@pipelex/sdk 0.23.0 → 0.25.0

package/dist/hooks/patch-envelope.js

@@ -0,0 +1,127 @@
+/**
+ * Reader of Codex's `apply_patch` envelope, as sections rather than bare
+ * header lines.
+ *
+ * A patch is `*** Begin Patch`, then one section per file, then `*** End
+ * Patch`. A section opens with `*** Add File: <path>`, `*** Update File:
+ * <path>` (optionally followed by `*** Move to: <path>`) or `*** Delete File:
+ * <path>`, and holds the lines under it until the next header. Its `+` lines
+ * are the lines the patch adds, which Codex writes to the file verbatim.
+ *
+ * The text read may be the patch alone (the `apply_patch` tool) or a whole
+ * shell script with the patch inside it (Codex's `Bash` tool), so the reader
+ * keeps each section's offset for the caller to place it in the script.
+ */
+const SECTION_HEADER = /^\*\*\* (Add File|Update File|Delete File):\s*(.*?)\s*$/;
+const MOVE_HEADER = /^\*\*\* Move to:\s*(.*?)\s*$/;
+const ENVELOPE_BOUNDARY = /^\*\*\* (Begin|End) Patch\b/;
+const MTHDS_PATH = /.\.mthds$/;
+const SECTION_KINDS = {
+    "Add File": "add",
+    "Update File": "update",
+    "Delete File": "delete",
+};
+/**
+ * Every section of every patch envelope in `text`, in order. The recognised
+ * headers are anchored at the start of a line, as the `pipelex-plugins`
+ * wrapper's pre-filter reads them; a `Move to` counts only inside an `Update
+ * File` section, and a `Begin Patch` or `End Patch` line closes the section
+ * before it and starts a new envelope, so script lines after a patch are not
+ * read as its content, and the patches of a script stay apart.
+ */
+export function readPatchSections(text) {
+    const sections = [];
+    let current = null;
+    let envelope = 0;
+    let lineStart = 0;
+    while (lineStart <= text.length) {
+        const newline = text.indexOf("\n", lineStart);
+        const lineEnd = newline === -1 ? text.length : newline;
+        const line = text.slice(lineStart, lineEnd).replace(/\r$/, "");
+        const header = SECTION_HEADER.exec(line);
+        const boundary = ENVELOPE_BOUNDARY.exec(line);
+        if (header) {
+            current = null;
+            if (header[2]) {
+                current = {
+                    kind: SECTION_KINDS[header[1]],
+                    path: header[2],
+                    moveTo: null,
+                    offset: lineStart,
+                    envelope,
+                    addedLines: [],
+                };
+                sections.push(current);
+            }
+        }
+        else if (boundary) {
+            current = null;
+            envelope++;
+        }
+        else if (current) {
+            const move = MOVE_HEADER.exec(line);
+            if (move) {
+                if (current.kind === "update" && current.moveTo === null && move[1]) {
+                    current.moveTo = move[1];
+                }
+            }
+            else if (line.startsWith("+")) {
+                current.addedLines.push(line.slice(1));
+            }
+        }
+        if (newline === -1) {
+            break;
+        }
+        lineStart = newline + 1;
+    }
+    return sections;
+}
+/**
+ * The `.mthds` files that survive the sections. Within one envelope the
+ * sections apply in order and in one directory: an added or updated file
+ * becomes a target, a moved file stops being one and its destination becomes
+ * one, and a deleted file stops being one. The envelopes of a shell script
+ * may run in other directories, in branches, or not at all, so across them a
+ * file survives when any envelope leaves it, and a delete in one never hides
+ * another's edit. Two paths name the same file when `keyOf` gives them the
+ * same key, which lets the caller compare paths resolved against the
+ * directory each section was applied in; by default the paths are compared
+ * as written.
