aboutsummaryrefslogtreecommitdiff
path: root/path-spec.txt
AgeCommit message (Collapse)Author
2023-10-12Move all text-only specifications into the OLD_TXT directory.Nick Mathewson
2021-12-08Clarify the behavior of some circuit timeout params.Nick Mathewson
In particular, this commit clarifies that the ranges of some parameters are such that choosing a very high value will, in effect, disable parts of the circuit timeout inference code.
2021-10-25Fix typos and cleanupDimitris Apostolou
2021-10-22better name for the num-primary-guards variableRoger Dingledine
2021-09-22path-spec: Add 'Table of contents'Oleh Franchuk
2021-09-22path-spec: Fix a typoOleh Franchuk
2021-09-07path-spec corrections from Mike.Nick Mathewson
2021-08-27cbtinitialtimeout also affects the default abandon timeout.Nick Mathewson
2021-08-27path-spec: describe the behavior of circuit_expire_building()Nick Mathewson
This is my attempt to extract a general concise rule here. (I am not sure that Tor adjusts both timeout_ms and close_ms, but Arti does indeed adjust both.)
2021-08-27Describe reload behavior of path-spec data.Nick Mathewson
2021-08-27Clarify Xm tiebreaker rules.Nick Mathewson
2021-08-27Note that only three-hop circuits are used as build time observations.Nick Mathewson
2021-08-27Explain when circuits are torn down, and how they become dirty.Nick Mathewson
2021-08-17Clarify tiebreaker rules when computing X_mNick Mathewson
2021-08-16Specify the exact "fraction of paths" calculation that Tor uses.Nick Mathewson
2021-07-13Remove tracking timestamps for incoming network activity.Nick Mathewson
This is never necessary, since we only count circuits who have completed at least one hop. Closes torspec#57
2021-05-13Clarify exit-first, then front-to-back selectionNick Mathewson
2021-02-18Update path-spec.txt for CBT estimator fixes for #40168.Mike Perry
Also clarify and improve wording of the timeout calculation section.
2020-11-16fix typos; more robust links between specsRoger Dingledine
2019-12-10convert text blocks into widely compatible "blockquote" syntaxHans-Christoph Steiner
This only adds newline characters to make the existing text blocks act like "blockquote" or "code block" syntax in Markdown, asciidoc, and others. This was accomplished by manually reviewing the output of this script: ```bash for f in *.txt; do cat $f | python -c "import sys,re;print(re.sub(r'(\n {0,3}[^ \n][^\n]*\n)( {4,}[^\n]*)', r'\1\n\2', sys.stdin.read()))" > ${f}.tmp mv ${f}.tmp $f done ```
2019-11-26remove trailing whitespace `sed -i 's, *$,,' *.txt`Hans-Christoph Steiner
2019-11-26fix erroneous header numbering punctuationHans-Christoph Steiner
The clear standard is trailing "." after each numeric section. This fixes the small handful of outliers. This makes it easy to convert these headers to common markup formats, for example: http://hyperpolyglot.org/lightweight-markup
2018-05-08Mention how HSLayer2Nodes and HSLayer3Nodes alter path restrictions.Mike Perry
2018-02-14Adjust cbtlearntimeout default to match the codeNick Mathewson
2018-02-14Document new consensus params for bug24769.Mike Perry
2017-02-01Explain more about primary guards and about building circuitsNick Mathewson
In path-spec: explain our rules (post-21242) for waiting to build circuits. In guard-spec: - explain what to do about missing descriptors - explain parallel use of multiple primary guards, based on parameters.
2017-01-31Start a new guard-spec.txt as a copy of prop271.Nick Mathewson
Remove the old guard section of path-spec, now that guard-spec is separate.
