langfingaz/anki
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
aaf5770db9
anki/rslib/Cargo.toml

82 lines
2.4 KiB
TOML
Raw Normal View History

split the rust code into a workspace
2019-12-23 05:42:14 +01:00
[package]
move the separate components back into this monorepo Earlier today I pushed a change that split this code up into multiple repos, but that has proved to complicate things too much. So we're back to a single repo, except the individual submodules are better separated than they were before. The README files need updating again; I will push them out soon. Aside from splitting out the different modules, the sound code has moved from from anki to aqt.
2020-01-02 10:43:19 +01:00
name = "anki"
bump version
2020-10-02 02:22:21 +02:00
version = "2.1.36" # automatically updated
split the rust code into a workspace
2019-12-23 05:42:14 +01:00
edition = "2018"
use support URL and point to source repo
2020-04-03 00:44:59 +02:00
authors = ["Ankitects Pty Ltd and contributors"]
move the separate components back into this monorepo Earlier today I pushed a change that split this code up into multiple repos, but that has proved to complicate things too much. So we're back to a single repo, except the individual submodules are better separated than they were before. The README files need updating again; I will push them out soon. Aside from splitting out the different modules, the sound code has moved from from anki to aqt.
2020-01-02 10:43:19 +01:00
license = "AGPL-3.0-or-later"
Created the GitHub Actions step Upload to PyPi to run only on tagged releases of anki: https://anki.tenderapp.com/discussions/ankidesktop/40025-anki-in-python-package-index
2020-04-02 04:01:48 +02:00
description = "Anki's Rust library code"
readme = "README.md"
use protobuf for python/rust bridge If we go ahead with betterproto on the Python side, it will mean bumping the minimum Python dependency to 3.7.
2019-12-24 04:33:44 +01:00
[dependencies]
vendor coarsetime so we can get the use after free fix
2020-07-13 05:34:37 +02:00
# pinned as any changes could invalidate sqlite indexes
unicase = "=2.6.0"
update rslib/rspy deps
2020-07-12 03:38:25 +02:00
nom = "5.1.2"
failure = "0.1.8"
media sync working, but unpolished
2020-02-01 10:05:22 +01:00
prost = "0.6.1"
update rslib/rspy deps
2020-07-12 03:38:25 +02:00
bytes = "0.5.5"
chrono = "0.4.13"
flatten->render; pass field content in This is paving the way to move the standard filters into Rust. Non-empty fields are now determined in Rust, using a single regex instead of the overkill stripHTMLMedia(). The old implementation has been moved into the Pystache code for now.
2020-01-10 05:59:29 +01:00
lazy_static = "1.4.0"
update rslib/rspy deps
2020-07-12 03:38:25 +02:00
regex = "1.3.9"
hex = "0.4.2"
blake3 = "0.3.5"
add a simple TTS field filter This simply wraps the field in extra text that the frontend will deal with. Also added some helpers for extracting and stripping audio and TTS tags from the rendered text.
2020-01-20 10:12:34 +01:00
htmlescape = "0.3.1"
add another implementation of media.addFile() and cleanFilename() et al Instead of adding an incrementing number in the conflict case, the file hash is appended.
2020-01-28 11:52:10 +01:00
sha1 = "0.6.0"
update rslib/rspy deps
2020-07-12 03:38:25 +02:00
unicode-normalization = "0.1.13"
add another implementation of media.addFile() and cleanFilename() et al Instead of adding an incrementing number in the conflict case, the file hash is appended.