+ */
+export function patchTargets(sections, keyOf = (path) => path) {
+    const envelopes = new Map();
+    for (const section of sections) {
+        envelopes.set(section.envelope, [...(envelopes.get(section.envelope) ?? []), section]);
+    }
+    const files = new Map();
+    const removed = new Set();
+    for (const envelope of envelopes.values()) {
+        const surviving = new Map();
+        for (const section of envelope) {
+            const key = keyOf(section.path, section);
+            if (section.kind === "delete" || section.moveTo !== null) {
+                surviving.delete(key);
+                removed.add(key);
+            }
+            if (section.moveTo !== null) {
+                surviving.set(keyOf(section.moveTo, section), { path: section.moveTo, section });
+            }
+            else if (section.kind !== "delete") {
+                surviving.set(key, { path: section.path, section });
+            }
+        }
+        for (const [key, { path, section }] of surviving) {
+            const file = files.get(key);
+            if (file) {
+                file.sections.push(section);
+            }
+            else {
+                files.set(key, { key, path, sections: [section], removedByPatch: false });
+            }
+        }
+    }
+    return Array.from(files.values())
+        .filter((file) => MTHDS_PATH.test(file.path))
+        .map((file) => ({ ...file, removedByPatch: removed.has(file.key) }));
+}
+//# sourceMappingURL=patch-envelope.js.map

package/dist/hooks/patch-envelope.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"patch-envelope.js","sourceRoot":"","sources":["../../src/hooks/patch-envelope.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;GAaG;AAkCH,MAAM,cAAc,GAAG,yDAAyD,CAAC;AACjF,MAAM,WAAW,GAAG,8BAA8B,CAAC;AACnD,MAAM,iBAAiB,GAAG,6BAA6B,CAAC;AACxD,MAAM,UAAU,GAAG,WAAW,CAAC;AAE/B,MAAM,aAAa,GAAyC;IAC1D,UAAU,EAAE,KAAK;IACjB,aAAa,EAAE,QAAQ;IACvB,aAAa,EAAE,QAAQ;CACxB,CAAC;AAEF;;;;;;;GAOG;AACH,MAAM,UAAU,iBAAiB,CAAC,IAAY;IAC5C,MAAM,QAAQ,GAAmB,EAAE,CAAC;IACpC,IAAI,OAAO,GAAwB,IAAI,CAAC;IACxC,IAAI,QAAQ,GAAG,CAAC,CAAC;IACjB,IAAI,SAAS,GAAG,CAAC,CAAC;IAClB,OAAO,SAAS,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC;QAChC,MAAM,OAAO,GAAG,IAAI,CAAC,OAAO,CAAC,IAAI,EAAE,SAAS,CAAC,CAAC;QAC9C,MAAM,OAAO,GAAG,OAAO,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,OAAO,CAAC;QACvD,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,SAAS,EAAE,OAAO,CAAC,CAAC,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;QAC/D,MAAM,MAAM,GAAG,cAAc,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACzC,MAAM,QAAQ,GAAG,iBAAiB,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAC9C,IAAI,MAAM,EAAE,CAAC;YACX,OAAO,GAAG,IAAI,CAAC;YACf,IAAI,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC;gBACd,OAAO,GAAG;oBACR,IAAI,EAAE,aAAa,CAAC,MAAM,CAAC,CAAC,CAAE,CAAE;oBAChC,IAAI,EAAE,MAAM,CAAC,CAAC,CAAC;oBACf,MAAM,EAAE,IAAI;oBACZ,MAAM,EAAE,SAAS;oBACjB,QAAQ;oBACR,UAAU,EAAE,EAAE;iBACf,CAAC;gBACF,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;YACzB,CAAC;QACH,CAAC;aAAM,IAAI,QAAQ,EAAE,CAAC;YACpB,OAAO,GAAG,IAAI,CAAC;YACf,QAAQ,EAAE,CAAC;QACb,CAAC;aAAM,IAAI,OAAO,EAAE,CAAC;YACnB,MAAM,IAAI,GAAG,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