summaryrefslogtreecommitdiff
path: root/doc/dev
Commit message (Collapse)AuthorAgeFilesLines
...
* | doc/dev/notes: keymgr-certificates: Apply suggestions from Diziet.gabi-2502024-11-141-2/+2
| |
* | doc/dev/notes: keymgr-certificates: Clarify code docs.Gabriela Moldovan2024-11-141-4/+8
| |
* | doc/dev/notes: keymgr-certificates: Fix typo.Gabriela Moldovan2024-11-141-1/+1
| |
* | doc/dev/notes: keymgr-certificates: Remove unreadable chunk from intro.Gabriela Moldovan2024-11-141-8/+3
| |
* | doc/dev/notes: keymgr-certificates: Remove unnecessary args.Gabriela Moldovan2024-11-141-4/+0
| | | | | | | | | | We don't actually need a `KeyType` argument (`KeyTypes` are obtained from the `EncodableKey::key_type()` impl of the key/cert).
* | doc/dev/notes: keymgr-certificates: Reorder args for clarity.Gabriela Moldovan2024-11-141-4/+4
| |
* | doc/dev/notes: keymgr-certificates: Move subject key specifier to trait.Gabriela Moldovan2024-11-141-2/+2
| | | | | | | | This helps us cut down on the number of args in `get_*_and_cert()`.
* | doc/dev/notes: keymgr-certificates: Say that cert paths are derived from key ↵Gabriela Moldovan2024-11-141-26/+39
| | | | | | | | paths.
* | doc/dev/notes: keymgr-certificates: Mention the uid we settled on in !2577.Gabriela Moldovan2024-11-141-5/+5
| |
* | doc/dev/notes: keymgr-certificates: Remove outdated design.Gabriela Moldovan2024-11-141-191/+33
| | | | | | | | | | | | | | | | We decided to go with the second proposed design, so we can safely remove this one now. This also updates the introduction to summarize how we arrived at this design.
* | doc/dev/notes: keymgr-certificates: Clarify confusing wording.Gabriela Moldovan2024-11-141-3/+2
| |
* | doc/dev/notes: keymgr-certificates: Make the signing_key optionally present ↵Gabriela Moldovan2024-11-141-3/+37
| | | | | | | | in keystore.
* | doc/dev/notes: keymgr-certificates: Remove redundant fields from ↵Gabriela Moldovan2024-11-141-5/+0
| | | | | | | | KeyCertificateSpecifier.
* | doc/dev/notes: keymgr-certificates: Clarify what KeyType/CertType is about.Gabriela Moldovan2024-11-141-0/+7
| |
* | doc/dev/notes: keymgr-certificates: Remove misleading statement.Gabriela Moldovan2024-11-141-1/+0
| |
* | doc/dev/notes: keymgr-certificates: Add a revised proposal.Gabriela Moldovan2024-11-141-0/+358
| | | | | | | | | | Based on Diziet's feedback in https://gitlab.torproject.org/tpo/core/arti/-/merge_requests/2565#note_3099397
* | doc/dev/notes: keymgr-certificates: Say that the original proposal is rejected.Gabriela Moldovan2024-11-141-7/+12
| |
* | doc/dev/notes: keymgr-certificates: Update key notation for clarity.Gabriela Moldovan2024-11-141-6/+8
| |
* | doc/dev/notes: keymgr-certificates: Add note about certificate management.Gabriela Moldovan2024-11-141-0/+177
| | | | | | | | | | | | | | This describes the storage format and on-disk key store paths of `relaysign_ed` certificates (which we'll soon need to support). Part of #1617
* | rpc-connect-sketch: clarify "no new critical members" principle.Nick Mathewson2024-11-131-5/+11
| |
* | rpc-connect-sketch: Per discussion, change non-localhost to "decline"Nick Mathewson2024-11-121-1/+1
| |
* | rpc-connect-sketc: Add question on TOML tolerance.Nick Mathewson2024-11-071-0/+6
| |
* | rpc-connect-sketch: more specific abort/decline documentationNick Mathewson2024-11-071-6/+12
| |
* | rpc-connect-sketch: more on philosophy of client error handlingNick Mathewson2024-11-071-2/+34
| |
* | rpc-cookie: shell expansion happens before the addr is canonicalNick Mathewson2024-11-071-0/+2
| |
* | rpc-connect: Clarify server behavior on disabled entries.Nick Mathewson2024-11-071-0/+2
|/ | | | | The behavior in question is to ignore the entries entirely, and not check their files at all.