2015-10-23Document our current guard selection algorithm in path-spec.txt.Isis Lovecruft
* ADDS new section, "ยง5.1. Guard selection algorithm", to path-spec.txt. * FIXES #17261: https://bugs.torproject.org/17261
2015-02-24Standardizing on the name 'server descriptor'Damian Johnson
We use a few names for the main self-published descriptor documents... https://trac.torproject.org/projects/tor/ticket/14987 Standardizing on the name 'server descriptor'. This change was made with a few quick runs of sed... % find ./* -type f -exec sed -i 's/router descriptor/server descriptor/' "{}" +; % find ./* -type f -exec sed -i 's/Router descriptor/Server descriptor/' "{}" +; % find ./* -type f -exec sed -i 's/Router Descriptor/Server Descriptor/' "{}" +; ... then reverting the attic changes, and scanning the rest to be sure it doesn't include any functional impact.
2014-12-30Actually, this patch is going into 0.2.6.2-alpha, not 0.2.6.1-alphaNick Mathewson
2014-12-20If the consensus doesn't contain exits, don't build exit pathsteor
If the consensus doesn't contain exits, we only build internal paths. This is enough to allow reachability tests (which can enable exits to bootstrap), and hidden services. If we subsequently receive a consensus with exits, start building exit paths. Update dir-spec and path-spec to document this. Update control-spec to document changes in controller bootstrap messages. Based on changes made in tor to resolve bug #13814.
2014-11-26Set default value for bw weights if absent from consensus.George Kadianakis
2013-04-18Define 'family' better in path-specNick Mathewson
Inspired by tor-talk thread "Family path selection"
2013-04-11Merge remote-tracking branch 'mikeperry/bug5178'Nick Mathewson
2013-03-27Describe path bias consensus parameters and clarify some wording.Mike Perry
2013-03-26Improve CBT network liveness section to match code.Mike Perry
We simplified this code a while back and forgot to update the spec.
2013-03-25Initial draft of path bias spec updates.Mike Perry
Still needs parameter description.
2011-11-26Trivial whitespace cleanupSebastian Hahn
2011-05-09Fix minor issuesRobert Ransom
2011-05-09In specs, do not say "server" when we mean "relay" or "node"Nick Mathewson
Fixes bug 2936.
2011-01-15Fix a typo spotted by RogerSebastian Hahn
2011-01-15Sanity-check consensus param valuesSebastian Hahn
We need to make sure that the worst thing that a weird consensus param can do to us is to break our Tor (and only if the other Tors are reliably broken in the same way) so that the majority of directory authorities can't pull any attacks that are worse than the DoS that they can trigger by simply shutting down. One of these worse things was the cbtnummodes parameter, which could lead to heap corruption on some systems if the value was sufficiently large. This commit fixes this particular issue and also introduces sanity checking for all consensus parameters.
2010-06-29Merge remote branch 'mikeperry/cbt-bugfixes3'Nick Mathewson
2010-06-25Make spec docs reference RFC2119Nick Mathewson
Now people will have no excuse for not knowing what "MUST" and "MAY" mean. Fixes bug 1310.
2010-06-09Update spec with new right-censored pareto estimators.Mike Perry
2010-05-10Bug 1296: Add option+logic to disable CBT learning.Mike Perry
There are now four ways that CBT can be disabled: 1. Network-wide, with the cbtdisabled consensus param. 2. Via config, with "LearnCircuitBuildTimeout 0" 3. Via config, with "AuthoritativeDirectory 1" 4. Via a state file write failure.
2010-05-10Update path-spec.txt with contents of proposal 151.Mike Perry
2010-02-22Update specs for weight computation and use.Mike Perry
2009-06-22path-spec: Clarify what we mean by "a server's bandwidth."Nick Mathewson
This just got a little complicated, since old clients use "clipped advertised bandwith" and new clients now use "consensus bandwidth" but fall back to "clipped advertised bandwidth".
2009-05-05Remove all svn metadata minus what I missed.Sebastian Hahn
Tor doesn't use SVN anymore, making $Revision$, $Id$ and $Date$ meaningless. Remove them without replacement.