2020-01-28 11:52:10 +01:00
tempfile = "3.1.0"
update rslib/rspy deps
2020-07-12 03:38:25 +02:00
serde = "1.0.114"
serde_json = "1.0.56"
use a shared async runtime instead of recreating each time
2020-06-02 01:35:27 +02:00
tokio = { version = "0.2.21", features = ["fs", "rt-threaded"] }
update rslib/rspy deps
2020-07-12 03:38:25 +02:00
serde_derive = "1.0.114"
rslib: disable bzip2 feature We can't link against libzip2.so.1 when building manylinux wheels, but the only dependency which required this was "zip" and we don't appear to make use of bzip2 in Anki. This fixes the following "make build" error on Linux: x maturin failed Caused by: Failed to ensure manylinux compliance Caused by: Your library is not manylinux compliant because it links the following forbidden libraries: ["libbz2.so.1"] Signed-off-by: Aleksa Sarai <cyphar@cyphar.com>
2020-09-02 09:02:10 +02:00
zip = { version = "0.5.6", default-features = false, features = ["deflate", "time"] }
update rslib/rspy deps
2020-07-12 03:38:25 +02:00
serde_tuple = "0.5.0"
vendor coarsetime so we can get the use after free fix
2020-07-13 05:34:37 +02:00
coarsetime = { git = "https://github.com/ankitects/rust-coarsetime.git", branch="old-mac-compat" }
update rslib/rspy deps
2020-07-12 03:38:25 +02:00
utime = "0.3.1"
gather field references in Rust; media check now mostly complete
2020-02-10 05:19:39 +01:00
serde-aux = "0.6.1"
update to latest fluent-rs and add basic locale-aware decimals - git version pinned at the moment until the concurrency fix lands in 0.10.2 - currently float values are hard-coded at 2 decimal places; we should switch to using NUMBER() in the future
2020-02-20 01:42:16 +01:00
unic-langid = { version = "0.8.0", features = ["macros"] }
add workaround for panic on 32 bit systems https://anki.tenderapp.com/discussions/beta-testing/1817-failed-to-generate-operands-out-of-fluentnumber-could-not-convert-string-to-integer
2020-02-25 03:57:35 +01:00
fluent = { git = "https://github.com/ankitects/fluent-rs.git", branch="32bit-panic" }
intl-memoizer = { git = "https://github.com/ankitects/fluent-rs.git", branch="32bit-panic" }
update to latest fluent-rs and add basic locale-aware decimals - git version pinned at the moment until the concurrency fix lands in 0.10.2 - currently float values are hard-coded at 2 decimal places; we should switch to using NUMBER() in the future
2020-02-20 01:42:16 +01:00
num-format = "0.4.0"
make logging explicit, and support logging to a file
2020-02-29 10:50:15 +01:00
slog = { version = "2.5.2", features = ["max_level_trace", "release_max_level_debug"] }
update rslib/rspy deps
2020-07-12 03:38:25 +02:00
slog-term = "2.6.0"
slog-async = "2.5.0"
make logging explicit, and support logging to a file
2020-02-29 10:50:15 +01:00
slog-envlogger = "2.2.0"
update rslib/rspy deps
2020-07-12 03:38:25 +02:00
serde_repr = "0.1.6"
num_enum = "0.5.0"
futures = "0.3.5"
the start of note adding, and note type changes
2020-04-16 09:02:59 +02:00
rand = "0.7.3"
update rslib/rspy deps
2020-07-12 03:38:25 +02:00
num-integer = "0.1.43"
template changes and card generation Cloze cards are not yet supported, missing decks are not handled, and more testing is still required.
2020-04-20 13:32:55 +02:00
itertools = "0.9.0"
update rslib/rspy deps
2020-07-12 03:38:25 +02:00
flate2 = "1.0.16"
pin-project = "0.4.22"
async-compression = { version = "0.3.5", features = ["stream", "gzip"] }
askama = "0.10.1"
panic fix was merged into hyper 0.13.7
2020-07-16 01:09:13 +02:00
hyper = "0.13.7"
rs: remove most &mut references in BackendService The previous implementation had some slightly questionable memory safety properties (older versions of PyO3 didn't uphold the Rust aliasing rules and would thus create multiple &mut references to #[pyclass] objects). This explains why Backend has internal Mutex<T>s even though all of its methods took &mut self. The solution is to simply make all methods take &self, which luckily doesn't pose too make issues -- most of the code inside Backend already has sufficient locking. The only two things which needed to be explicitly handled where: 1. "self.runtime" which was fairly easy to handle. All usages of the Runtime only require an immutable reference to create a new Handle, so we could switch to OnceCell which provides lazy-initialisation semantics without needing a more heavy-handed Mutex<tokio::runtime::Handle>. 2. "self.sync_abort" was simply wrapped in a Mutex<>, though some of the odd semantics of sync_abort (not being able to handle multiple processes synchronising at the same time) become pretty obvious with this change (for now we just log a warning in that case). In addition, switch to an RAII-style guard to make sure we don't forget to clear the abort_handle. As a result, we now no longer break Rust's aliasing rules and we can build with newer versions of PyO3 which have runtime checks for these things (and build on stable Rust). Signed-off-by: Aleksa Sarai <cyphar@cyphar.com>
2020-09-04 08:20:39 +02:00
once_cell = "1.4.1"
scopeguard = "1.1.0"
use protobuf for python/rust bridge If we go ahead with betterproto on the Python side, it will mean bumping the minimum Python dependency to 3.7.
