Merge pull request #8618 from sagemathinc/latex-output-fixes4

frontend/latex: add more file types to the list of dependencies

Author: haraldschilly

src/packages/frontend/cspell.json

@@ -46,6 +46,7 @@
     "respawns",
     "revealjs",
     "Rmarkdown",
+    "rtex",
     "rtypes",
     "Sagemath",
     "sagetex",

src/packages/frontend/frame-editors/latex-editor/latex-log-parser.ts

@@ -23,6 +23,8 @@ import { trimEnd } from "lodash";
 
 import { normalize as path_normalize } from "path";
 
+import { filename_extension } from "@cocalc/util/misc";
+
 // Define some constants
 const LOG_WRAP_LIMIT = 79;
 const LATEX_WARNING_REGEX = /^LaTeX Warning: (.*)$/;
@@ -33,6 +35,26 @@ const LINES_REGEX = /lines? ([0-9]+)/;
 // This is used to parse the package name from the package warnings
 const PACKAGE_REGEX = /^(?:Package|Class|Module) (\b.+\b) Warning/;
 
+// Whitelist of text file extensions that can be used with \input{} or \include{}
+const ALLOWED_DEP_EXTENSIONS = [
+  "bbx",
+  "bib",
+  "bst",
+  "cbx",
+  "cfg",
+  "cls",
+  "def",
+  "lbx",
+  "md",
+  "pgf",
+  "rnw",
+  "rtex",
+  "sty",
+  "tex",
+  "tikz",
+  "txt",
+] as const;
+
 class LogText {
   private lines: string[];
   private row: number;
@@ -224,14 +246,25 @@ export class LatexParser {
 
   addDeps(line: string): void {
     line = line.trim();
-    // ignore absolute files
+    // ignore absolute files (starting with /)
     if (line[0] === "/") return;
     if (line[line.length - 1] === "\\") {
       line = line.slice(0, line.length - 1);
     }
-    // we only want to know about tex and bib files
-    const pl = line.toLowerCase(); // could be name.TEX
-    if (!pl.endsWith(".tex") && !pl.endsWith(".bib")) return;
+    // Skip files that contain a colon (like "master.pdf :")
+    if (line.includes(":")) return;
+
+    // Get the file extension (returns empty string if no extension)
+    const ext = filename_extension(line).toLowerCase();
+
+    // If there's an extension, check if it's in the whitelist
+    if (ext) {
+      if (!ALLOWED_DEP_EXTENSIONS.includes(ext as any)) {
+        return;
+      }
+    }
+    // If no extension, include it (files without extensions are allowed)
+
     this.deps.push(line);
   }
 
@@ -306,7 +339,7 @@ export class LatexParser {
     const packageMatch = this.currentLine.match(PACKAGE_REGEX);
     if (!packageMatch) return;
     const packageName = packageMatch[1];
-    // Regex to get rid of the unnecesary (packagename) prefix in most multi-line warnings
+    // Regex to get rid of the unnecessary (packagename) prefix in most multi-line warnings
     const prefixRegex = new RegExp(`(?:\\(${packageName}\\))*[\\s]*(.*)`, "i");
     // After every warning message there's a blank line, let's use it
     while (!!(this.currentLine = this.log.nextLine())) {

src/packages/frontend/misc/latex-log-parser.test.ts

@@ -1,4 +1,5 @@
 import fs from "fs";
+
 import { LatexParser } from "@cocalc/frontend/frame-editors/latex-editor/latex-log-parser";
 
 describe("latex-log-parser", () => {
@@ -14,7 +15,7 @@ describe("latex-log-parser", () => {
     expect(parsed1.deps[1]).toEqual("subfile.tex");
   });
 
-  // This is an abbrivated log of https://github.com/sagemathinc/cocalc/issues/8089
+  // This is an abbreviated log of https://github.com/sagemathinc/cocalc/issues/8089
   test("log2", () => {
     const log2 = fs.readFileSync("./misc/latex-logs/log2.txt", "utf-8");
     const parsed2 = new LatexParser(log2, {
@@ -23,7 +24,27 @@ describe("latex-log-parser", () => {
     const err0 = parsed2.all[0];
     expect(err0.file).toEqual("ch_Euclidean.tex");
     expect(err0.message).toEqual("Marginpar on page 1 moved.");
-    expect(parsed2.deps).toEqual(["01.tex", "02.5.tex"]);
+    expect(parsed2.deps).toEqual(["./livre1.bst", "01.tex", "02.5.tex"]);
     expect(parsed2.files).toEqual(["livre1.tex", "ch_Euclidean.tex"]);
   });
+
+  // Test for non-.tex/.bib files in dependencies (e.g., .txt files used via \input{})
+  test("log3", () => {
+    const log3 = fs.readFileSync("./misc/latex-logs/log3.txt", "utf-8");
+    const parsed3 = new LatexParser(log3, {
+      ignoreDuplicates: true,
+    }).parse();
+    expect(parsed3.deps).toEqual([
+      "../subdir0/file2.md",
+      "long-running.tex",
+      "long-running.txt",
+      "subdir/file_9",
+    ]);
+    // No .tex or .bib files in parentheses, so files array is empty
+    expect(parsed3.files).toEqual([]);
+    expect(parsed3.errors).toEqual([]);
+    expect(parsed3.warnings).toEqual([]);
+    expect(parsed3.typesetting).toEqual([]);
+    expect(parsed3.all).toEqual([]);
+  });
 });

src/packages/frontend/misc/latex-logs/log3.txt

@@ -0,0 +1,27 @@
+(./long-running.txt) (./subdir/file_9)
+(../subdir0/file2.md) [1{/var/lib/texmf/fonts/map/pdftex/updmap/pdftex.map}]
+(./long-running.aux) )</usr/share/texlive/texmf-dist/fonts/type1/public/amsfont
+s/cm/cmr10.pfb>
+Output written on long-running.pdf (1 page, 15440 bytes).
+SyncTeX written on long-running.synctex.gz.
+Transcript written on long-running.log.
+Latexmk: Examining 'long-running.log'
+=== TeX engine is 'pdfTeX'
+Latexmk: applying rule 'pdflatex'...
+Latexmk: All targets (long-running.pdf) are up-to-date
+#===Dependents, and related info, for long-running.tex:
+long-running.pdf :\
+../subdir0/file2.md\
+/etc/texmf/web2c/texmf.cnf\
+/usr/share/texlive/texmf-dist/fonts/type1/public/amsfonts/cm/cmr10.pfb\
+/usr/share/texlive/texmf-dist/tex/latex/base/article.cls\
+/usr/share/texlive/texmf-dist/tex/latex/base/size10.clo\
+/usr/share/texlive/texmf-dist/tex/latex/l3backend/l3backend-pdftex.def\
+/usr/share/texlive/texmf-dist/web2c/texmf.cnf\
+/usr/share/texmf/web2c/texmf.cnf\
+/var/lib/texmf/fonts/map/pdftex/updmap/pdftex.map\
+/var/lib/texmf/web2c/pdftex/pdflatex.fmt\
+long-running.tex\
+long-running.txt\
+subdir/file_9
+#===End dependents for long-running.tex: