summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAge
* Mention in INSTALL that tlsdate is expected to work on FreeBSD 9.2 and 11 as ↵HEADmasterFabian Keil2013-11-21
| | | | well
* Mention the recent FreeBSD-specific changes in the CHANGELOGFabian Keil2013-11-21
|
* Use the strchrnul() replacement on FreeBSD versions that need itFabian Keil2013-11-21
| | | | This gets tlsdate and tlsdate-helper working on FreeBSD 9.2-RELEASE.
* Include <sys/stat.h> in src/util.cFabian Keil2013-11-21
| | | | | | | | | | | On FreeBSD 11-CURRENT this fixes the compilation failure: src/util.c: In function 'file_write': src/util.c:256: error: 'S_IRUSR' undeclared (first use in this function) src/util.c:256: error: (Each undeclared identifier is reported only once src/util.c:256: error: for each function it appears in.) src/util.c:256: error: 'S_IWUSR' undeclared (first use in this function) *** Error code 1
* Include <unistd.h> in src/util.hFabian Keil2013-11-21
| | | | | | | | | | | | | On FreeBSD 11-CURRENT this fixes the compilation failure: In file included from ./src/proxy-bio.h:16, from src/proxy-bio.c:38: ./src/util.h:46: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'wait_with_timeout' ./src/util.h:50: warning: 'struct timeval' declared inside parameter list ./src/util.h:50: warning: its scope is only this definition or declaration, which is probably not what you want ./src/util.h:51: warning: 'struct timeval' declared inside parameter list ./src/util.h:57: warning: 'struct timeval' declared inside parameter list *** Error code 1
* Add note about servers with weak DH group parametersJacob Appelbaum2013-11-07
|
* remove unused code; valgrind complained about this lazy sloptlsdate-0.0.7Jacob Appelbaum2013-11-02
|
* Bump to version 0.0.7Jacob Appelbaum2013-11-02
|
* Add verbose change to CHANGELOGJacob Appelbaum2013-11-02
|
* Update README to make it timeless, so to speakJacob Appelbaum2013-11-02
|
* Add verb_debug and give -vv meaningJacob Appelbaum2013-11-01
|
* Add tlsdated.service to MakefileJacob Appelbaum2013-10-31
|
* update CHANGELOGJacob Appelbaum2013-10-31
|
* Add /usr/lib/x86_64-linux-gnu/ to AppArmor profile\Jacob Appelbaum2013-10-31
| | | | This partially addresses Debian bug #704680
* add systemd service file for tlsdatedJacob Appelbaum2013-10-31
|
* Update .gitignore to reflect ctags/vim usageJacob Appelbaum2013-10-30
|
* Add README note about HTTP Date parsingJacob Appelbaum2013-10-30
|
* move defines into the proper headerJacob Appelbaum2013-10-30
|
* Add missing \nJacob Appelbaum2013-10-30
|
* Define a max length for the HTTP Date: headerIngy döt Net2013-10-30
|
* Merge pull request #121 from nmathewson/http_date\033]2;Jacob Appelbaum\0072013-10-30
|\ | | | | Rudimentary support for HTTP Date headers
| * Restrict loggable date fields even harder.Nick Mathewson2013-10-14
| |
| * Add a changelogNick Mathewson2013-10-10
| |
| * Make user-agent a configuration optionNick Mathewson2013-10-10
| |
| * Update manpage with http optionNick Mathewson2013-10-10
| |
| * In verbose mode, display TLS time as well as HTTPS timeNick Mathewson2013-10-10
| |
| * Document magic "1900" valueNick Mathewson2013-10-10
| |
| * Don't log raw dates, in case they contain junk.Nick Mathewson2013-10-10
| |
| * Describe what an HTTP date isNick Mathewson2013-10-10
| |
| * Change check for RTC to be configure based.Nick Mathewson2013-10-10
| | | | | | | | | | Previously I had disabled it with #ifdef __linux__, but that's not very good autoconf style.
| * Rudimentary support for HTTP Date headersNick Mathewson2013-09-18
|/ | | | | | | | | | | | | Since I'm going on a personal crusade to kill off gmt_unix_time, I should provide an alternative. That alternative can be the Date header from HTTP -- unlike gmt_unix_time, the Date header is required by the RFC to actually be an accurate clock-like clock, and nobody is trying to get rid of it. This code is pretty hack-ish and does some nonportable stuff, like using memmem() and timegm(). It's not super-tolerant of non-standards-compliant HTTP servers. I hope I didn't make any pointer mistakes.
