h, which break c-ares on my Sparc64. Bit-wise
518 operations in C operate on logical values. And in any event the octets are
519 already in big-endian (aka network) byte order so they're being reversed
520 (thus the source of the breakage).
524 - William Ahern handles EAGAIN/EWOULDBLOCK errors in most of the I/O calls
527 TODO: Handle one last EAGAIN for a UDP socket send(2) in
532 - Bram Matthys brought my attention to a libtool peculiarity where detecting
533 things such as C++ compiler actually is a bad thing and since we don't need
534 that detection I added a work-around, much inspired by a previous patch by
535 Paolo Bonzini. This also shortens the configure script quite a lot.
539 - Nick Mathewson added the ARES_OPT_SOCK_STATE_CB option that when set makes
540 c-ares call a callback on socket state changes. A better way than the
541 ares_getsock() to get full control over the socket state.
545 - Alexander Lazic improved the getservbyport_r() configure check.
549 - Alexander Lazic pointed out that the buildconf should use the ACLOCAL_FLAGS
550 variable for easier controlling what it does and how it runs.
554 - James Bursa fixed c-ares to find the hosts file on RISC OS, and made it
555 build with newer gcc versions that no longer defines "riscos".
559 - Daniel Stenberg added ares_getsock() that extracts the set of sockets to
560 wait for action on. Similar to ares_fds() but not restricted to using
561 select() for the waiting.
565 - Yang Tse fixed some send() / recv() compiler warnings
569 - Added constants that will be used by ares_getaddrinfo
571 - Made ares_getnameinfo use the reentrant getservbyport (getservbyport_r) if it
572 is available to ensure it works properly in a threaded environment.
576 - configure fix for detecting a member in the sockaddr_in6 struct which failed
577 on ipv6-enabled HP-UX 11.00
579 Version 1.3.0 (August 29, 2005)
583 - Alfredo Tupone provided a fix for the Windows code in get_iphlpapi_dns_info()
584 when getting the DNS server etc.
588 - Added some checks for the addrinfo structure.
594 Make UDP sockets non-blocking. I've confirmed that at least on Linux 2.4 a
595 read event can come back from poll() on a valid SOCK_DGRAM socket but
596 recv(2) will still block. This patch doesn't ignore EAGAIN in
597 read_udp_packets(), though maybe it should. (This patch was edited by Daniel
598 Stenberg and a new configure test was added (imported from curl's configure)
599 to properly detect what non-blocking socket approach to use.)
601 I'm not quite sure how this was happening, but I've been seeing PTR queries
602 which seem to return empty responses. At least, they were empty when calling
603 ares_expand_name() on the record. Here's a patch which guarantees to
604 NUL-terminate the expanded name. The old behavior failed to NUL-terminate if
605 len was 0, and this was causing strlen() to run past the end of the buffer
606 after calling ares_expand_name() and getting ARES_SUCCESS as the return
607 value. If q is not greater than *s then it's equal and *s is always
608 allocated with at least one byte.
612 - Added ares_getnameinfo which mimics the getnameinfo API (another feature
613 that could use testing).
617 - Added an inet_ntop function from BIND for systems that do not have it.
621 - Made sortlist support IPv6 (this can probably use some testing).
623 - Made sortlist support CIDR matching for IPv4.
627 - Added preliminary IPv6 support to ares_gethostbyname. Currently, sortlist
628 does not work with IPv6. Also provided an implementation of bitncmp from
629 BIND for systems that do not supply this function. This will be used to add
630 IPv6 support to sortlist.
632 - Made ares_gethostbyaddr support IPv6 by specifying AF_INET6 as the family.
633 The function can lookup IPv6 addresses both from files (/etc/hosts) and
638 - Tupone Alfredo fixed includes of arpa/nameser_compat.h to build fine on Mac
643 - Dominick Meglio: Provided implementations of inet_net_pton and inet_pton
644 from BIND for systems that do not include these functions.
648 - Dominick Meglio added ares_parse_aaaa_reply.c and did various
649 adjustments. The first little steps towards IPv6 support!
653 - Fixed the VC project and makefile to use ares_cancel and ares_version
657 - The released ares_version.h from 1.2.1 says 1.2.0 due to a maketgz flaw.
660 Version 1.2.1 (October 20, 2004)
664 - Henrik Stoerner fix: got a report that Tru64 Unix (the unix from Digital
665 when they made Alpha's) uses /etc/svc.conf for the purpose fixed below for
666 other OSes. He made c-ares check for and understand it if present.
668 - Now c-ares will use local host name lookup _before_ DNS resolving by default
669 if nothing else is told.
673 - Henrik Stoerner: found out that c-ares does not look at the /etc/host.conf
674 file to determine the sequence in which to search /etc/hosts and DNS. So on
675 systems where this order is defined by /etc/host.conf instead of a "lookup"
676 entry in /etc/resolv.conf, c-ares will always default to looking in DNS
677 first, and /etc/hosts second.
681 1) resolv.conf (for the "lookup" line);
682 2) nsswitch.fon (for the "hosts:" line);
683 3) host.conf (for the "order" line).
687 - Dominick Meglio patched: C-ares on Windows assumed that the HOSTS file is
688 located in a static location. It assumed
689 C:\Windows\System32\Drivers\Etc. This is a poor assumption to make. In fact,
690 the location of the HOSTS file can be changed via a registry setting.
692 There is a key called DatabasePath which specifies the path to the HOSTS
694 http://www.microsoft.com/technet/itsolutions/network/deploy/depovg/tcpip2k.mspx
696 The patch will make c-ares correctly consult the registry for the location
701 - Gisle Vanem fixed the MSVC build files.
705 - Gisle Vanem made c-ares build and work with his Watt-32 TCP/IP stack.
