mirrors/arti.git/crates/tor-netdoc/src/test2.rs, branch arti-v2.1.0mirror of https://gitlab.torproject.org/tpo/core/arti
http://git.dilluti0n.com/mirrors/arti.git/atom?h=arti-v2.1.02026-01-14T14:45:54Ztor-netdoc: derives: New `skip` document field option2026-01-14T14:45:54ZIan Jacksonijackson@chiark.greenend.org.uk2026-01-07T12:39:59Zurn:sha1:d209f26e1c01c5c337418f3c99c85aec31613c66
This lets us having document items that are "manually non exhaustive"
which is necessary for struct literal constructors.
tor-netdoc: test2: Make t_ok (which takes no boundaries) take &D not &[D]2025-12-18T17:30:44ZIan Jacksonijackson@chiark.greenend.org.uk2025-12-17T16:18:01Zurn:sha1:39e1c2de26ce626bdd84550aaefb43daa8782a34tor-netdoc: test2: Test parse_netdoc_multiple_with_offsets2025-12-18T17:30:44ZIan Jacksonijackson@chiark.greenend.org.uk2025-12-17T16:16:04Zurn:sha1:f5de3335e79c1b23da7b62d8074df2704bcfa2ec
* Make a fancy version of t_ok, t_ok_multi, that takes the expected boundary byte offsets.
* t_ok is now just for single-document files and is implemented in terms of t_ok_multi.
* Use t_ok_multi for the one test case with multiple documents.
tor-netdoc: test2: Add a test file with more than one document2025-12-18T17:28:36ZIan Jacksonijackson@chiark.greenend.org.uk2025-12-17T16:04:54Zurn:sha1:3ec2461f501f6ea21279acfa4d2fa87649eb785ator-netdoc: rustfmt2025-12-18T16:44:55ZIan Jacksonijackson@chiark.greenend.org.uk2025-12-16T16:22:55Zurn:sha1:653d6b74784ee5ab33f065f87aeea9dbf1157474tor-netdoc: Use $P for prelude in one place2025-12-18T16:44:55ZIan Jacksonijackson@chiark.greenend.org.uk2025-12-10T14:38:46Zurn:sha1:5ec68ca2391068f90683fc406fb6f38bd4074050
This was missed, apparently.
tor-netdoc: test2: check round-trip encoding2025-12-10T11:05:28ZIan Jacksonijackson@chiark.greenend.org.uk2025-12-01T13:07:28Zurn:sha1:10cb24c401baead05296e7e4fd8251bd839fc0f4
The test case documents are often not in canonical form. We don't
want to change them to be in canonical form because we want to test
parsing edge cases. So, we provide a mangling scheme with in-document
notes, and annotate all the documents.
Apply suitable annotations to all the documents.
Test that the encoded forms are as expected.
This reverts commit 05e7b8b60061a7e8a8d004a822cc213d3ea22d47.
tor-netdoc: test2: implement encoding traits for all test types2025-12-10T11:05:28ZIan Jacksonijackson@chiark.greenend.org.uk2025-11-25T12:07:03Zurn:sha1:04d39e31038472344e3c37e8c2d7becdc45e7ffc
Make this whole module conditional on the encode feature.
Apply the derives. Provide the necessary manual implementations for
the types where we use `with`.
tor-netdoc: test2: Don't ever have raw String in defaulted arguments2025-12-04T12:54:14ZIan Jacksonijackson@chiark.greenend.org.uk2025-12-01T11:14:03Zurn:sha1:0be1ad7b438470bb9871fd4cbdf42597d4d76c23
These cannot be encoded. So that is logically incoherent.
(Perhaps String ought not to be NormalItemArgument, but let's not
tackle that now.)
parse2: ItemValueParseable, netdoc(rest, with): rework2025-12-02T17:34:21ZIan Jacksonijackson@chiark.greenend.org.uk2025-11-26T14:12:20Zurn:sha1:c44465c1d7f397e04fa3f2fd8cd55f26968c179b
The path in `with` must always be a module, not a function, because we
need to use it for encoding too.
In the case of `rest`, it's probably sensible to require a different
function name, given that the semantics are subtly different to those
of `rest` without `with`, since the latter can only get a single word
but with `rest` you get the whole line.