* Merge pull request #119 from redpig/showtime=raw\033]2;Jacob Appelbaum\0072013-09-14
|\ | | | | tlsdate: add optional --showtime support for "raw"
| * tlsdate, docs: make -V taken an optional arg to simplify showtimeWill Drewry2013-09-13
| | | | | | | | | | Update the man pages and simplify the processing of -V and --showtime to share the optional argument path.
| * tlsdate: add optional --showtime support for "raw"Will Drewry2013-09-13
| | | | | | | | | | | | | | | | | | | | | | | | | | Add support for optional arguments to --showtime, "raw" and "human". Human readable output is what --showtime and -V do already. Passing in --showtime=raw will tell tlsdate-helper to write the raw unix time directly to stdout in host byte order. The purpose of this change is to provide easier integration of privilege isolation for time setting. If tlsdate is called from a wrapper, like tlsdated, it should be possible to pass the time back to a privileged time-setter with the minimum of needed parsing.
* | Merge pull request #118 from elly/event-loop\033]2;Jacob Appelbaum\0072013-09-14
|\ \ | |/ |/| Refactor event loop.
| * Refactor event loop.elly2013-07-17
|/ | | | | | | | | | Refactor the event loop to be modular and testable. Also, add support for detecting corruption of the realtime clock, as can be caused by suspend/resume cycles without an rtc battery. The event loop is now driven by a tree of events, which are either sources (currently suspend/resume events, periodic events, and network route events) or composite events. Signed-off-by: Elly Fong-Jones <elly@leptoquark.net>
* Merge pull request #117 from elly/integration-changelogJacob Appelbaum2013-07-08
|\ | | | | Update CHANGELOG for integration tests.
| * Update CHANGELOG for integration tests.elly2013-07-08
|/
* Merge pull request #116 from elly/integrationJacob Appelbaum2013-07-08
|\ | | | | Add integration testing.
| * Add integration testing.elly2013-07-08
|/ | | | | | These new integration tests mostly focus on tlsdated's subprocess code. Signed-off-by: Elly Fong-Jones <elly@leptoquark.net>
* Merge branch 'master' of github.com:ioerror/tlsdateJacob Appelbaum2013-06-30
|\
| * Merge pull request #114 from elly/refactorJacob Appelbaum2013-06-24
| |\ | | | | | | Fix subprocess watching.
| | * Fix subprocess watching.elly2013-06-21
| |/ | | | | | | | | | | | | Currently, the subprocess watching code polls with a delay between attempts. Instead, introduce wait_with_timeout() and use it, simplifying this code. Signed-off-by: Elly Fong-Jones <elly@leptoquark.net>
* | add verbosity for privilege dropping attemptJacob Appelbaum2013-06-29
| |
* | David Hill noticed that we sometimes leave &warp_time uninitalizedJacob Appelbaum2013-06-29
|/
* update policy per issue 109Jacob Appelbaum2013-05-11
|
* Add Elly as an authorJacob Appelbaum2013-05-11
|
* Merge pull request #110 from elly/unittest-rmrfJacob Appelbaum2013-05-11
|\ | | | | tlsdated: remove rmrf()
| * tlsdated: remove rmrf()elly2013-05-09
|/ | | | | | | | | | | | | This function, while handy and presently used safely, could become dangerous if someone later adds a call to it that passes in attacker-controlled input. Therefore, hardcode calls to unlink(2) and rmdir(2) for the files tlsdated presently creates. This has the pleasant (?) side-effect of breaking unit-tests whenever tlsdated creates files we don't explicitly list on disk. BUG=None TEST=unit Signed-off-by: Elly Fong-Jones <ellyjones@chromium.org>
* Fix from Paul Wouters to buildJacob Appelbaum2013-05-07
|
generated by cgit v1.2.1 (git 2.18.0) at 2024-02-02 09:27:19 +0000