Issues

Two rules that undo each other pass silently from one term and fail as looping from the other, since `rewrite` starts with an empty seen set (src/Rewriter.hs:266)

objectionary/phino#2250 · opened

on Nov 1, 2026

1

`eq: [𝑒1, 5]` and `gt: [𝑒1, 3]` load fine but never hold, since `eq` and `gt` answer false for a number against an expression (src/Rule.hs:174)

objectionary/phino#2249 · opened

on Nov 1, 2026

1

Under the `C` locale a non-ASCII argument such as `--hide Φ.bar` is garbled, and the error about it crashes on stderr

objectionary/phino#2222 · opened

on Oct 30, 2026

`--inside` opens the judgment's block twice in `--protocol`, because `aimed` moves `_site` to `Φ.a🌵0` but leaves `_opened` at `Φ` (src/CLI/Helpers.hs:201)

objectionary/phino#2207 · opened

on Oct 29, 2026

1

`--acyclic=plausible` never cuts a recursion through a side of a fork, because `entrance` in `src/Morph.hs` skips a formation without λ

objectionary/phino#2206 · opened

on Oct 28, 2026

2

The fast replacer rewrites a formation the pattern does not match, when a binding meta of its ends is used again in the pattern

objectionary/phino#2180 · opened

on Oct 27, 2026

README shows `phino morph --max-steps` failing with "Dataization did not finish", but phino now prints "Morphing did not finish"

objectionary/phino#2137 · opened

on Oct 26, 2026

MorphRule, DataizeRule and ContextualizeRule silently accept unknown YAML keys, unlike rewrite `Rule`

objectionary/phino#2107 · opened

on Oct 23, 2026

1

`--output=xmir` writes a Δ text node with a bare `&`, so the XMIR is not well-formed and will not reparse

objectionary/phino#2102 · opened

on Oct 23, 2026

1

`join` renames a duplicated attribute but leaves `ξ.ρ.a` in a nested formation pointing at the old attribute

objectionary/phino#2097 · opened

on Oct 23, 2026

1

`matches` and `sed` kill `phino` with a segmentation fault (exit code 139) when `(a|b)*` is applied to a string of about 9000 characters

objectionary/phino#2096 · opened

on Oct 23, 2026

2

`sed` with the `g` flag takes quadratic time in the number of matches, so a replacement over 160000 characters takes a minute

objectionary/phino#2095 · opened

on Oct 23, 2026

1

`--canonize` with `--sequence` numbers the functions again in every step, so `Fn2` of one step is called `Fn1` in the next

objectionary/phino#2091 · opened

on Oct 23, 2026

1

`--label` accepts a period, but `--label=a.b` stops pdflatex, because the `eolang` package rewrites the period inside `\label{a.b}`

objectionary/phino#2090 · opened

on Oct 23, 2026

1

`--output=latex` writes an attribute name such as `a_D` or `x-Q` as it is, and the `eolang` package turns the capital letter after `_` or `-` into a symbol that breaks pdflatex

objectionary/phino#2089 · opened

on Oct 23, 2026

1

`explain` writes the `where` functions `number` and `string` as the TeX primitives `\number` and `\string`, and `sum` and `tau` as the symbols `\sum` and `\tau`

objectionary/phino#2087 · opened

on Oct 23, 2026

1

`--sequence` with `--output=latex` writes the rule name into `\nameref{r:...}`, and a name with `_`, `%`, `.`, `,`, `?`, `(` or a space stops pdflatex with `Missing \endcsname inserted`

objectionary/phino#2086 · opened

on Oct 23, 2026

1

`--output=latex` writes `φ` and `ρ` inside an attribute or `λ` name as `\phiTerminal{...}`, which pdflatex rejects in the typewriter text of the name

objectionary/phino#2085 · opened

on Oct 23, 2026

1

A `where` step with the wrong number or kind of `args` loads fine, and the rule fails only when it matches, for example `contextualize` with one argument

objectionary/phino#2083 · opened

on Oct 23, 2026

A `where` step whose function cannot return the kind of its `meta` loads fine, and the rule fails with "inappropriate term" only when it matches

objectionary/phino#2082 · opened

on Oct 23, 2026

A rule stops with a match left when an enclosing formation matches too and its result equals itself, because `swallowed` drops the nested match

objectionary/phino#2081 · opened

on Oct 23, 2026

The salty printer takes quadratic time in the number of arguments of one call, so 16000 arguments take 27 seconds, while `--sweet` takes 0.1

