Commit graph

14 commits

Author SHA1 Message Date
Vincent Ambo
012da7d0e1 feat(tazjin/german-string): add persistent construction from statics
Makes it possible to construct persistent German Strings from data with 'static
lifetime. This doesn't leak any additional data.

Change-Id: Ifb3a36ff72b106724883302e4c2195ac54acd352
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12263
Tested-by: BuildkiteCI
Reviewed-by: tazjin <tazjin@tvl.su>
Autosubmit: tazjin <tazjin@tvl.su>
2024-08-22 14:02:15 +00:00
Vincent Ambo
95ebcc24b0 refactor(tazjin/german-string): add type for encoding storage class
Adds a StorageClassPtr type that is set up to be able to steal bits from an
aligned pointer to encode the storage class of a German String.

Change-Id: I64591174eac1ebcb73e624a59bd107ba1e02c69d
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12262
Reviewed-by: tazjin <tazjin@tvl.su>
Autosubmit: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
2024-08-22 14:02:15 +00:00
Vincent Ambo
ed4f68b1d3 feat(tazjin/german-string): add constructor method from owned data
Makes it possible to construct a GermanString from an owned byte vector, without
having to clone the data.

This is done by "disowning" the vector using ManuallyDrop to access its internal
pointer. For transient strings, this memory is then owned (and freed) by the
GermanString instance.

Small strings are copied out of the heap and stored inline as before, to avoid
any dereferencing operations.

Change-Id: I754736099f71d646d430aed73e558a5a7626c394
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12249
Autosubmit: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
Reviewed-by: tazjin <tazjin@tvl.su>
2024-08-19 13:19:40 +00:00
Vincent Ambo
0d6d31bc66 feat(tazjin/german-string): add pointer equality check in Eq
This can short-circuit two large string comparisons.

Change-Id: If45e7cf33921fe571482dc710c27ef8cc7c70885
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12245
Autosubmit: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
Reviewed-by: tazjin <tazjin@tvl.su>
2024-08-19 13:19:40 +00:00
Vincent Ambo
a521f9e5f8 chore(tazjin/german-string): add Nix build
Change-Id: I1e8c500c9bb4ba365ba32f53f60bcf657f62ed62
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12243
Reviewed-by: tazjin <tazjin@tvl.su>
Autosubmit: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
2024-08-19 13:19:40 +00:00
Vincent Ambo
56d4f01e9f feat(tazjin/german-string): add [Partial]Ord implementation
Change-Id: I8002ec63be45a15160387c21ed1fa8721a7c3eb4
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12242
Tested-by: BuildkiteCI
Reviewed-by: tazjin <tazjin@tvl.su>
Autosubmit: tazjin <tazjin@tvl.su>
2024-08-19 13:19:40 +00:00
Vincent Ambo
86f467cee6 test(tazjin/german-string): add roundtrip proptests
Change-Id: Iee392368252d5c6e96d879b213aee34a301d13d6
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12241
Reviewed-by: tazjin <tazjin@tvl.su>
Autosubmit: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
2024-08-19 13:19:40 +00:00
Vincent Ambo
64a085cf52 feat(tazjin/german-string): add Eq impl & corresponding proptests
Change-Id: I66a258fad5d4e249268b9d2743d46b30c5ac1dac
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12240
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
Autosubmit: tazjin <tazjin@tvl.su>
2024-08-19 13:19:40 +00:00
Vincent Ambo
ab6a4815ff feat(tazjin/german-string): add Debug implementation
Change-Id: I9a4a8a4815ac210c402cfc2feb1fe0606affb327
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12239
Autosubmit: tazjin <tazjin@tvl.su>
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
2024-08-19 13:19:40 +00:00
Vincent Ambo
2284c00417 feat(tazjin/german-string): PartialEq implementation for GermanString
This is where one of the advantages of this string representation starts to
shine: For small strings there's no derefencing any heap memory at all, and for
the long representation we can compare the prefix before moving on.

Change-Id: Iac333a52e8d7c9dd09e33dbcf51754e321c880e6
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12238
Reviewed-by: tazjin <tazjin@tvl.su>
Autosubmit: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
2024-08-19 13:19:25 +00:00
Vincent Ambo
9aa1c2dd19 fix(tazjin/german-string): add maximum length check
Change-Id: I4fd9b2a31749d65632b662807c4e703a9df9ec9e
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12237
Reviewed-by: tazjin <tazjin@tvl.su>
Autosubmit: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
2024-08-19 11:50:11 +00:00
Vincent Ambo
d8640b6e67 feat(tazjin/german-string): add data accessors for &str and &[u8]
Change-Id: I992e625861f79ef6d9993e8caee4e02d3fc5557e
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12236
Tested-by: BuildkiteCI
Autosubmit: tazjin <tazjin@tvl.su>
Reviewed-by: tazjin <tazjin@tvl.su>
2024-08-19 11:50:11 +00:00
Vincent Ambo
4d5abbe232 feat(tazjin/german-string): add Drop impl for transient strings
All of these strings are currently transient (the storage class is not yet
represented anywhere), and the ones that are heap allocated need to be
deallocated when the transient string dies.

Change-Id: Iba0ca926df5db7594f304c5d5318db9d69c6f42c
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12235
Autosubmit: tazjin <tazjin@tvl.su>
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
2024-08-19 11:50:11 +00:00
Vincent Ambo
ef75a6300b feat(tazjin/german-string): init initial representation
This adds an initial implementation of the so-called "German Strings" in Rust.

https://cedardb.com/blog/german_strings/

This implementation is *far from* complete, the only thing that can be done
right now is construct a string, and even that I'm not fully happy with.

Change-Id: I2697932a0ef373be76ffd14d59677493a5783b58
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12234
Autosubmit: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
Reviewed-by: tazjin <tazjin@tvl.su>
2024-08-19 11:42:56 +00:00