709 - Harshal Pradhan made a minor syntax change in ares_init.c to make it build
714 - Made the lib get built static only if --enable-debug is used.
718 Basically in loops like handle_errors(), 'query->next' was assigned a local
719 variable and then query was referenced after the memory was freed by
720 next_server(). I've changed that so next_server() and end_query() returns
721 the next query. So callers should use this ret-value.
723 The next problem was that 'server->tcp_buffer_pos' had a random value at
724 entry to 1st recv() (luckily causing Winsock to return ENOBUFS).
726 I've also added a ares_writev() for Windows to streamline the code a bit
730 - Fixed a few variable return types for some system calls. Made configure
731 check for ssize_t to make it possible to use that when receiving the send()
732 error code. This is necessary to prevent compiler warnings on some systems.
734 - Made configure create config.h, and all source files now include setup.h that
735 might include the proper config.h (or a handicrafted alternative).
737 - Switched to 'ares_socket_t' type for sockets in ares, since Windows don't
740 - automake-ified and libool-ified c-ares. Now it builds libcares as a shared
741 lib on most platforms if wanted. (This bloated the size of the release
742 archive with another 200K!)
744 - Makefile.am now uses Makefile.inc for the c sources, h headers and man
745 pages, to make it easier for other makefiles to use the exact same set of
748 - Adjusted 'maketgz' to use the new automake magic when building distribution
751 - Anyone desires HTML and/or PDF versions of the man pages in the release
755 - Günter Knauf made c-ares build and run on Novell Netware.
758 - Gisle Vanem provided Makefile.dj to build with djgpp, added a few more djgpp
759 fixes and made ares not use 'errno' to provide further info on Windows.
762 - Gisle Vanem made it build with djgpp and run fine with the Watt-32 stack.
765 - Gisle Vanem's init patch for Windows:
767 The init_by_resolv_conf() function fetches the DNS-server(s)
768 from a series of registry branches.
770 This can be wrong in the case where DHCP has assigned nameservers, but the
771 user has overridden these servers with other prefered settings. Then it's
772 wrong to use the DHCPNAMESERVER setting in registry.
774 In the case of no global DHCP-assigned or fixed servers, but DNS server(s)
775 per adapter, one has to query the adapter branches. But how can c-ares know
776 which adapter is valid for use? AFAICS it can't. There could be one adapter
777 that is down (e.g. a VPN adapter).
779 So it's better to leave this to the IP Helper API (iphlapi) available in
780 Win-98/2000 and later. My patch falls-back to the old way if not available.
783 - James Bursa fixed an init issue for RISC OS.
786 - Nico Stappenbelt reported that when processing domain and search lines in
787 the resolv.conf file, the first entry encountered is processed and used as
788 the search list. According to the manual pages for both Linux, Solaris and
789 Tru64, the last entry of either a domain or a search field is used.
791 This is now adjusted in the code
793 Version 1.2.0 (April 13, 2004)
796 - Updated various man pages to look nicer when converted to HTML on the web
800 - Dirk Manske provided a new function that is now named ares_cancel(). It is
801 used to cancel/cleanup a resolve/request made using ares functions on the
802 given ares channel. It does not destroy/kill the ares channel itself.
804 - Dominick Meglio cleaned up the formatting in several man pages.
807 - Dominick Meglio's new ares_expand_string. A helper function when decoding
808 incoming DNS packages.
810 - Daniel Stenberg modified the Makefile.in to use a for loop for the man page
811 installation to improve overview and make it easier to add man pages.
813 Version 1.1.0 (March 11, 2004)
816 - Gisle Vanem improved build on Windows.
819 - Dan Fandrich found a flaw in the Feb 22 fix.
821 - Added better configure --enable-debug logic (taken from the curl configure
822 script). Added acinclude.m4 to the tarball.
825 - Removed ares_free_errmem(), the function, the file and the man page. It was
826 not used and it did nothing.
828 - Fixed a lot of code that wasn't "64bit clean" and thus caused a lot of
829 compiler warnings on picky compilers.
832 - Dominick Meglio made ares init support multiple name servers in the
833 NameServer key on Windows.
836 - Modified ares_private.h to include libcurl's memory debug header if
837 CURLDEBUG is set. This makes all the ares-functions supervised properly by
838 the curl test suite. This also forced me to add inclusion of the
839 ares_private.h header in a few more files that are using some kind of
840 memory-related resources.
842 - Made the makefile only build ahost and adig if 'make demos' is used.
845 - Dirk Manske made ares_version.h installed with 'make install'
848 - ares_free_errmem() is subject for removal, it is simply present for future
849 purposes, and since we removed the extra parameter in strerror() it won't
851 - configure --enable-debug now enables picky compiler options if gcc is used
852 - fixed several compiler warnings --enable-debug showed and Joerg Mueller-Tolk
855 Version 1.0.0 (February 3, 2004)
858 - now we produce the libcares.a library instead of the previous libares.a
859 since we are no longer compatible
863 - ares_strerror() has one argument less. This is the first official
864 modification of the existing provided ares API.
868 - Dirk Manske fixed how the socket is set non-blocking.
872 - Dominick Meglio made the private gettimeofday() become ares_gettimeofday()
873 instead in order to not pollute the name space and risk colliding with
874 other libraries' versions of this function.
876 * October 24, 2003. Daniel Stenberg
878 Added ares_version().
880 Version 1.0-pre1 (8 October 2003)
882 - James Bursa made it run on RISC OS
884 - Dominick Meglio made it run fine on NT4
886 - Duncan Wilcox made it work fine on Mac OS X
888 - Daniel Stenberg adjusted the windows port
890 - liren at vivisimo.com made the initial windows port
892 * Imported the sources from ares 1.1.1