* Grammar fix.Nick Mathewson2024-10-311-1/+1
|
* RPC spec: note that embedded/owned need options.Nick Mathewson2024-10-311-1/+6
|
* RPC spec: clarify what is escapedNick Mathewson2024-10-311-2/+7
| | | | | Specifically, only literal connect points (TOML strings) are escaped.
* RPC specs: Fix an inline TOML table syntax.Nick Mathewson2024-10-311-1/+1
|
* rpc: clarify percent-encoding rules.Nick Mathewson2024-10-281-5/+10
|
* rpc: refer to TOML tables, not sections.Nick Mathewson2024-10-281-5/+5
|
* rpc: centralize treatment of unrecognized sections/members.Nick Mathewson2024-10-281-2/+2
|
* rpc: clarify "hidden file" meaning.Nick Mathewson2024-10-281-2/+3
|
* rpc: tweak "connect string" terminologyNick Mathewson2024-10-282-75/+78
| | | | | With all the fanciness we've built up here here, it seems that "connect point" is a better term than "connect point".
* rpc-connect-sketch: convert to toml.Nick Mathewson2024-10-271-50/+61
| | | | | (We've decided to make this change because these files are meant to be written by humans.)
* Move RPC documents into a new locationNick Mathewson2024-10-243-0/+0
| | | | This will be the start of an mdbook document.
* Merge branch 'rpc-server-sketch' into 'main'Nick Mathewson2024-10-242-15/+201
|\ | | | | | | | | | | | | Sketch RPC server and superuser operations. Closes #1702 See merge request tpo/core/arti!2563
| * rpc-sketch: clean up some ungrammatical/hypothetical language.Nick Mathewson2024-10-241-2/+2
| |
| * rpc sketch: show how to override a connect point option.Nick Mathewson2024-10-241-0/+5
| |
| * rpc-sketch: specify and permit directories full of connect files.Nick Mathewson2024-10-241-10/+14
| | | | | | | | | | | | | | | | | | | | | | The main purpose of this commit is to RPC servers to configure connect points individually, while allowing clients to look at a configuration location that is full of multiple connect points. This change permits us to discard allow_missing_file, which is nice. This commit still pretends that the file extension is "json"; that will be fixed in #1711.
| * rpc-connect: revise description of superuser and fs-mistrust options.Nick Mathewson2024-10-231-5/+17
| |
| * rpc-connect: Rewrite server behavior section.Nick Mathewson2024-10-231-16/+58
| | | | | | | | | | | | | | | | | | | | | | | | | | | | The most salient change here is that _everything_, including the default connect string, is a configuration option, so that `--system-arti` (hypothetical) can override it. In order to make the default connect file locations behave like regular options, I've added a `allow_missing_file` sub-option that makes the absence of a file tolerable. There is a behavioral change here: Specifying your own `rpc.listen` entries in the arti.toml _does not_ disable checking the default connect file locations.
| * rpc-connect: Mention #1710 when talking about system-arti case.Nick Mathewson2024-10-231-0/+6
| |
| * rpc-connect-sketch: normalize some jsonNick Mathewson2024-10-211-6/+6
| |
| * Apply 2 suggestion(s) to 2 file(s)Nick Mathewson2024-10-212-2/+2
| | | | | | Co-authored-by: Ian Jackson <[email protected]>
| * rpc-*: sketch superuser operations.Nick Mathewson2024-10-212-1/+39
| |
| * rpc-connect-sketch: Sketch RPC server operationNick Mathewson2024-10-211-0/+44
| |
| * rpc-connect-sketch: Fix ambiguous use of "embedded"Nick Mathewson2024-10-211-1/+1
| |
| * rpc-connect-sketch: more sketching about embedded arti.Nick Mathewson2024-10-211-9/+25
| |