YACpC,IAAI,IAAI,EAAE,CAAC;gBACT,IAAI,OAAO,CAAC,IAAI,KAAK,QAAQ,IAAI,OAAO,CAAC,MAAM,KAAK,IAAI,IAAI,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC;oBACpE,OAAO,CAAC,MAAM,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC;gBAC3B,CAAC;YACH,CAAC;iBAAM,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC;gBAChC,OAAO,CAAC,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;YACzC,CAAC;QACH,CAAC;QACD,IAAI,OAAO,KAAK,CAAC,CAAC,EAAE,CAAC;YACnB,MAAM;QACR,CAAC;QACD,SAAS,GAAG,OAAO,GAAG,CAAC,CAAC;IAC1B,CAAC;IACD,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED;;;;;;;;;;;GAWG;AACH,MAAM,UAAU,YAAY,CAC1B,QAAiC,EACjC,QAAyD,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI;IAEvE,MAAM,SAAS,GAAG,IAAI,GAAG,EAA0B,CAAC;IACpD,KAAK,MAAM,OAAO,IAAI,QAAQ,EAAE,CAAC;QAC/B,SAAS,CAAC,GAAG,CAAC,OAAO,CAAC,QAAQ,EAAE,CAAC,GAAG,CAAC,SAAS,CAAC,GAAG,CAAC,OAAO,CAAC,QAAQ,CAAC,IAAI,EAAE,CAAC,EAAE,OAAO,CAAC,CAAC,CAAC;IACzF,CAAC;IACD,MAAM,KAAK,GAAG,IAAI,GAAG,EAAyB,CAAC;IAC/C,MAAM,OAAO,GAAG,IAAI,GAAG,EAAU,CAAC;IAClC,KAAK,MAAM,QAAQ,IAAI,SAAS,CAAC,MAAM,EAAE,EAAE,CAAC;QAC1C,MAAM,SAAS,GAAG,IAAI,GAAG,EAAmD,CAAC;QAC7E,KAAK,MAAM,OAAO,IAAI,QAAQ,EAAE,CAAC;YAC/B,MAAM,GAAG,GAAG,KAAK,CAAC,OAAO,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC;YACzC,IAAI,OAAO,CAAC,IAAI,KAAK,QAAQ,IAAI,OAAO,CAAC,MAAM,KAAK,IAAI,EAAE,CAAC;gBACzD,SAAS,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;gBACtB,OAAO,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;YACnB,CAAC;YACD,IAAI,OAAO,CAAC,MAAM,KAAK,IAAI,EAAE,CAAC;gBAC5B,SAAS,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,MAAM,EAAE,OAAO,CAAC,EAAE,EAAE,IAAI,EAAE,OAAO,CAAC,MAAM,EAAE,OAAO,EAAE,CAAC,CAAC;YACnF,CAAC;iBAAM,IAAI,OAAO,CAAC,IAAI,KAAK,QAAQ,EAAE,CAAC;gBACrC,SAAS,CAAC,GAAG,CAAC,GAAG,EAAE,EAAE,IAAI,EAAE,OAAO,CAAC,IAAI,EAAE,OAAO,EAAE,CAAC,CAAC;YACtD,CAAC;QACH,CAAC;QACD,KAAK,MAAM,CAAC,GAAG,EAAE,EAAE,IAAI,EAAE,OAAO,EAAE,CAAC,IAAI,SAAS,EAAE,CAAC;YACjD,MAAM,IAAI,GAAG,KAAK,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;YAC5B,IAAI,IAAI,EAAE,CAAC;gBACT,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;YAC9B,CAAC;iBAAM,CAAC;gBACN,KAAK,CAAC,GAAG,CAAC,GAAG,EAAE,EAAE,GAAG,EAAE,IAAI,EAAE,QAAQ,EAAE,CAAC,OAAO,CAAC,EAAE,cAAc,EAAE,KAAK,EAAE,CAAC,CAAC;YAC5E,CAAC;QACH,CAAC;IACH,CAAC;IACD,OAAO,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,EAAE,CAAC;SAC9B,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;SAC5C,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,EAAE,GAAG,IAAI,EAAE,cAAc,EAAE,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC,CAAC;AACzE,CAAC"}

package/dist/hooks/shell-script.d.ts

@@ -0,0 +1,92 @@
+/**
+ * A reading of a POSIX shell script that tells, for an offset in it, which
+ * directory the patch command reading that offset runs in.
+ *
+ * Codex reports a patch run through the shell as a `Bash` call carrying the
+ * whole script, and a relative path in that patch is relative to wherever the
+ * script had moved when `apply_patch` ran. This module lexes the script far
+ * enough to follow `cd`: commands, quotes, heredocs, substitutions, scopes,
+ * branches and loops. It is not a shell parser. Past what it reads, it
+ * answers "unknown" rather than guessing, and a script it cannot lex at all
+ * is "unparsed".
