• ANNOUNCE: tclpdf 1.7.2 released - provenance and licence audit, notices in place

    From =?UTF-8?Q?Alexander_Sch=C3=B6pe?=@ete-sep@mxbo.de to comp.lang.tcl on Sun Nov 1 11:50:38 2026
    From Newsgroup: comp.lang.tcl

    ANNOUNCE: tclpdf 1.7.2 released - provenance and licence audit, notices
    in place ===============================================================================

    tclpdf is a pure Tcl extension for creating PDF: graphics, embedded and
    colour fonts, JPEG, PNG, TIFF and SVG, rich text in paragraphs and
    tables (run list, tags or Markdown), layers, form fields, tagged
    documents up to PDF/UA-2, encryption, signatures, timestamps and ZUGFeRD
    / Factur-X invoices in PDF/A-3B. It can import pages of existing PDFs
    and update files incrementally. By Alexander Schoepe, for Tcl 8.6.11+
    and Tcl 9; tdom, http and tls are optional (XMP metadata, timestamps),
    Tk is not used.

    Built against ISO 32000-1/-2, PDF/A, PDF/UA, EN 16931 with ZUGFeRD /
    Factur-X, ETSI EN 319 142-1, RFC 5652 and RFC 3161. veraPDF,
    Mustangproject, qpdf and pdfsig check every document on every run; 4672
    tests pass under Tcl 8.6.18 and 9.0.4.

    1.7.1 and 1.7.2 add no feature. They are the outcome of an audit of the
    whole tree for origin and licences, made because a package written with
    AI assistance has to show what it is made of. The result, in one
    sentence: no code from another project and no copyleft code in the
    package; the matches that exist are constants and names fixed by the standards, and the one dictionary value once taken from a pdf4tcl fork
    had been removed as non-conforming in August. What the audit did find
    were missing notices for the third-party data the package ships and the
    build files it derives from; 1.7.1 and 1.7.2 put them in place. How the package was made, and what the audit checked and found, are two tech notes:

    Provenance of tclpdf: https://fossil.sowaswie.de/tclpdf/technote/cb5371ca27 What a licence audit of tclpdf found: https://fossil.sowaswie.de/tclpdf/technote/d5570cffc2

    Wiki (manual, tables, markup, accessibility, test tools): https://fossil.sowaswie.de/tclpdf/wcontent
    Examples, every script with its finished PDF: https://fossil.sowaswie.de/tclpdf/wiki?name=examples
    Tech notes: https://fossil.sowaswie.de/tclpdf/timeline?y=e
    Download / Repository: https://fossil.sowaswie.de/tclpdf

    Author: Alexander Schoepe <alx.tcl@sowaswie.de>


    The audit, in short:

    * ScanCode Toolkit over the whole tree, SCANOSS snippet matching of
    every own file against the public code base, a manual review of the risk
    areas (font parsing and subsetting, image decoding, encryption, signing,
    XMP, generated tables), and a line-level comparison against pdf4tcl,
    tcllib, jsPDF and libhyphen. No foreign code in the modules, tests,
    examples or manual; the four snippet matches lie in the build
    scaffolding and are TEA template text or the author's own code from his
    other projects.
    * Every module that changed since a release carries a new version, and
    the module versions are held against every release tag by the test suite (since 1.6.1).

    Changed since 1.7:

    * 1.7.1: afmData, bidiData and joiningData carry the terms of their
    sources in the header (Adobe's core font metrics, the Unicode Character Database, text in tools/ucd/LICENSE); the Tcl licence travels with the
    TEA build files as tclconfig/license.terms; tools/mkversion.tcl names
    Fossil's mkversion.c it was ported from, with its BSD-2-Clause notice;
    the URW NOTICE states the measured fsType per face; a test that
    extracted a timestamp token lost a byte one time in sixteen and is fixed.
    * 1.7.2: the pandoc template behind the HTML manual carries pandoc's
    copyright and its BSD-3-Clause text; OCR-LICENSE.txt quotes the terms
    for the OCR A base (free use, no fee for distribution); Makefile.in
    keeps the TEA copyright lines; the asset README carries the fonts' own copyright years and rows for every face.

    Notes:

    * Installing: the release archives tclpdf1.7.2.zip and
    tclpdf1.7.2.tar.gz hold one directory tclpdf1.7.2 with the modules, pkgIndex.tcl and the ICC profiles - nothing to build; extract it into a directory on your auto_path and "package require tclpdf" finds it. The
    source archive builds the TEA way in the source directory: ./configure,
    make test, make install.
    * No font is part of the package. The faces under examples/assets/fonts
    travel in the source archive alone; "make install" installs none. What
    is installed besides the MIT-licensed modules carries its own terms in
    the files: the ICC profiles, the Adobe-derived widths, the Unicode tables.
    * Not coming, each for a measured reason: a heading inside a table cell, nested lists, runs that change the size or the colour, PDF/X, RC4 and
    the older encryption revisions, EPS, GIF and BMP.
    --- Synchronet 3.22a-Linux NewsLink 1.2