2019-12-24 04:33:44 +01:00
handle fields with embedded nuls
2020-03-24 05:53:33 +01:00
[target.'cfg(target_vendor="apple")'.dependencies.rusqlite]
update rusqlite
2020-05-16 11:49:08 +02:00
version = "0.23.1"
handle fields with embedded nuls
2020-03-24 05:53:33 +01:00
features = ["trace", "functions", "collation"]
comply with manylinux wheel requirements, and vendor sqlite on Windows
2020-02-05 06:01:40 +01:00
handle fields with embedded nuls
2020-03-24 05:53:33 +01:00
[target.'cfg(not(target_vendor="apple"))'.dependencies.rusqlite]
update rusqlite
2020-05-16 11:49:08 +02:00
version = "0.23.1"
handle fields with embedded nuls
2020-03-24 05:53:33 +01:00
features = ["trace", "functions", "collation", "bundled"]
comply with manylinux wheel requirements, and vendor sqlite on Windows
2020-02-05 06:01:40 +01:00
vendor reqwest for now, to support read/write timeouts https://github.com/seanmonstar/reqwest/issues/870 adding back now that fix has been applied to a fork of tokio-io-timeout
2020-08-26 05:24:51 +02:00
[target.'cfg(linux)'.dependencies.reqwest]
git = "https://github.com/ankitects/reqwest.git"
use commit hash in reqwests dependency
2020-08-26 10:49:23 +02:00
rev = "57665e2c2a39db85723ba860f1b570a608bb73f9"
vendor reqwest for now, to support read/write timeouts https://github.com/seanmonstar/reqwest/issues/870 adding back now that fix has been applied to a fork of tokio-io-timeout
2020-08-26 05:24:51 +02:00
features = ["json", "socks", "stream", "native-tls-vendored"]
comply with manylinux wheel requirements, and vendor sqlite on Windows
2020-02-05 06:01:40 +01:00
vendor reqwest for now, to support read/write timeouts https://github.com/seanmonstar/reqwest/issues/870 adding back now that fix has been applied to a fork of tokio-io-timeout
2020-08-26 05:24:51 +02:00
[target.'cfg(not(linux))'.dependencies.reqwest]
git = "https://github.com/ankitects/reqwest.git"
use commit hash in reqwests dependency
2020-08-26 10:49:23 +02:00
rev = "57665e2c2a39db85723ba860f1b570a608bb73f9"
vendor reqwest for now, to support read/write timeouts https://github.com/seanmonstar/reqwest/issues/870 adding back now that fix has been applied to a fork of tokio-io-timeout
2020-08-26 05:24:51 +02:00
features = ["json", "socks", "stream"]
comply with manylinux wheel requirements, and vendor sqlite on Windows
2020-02-05 06:01:40 +01:00
use protobuf for python/rust bridge If we go ahead with betterproto on the Python side, it will mean bumping the minimum Python dependency to 3.7.
2019-12-24 04:33:44 +01:00
[build-dependencies]
Update prost-build version to 0.6 This is the same version as Prost, and lets us reduce the number of deps we have to compile
2020-02-27 04:48:16 +01:00
prost-build = "0.6.1"
update rslib/rspy deps
2020-07-12 03:38:25 +02:00
fluent-syntax = "0.9.3"
move the separate components back into this monorepo Earlier today I pushed a change that split this code up into multiple repos, but that has proved to complicate things too much. So we're back to a single repo, except the individual submodules are better separated than they were before. The README files need updating again; I will push them out soon. Aside from splitting out the different modules, the sound code has moved from from anki to aqt.
2020-01-02 10:43:19 +01:00
full upload/download HTTP code
2020-05-27 11:32:05 +02:00
[dev-dependencies]
env_logger = "0.7.1"
Reference in New Issue Copy Permalink