+ *
+ * - Text is placed only where a patch command reads it: the words and
+ *   heredoc bodies of an `apply_patch` command, and the text of a command
+ *   feeding one through a pipeline or a substitution in its words. Text that
+ *   any other command holds, such as a variable, a file, `bash -c` or `eval`,
+ *   is unknown, since this reading does not follow whatever reads it later.
+ * - `cd DIR` with one literal operand (optionally after `-L`, `-P` or `--`)
+ *   moves the directory, and is assumed to succeed; any other `cd`, and
+ *   `pushd`, `popd`, `eval`, `source`, `.` or a function definition, makes
+ *   it unknown. An absolute `cd` makes an unknown directory known again.
+ * - The reading follows the path on which each command succeeds, so a
+ *   negated `cd` fails on it. What runs only on the other path (after `||`,
+ *   in `elif` and `else`, after a negated `cd` and `&&`, and on the path a
+ *   loop's condition takes to end or skip it) starts where the shell stood
+ *   when nothing had moved it, and from an unknown directory otherwise.
+ * - `( … )`, `$( … )`, `<( … )` and `>( … )` are scopes: a `cd` inside one
+ *   ends at its closing parenthesis. A list sent to the background, and every
+ *   member of a pipeline but the last, run in subshells in bash and zsh
+ *   alike. zsh runs a pipeline's last member in the current shell and bash
+ *   does not, so a directory change there makes the directory after it
+ *   unknown.
+ * - `{ … }` and the separators `;`, newline and `&&` are followed in order.
+ *   Where only one of several paths runs, after `||`, between the branches of
+ *   `if` and `case`, and across the passes of a loop, the directory is kept
+ *   when every path leaves it the same, and is unknown otherwise.
+ * - Heredoc bodies, quoted strings, comments and the expression of a `[[ … ]]`
+ *   are never read as commands. A heredoc body belongs to the command that
+ *   opened it.
+ * - The text a command reads is the script's after the shell's quoting, which
+ *   the reading also tells: a patch in double quotes or in an unquoted heredoc
+ *   has its backslash escapes removed and its `$` and backtick expansions run.
+ */
+/** Where the patch command reading an offset runs. */
+export type Placement = {
+    kind: "directory";
+    path: string;
+}
+/** A patch command reads the offset from somewhere this reading cannot follow. */
+ | {
+    kind: "unknown";
+}
+/** A command holds the offset, and no patch command reads it: a variable, a file, `bash -c`. */
+ | {
+    kind: "unread";
+}
+/** No command holds the offset: it is in a comment or between commands. */
+ | {
+    kind: "outside";
+};
+/** How the shell turns a stretch of the script into the text a command reads. */
+export type Quoting = 
+/** As written: in a quoted heredoc, or plain. */
+"verbatim"
+/** In single quotes: as written, up to the `'` that closes them. */
+ | "single-quoted"
+/** In double quotes: a backslash escapes `$`, a backtick, `"` or itself, and `$` and backticks expand. */
+ | "double-quoted"
+/** In an unquoted heredoc: a backslash escapes `$`, a backtick or itself, and `$` and backticks expand. */
+ | "heredoc";
+/** A script that could be lexed, able to place any offset in it. */
+export interface ShellReading {
+    directoryAt(offset: number): Placement;
+    quotingAt(offset: number): Quoting;
+}
+/**
+ * Read `script` as a POSIX shell script run from `sessionDir`, or return
+ * `"unparsed"` when it holds an unterminated quote, substitution or backtick,
+ * a heredoc inside backticks, a parenthesis that closes nothing, or any other
+ * syntax error this reading can see.
+ */
+export declare function readShellScript(script: string, sessionDir: string): ShellReading | "unparsed";
+/**
+ * The lines of a patch section as the patch command reads them, given the
+ * quoting its header is in, less any line an expansion makes unknown. A line
+ * in quotes may close them and open others, as `'\''` writes an apostrophe in
+ * single quotes, and is read through each. When a line ends in other quoting
+ * than it began in, or in a backslash that joins it to the next, the lines
+ * after it are read in quoting this reading does not follow, so no line of
+ * the section is known and none is given.
+ */
+export declare function linesAsRead(lines: readonly string[], quoting: Quoting): string[];