diff options
author | sterni <sternenseemann@systemli.org> | 2022-09-21T11·47+0200 |
---|---|---|
committer | sterni <sternenseemann@systemli.org> | 2022-09-21T14·23+0000 |
commit | 083fc1dbe59e28bb25cc1dc7405ee45d9d9244b7 (patch) | |
tree | e968148f948aa1bbe0e9de2ed17c6b7bcee35b3f /tvix/eval/src | |
parent | 9a8a6a33f9265a3844e91f2c1aab0b28ac46decf (diff) |
fix(tvix/eval): compare versions with an extra empty component r/4950
This is necessary because builtins.compareVersions compares versions in a subtly not-quite-but-still-lexicographical way: `pre` for example can have an effect if it is post-fixed: `2.3 < 2.3pre`. This is a violation of the rule that in a lexicographical ordering, the longer string is considered greater if they are otherwise equal. builtins.compareVersion is comparing lexicographically though, if you do the following transformation beforehand: 2.3 --split--> [ "2" "3" ] --append--> [ "2" "3" "" ] 2.3pre --split--> [ "2" "3" "pre" ] --append--> [ "2" "3" "pre" "" ] Comparing the transformed version is then done lexicographically: 2.3 < 2.3.0pre since [ "2" "3" "" ] < [ "2" "3" "0" "pre" ] Here, the `pre` rule never comes into effect because no comparison on it happens, instead we use the longer string rule of a lexicographical comparison. In the C++ codebase, the reason for this behavior is that the iterator-esque construct they use always yields the empty string before it exposes it has been fully consumed. This is probably intentional to support the postfixed `pre` which is, for example, used by NixOS versions (e.g. unstable post 22.05 is 22.11-pre). We replicate this behavior using the `Chain` iterator in `VersionPartsIter::new_for_cmp`. Change-Id: I021c69aa27b0b7deb949dffe50ed18b6de3a7b1f Reviewed-on: https://cl.tvl.fyi/c/depot/+/6720 Tested-by: BuildkiteCI Reviewed-by: tazjin <tazjin@tvl.su>
Diffstat (limited to 'tvix/eval/src')
-rw-r--r-- | tvix/eval/src/builtins/mod.rs | 4 | ||||
-rw-r--r-- | tvix/eval/src/builtins/versions.rs | 15 |
2 files changed, 17 insertions, 2 deletions
diff --git a/tvix/eval/src/builtins/mod.rs b/tvix/eval/src/builtins/mod.rs index 0b1fe540fed9..55135c4d8d45 100644 --- a/tvix/eval/src/builtins/mod.rs +++ b/tvix/eval/src/builtins/mod.rs @@ -103,9 +103,9 @@ fn pure_builtins() -> Vec<Builtin> { }), Builtin::new("compareVersions", &[true, true], |args, _| { let s1 = args[0].to_str()?; - let s1 = VersionPartsIter::new(s1.as_str()); + let s1 = VersionPartsIter::new_for_cmp(s1.as_str()); let s2 = args[1].to_str()?; - let s2 = VersionPartsIter::new(s2.as_str()); + let s2 = VersionPartsIter::new_for_cmp(s2.as_str()); match s1.cmp(s2) { std::cmp::Ordering::Less => Ok(Value::Integer(-1)), diff --git a/tvix/eval/src/builtins/versions.rs b/tvix/eval/src/builtins/versions.rs index ee75eeaeed3b..e149cc4dba37 100644 --- a/tvix/eval/src/builtins/versions.rs +++ b/tvix/eval/src/builtins/versions.rs @@ -1,4 +1,5 @@ use std::cmp::Ordering; +use std::iter::{once, Chain, Once}; use std::ops::RangeInclusive; /// Version strings can be broken up into Parts. @@ -69,6 +70,20 @@ impl<'a> VersionPartsIter<'a> { version, } } + + /// Create an iterator that yields all version parts followed by an additional + /// `VersionPart::Word("")` part (i.e. you can think of this as + /// `builtins.splitVersion version ++ [ "" ]`). This is necessary, because + /// Nix's `compareVersions` is not entirely lexicographical: If we have two + /// equal versions, but one is longer, the longer one is only considered + /// greater if the first additional part of the longer version is not `pre`, + /// e.g. `2.3 > 2.3pre`. It is otherwise lexicographical, so peculiar behavior + /// like `2.3 < 2.3.0pre` ensues. Luckily for us, this means that we can + /// lexicographically compare two version strings, _if_ we append an extra + /// component to both versions. + pub fn new_for_cmp(version: &'a str) -> Chain<Self, Once<VersionPart>> { + Self::new(version).chain(once(VersionPart::Word(""))) + } } impl<'a> Iterator for VersionPartsIter<'a> { |