objectionary/phino#2080 · opened

on Oct 23, 2026

`--output=xmir` writes a meta attribute `!t` as `as="𝜏"` or `base="Φ.f.𝜏"`, while it refuses every other meta

objectionary/phino#2078 · opened

on Oct 23, 2026

Two `λ` atoms inside application arguments of different bindings get the same derived function name, and the XMIR writer gives both the `@atom` of the last one

objectionary/phino#2076 · opened

on Oct 23, 2026

The printer moves a `Δ` or `λ` that stands between two voids behind them, so `x ↦ ⟦ a ↦ ∅, Δ ⤍ 00-, b ↦ ∅ ⟧` is printed with the order changed

objectionary/phino#2074 · opened

on Oct 23, 2026

Navigation Menu

Repository navigation

Issues

Two rules that undo each other pass silently from one term and fail as looping from the other, since `rewrite` starts with an empty seen set (src/Rewriter.hs:266)

`eq: [𝑒1, 5]` and `gt: [𝑒1, 3]` load fine but never hold, since `eq` and `gt` answer false for a number against an expression (src/Rule.hs:174)

Under the `C` locale a non-ASCII argument such as `--hide Φ.bar` is garbled, and the error about it crashes on stderr

`--inside` opens the judgment's block twice in `--protocol`, because `aimed` moves `_site` to `Φ.a🌵0` but leaves `_opened` at `Φ` (src/CLI/Helpers.hs:201)

`--acyclic=plausible` never cuts a recursion through a side of a fork, because `entrance` in `src/Morph.hs` skips a formation without λ

The fast replacer rewrites a formation the pattern does not match, when a binding meta of its ends is used again in the pattern

README shows `phino morph --max-steps` failing with "Dataization did not finish", but phino now prints "Morphing did not finish"

MorphRule, DataizeRule and ContextualizeRule silently accept unknown YAML keys, unlike rewrite `Rule`

`--output=xmir` writes a Δ text node with a bare `&`, so the XMIR is not well-formed and will not reparse

`join` renames a duplicated attribute but leaves `ξ.ρ.a` in a nested formation pointing at the old attribute

`matches` and `sed` kill `phino` with a segmentation fault (exit code 139) when `(a|b)*` is applied to a string of about 9000 characters

`sed` with the `g` flag takes quadratic time in the number of matches, so a replacement over 160000 characters takes a minute

`--canonize` with `--sequence` numbers the functions again in every step, so `Fn2` of one step is called `Fn1` in the next

`--label` accepts a period, but `--label=a.b` stops pdflatex, because the `eolang` package rewrites the period inside `\label{a.b}`

`--output=latex` writes an attribute name such as `a_D` or `x-Q` as it is, and the `eolang` package turns the capital letter after `_` or `-` into a symbol that breaks pdflatex

`explain` writes the `where` functions `number` and `string` as the TeX primitives `\number` and `\string`, and `sum` and `tau` as the symbols `\sum` and `\tau`

`--sequence` with `--output=latex` writes the rule name into `\nameref{r:...}`, and a name with `_`, `%`, `.`, `,`, `?`, `(` or a space stops pdflatex with `Missing \endcsname inserted`

`--output=latex` writes `φ` and `ρ` inside an attribute or `λ` name as `\phiTerminal{...}`, which pdflatex rejects in the typewriter text of the name

A `where` step with the wrong number or kind of `args` loads fine, and the rule fails only when it matches, for example `contextualize` with one argument

A `where` step whose function cannot return the kind of its `meta` loads fine, and the rule fails with "inappropriate term" only when it matches

A rule stops with a match left when an enclosing formation matches too and its result equals itself, because `swallowed` drops the nested match

The salty printer takes quadratic time in the number of arguments of one call, so 16000 arguments take 27 seconds, while `--sweet` takes 0.1

`--output=xmir` writes a meta attribute `!t` as `as="𝜏"` or `base="Φ.f.𝜏"`, while it refuses every other meta

Two `λ` atoms inside application arguments of different bindings get the same derived function name, and the XMIR writer gives both the `@atom` of the last one

The printer moves a `Δ` or `λ` that stands between two voids behind them, so `x ↦ ⟦ a ↦ ∅, Δ ⤍ 00-, b ↦ ∅ ⟧` is printed with the order changed