ChangeLog for SDoP
------------------

Version 0.61
------------

 1. Added options to ./configure to force disabling of JPEG and PNG support.

 2. Some #ifdefs for JPEG and PNG support were missing.

 3. Arrange to skip JPEG and PNG tests in "make test" when the appropriate
    support is not compiled.


Version 0.60
------------

 1. Added support for subscript_small and superscript_small, both defaulting to
    "yes". At present, the small font is the footnote font. These parameters
    are ignored in titles.

 2. Added support for subscript_down and superscript_up, both defaulting to 33.

 3. Do not check for unsupported processing instructions unless output is
    generated, because when SDoP stops early after errors, not all of them will
    have been looked at.

 4. Added some words to the hyphenation list.

 5. If two index entries for the same item both had processing parameters in
    their content, the second gave rise to incorrect "unrecognized SDoP
    processing parameter" errors, because it was never actually processed.

 6. Index entries that differed only in font sizes, colours, or in having some
    characters subscripted or superscripted were not being treated as
    different.

 7. Allow spaces after commas in settings such as "yes, yes, no".

 8. Impose a minimum page length of 108 points (1.5").

 9. Added support for PNG files, but I have only tested a couple, so it may
    be incomplete.

10. Changed the fontmetrics files in the distribution to a set obtained with
    clear licensing conditions, and with appropriate licence files. The
    replacement files have more kerning data, which might be helpful. However,
    they lacked entries for the Euro character (which is in the fonts). I added
    Euro entries to the Helvetica, Courier, and Times fonts, copying from the
    previous AFM files, and noting in a comment that I had done so.

11. The name of the AFM file for the AvantGarde italic font was incorrect in
    the code ("Oblique" instead of "BookOblique").

12. The numeration in an ordered list item took on the font characteristics
    (size, italic, etc.) of the first character, which is clearly silly. Now
    the numeration uses the default font and size for the main text.

13. SDoP was reading all characters between & and ; when processing entities.
    Now it accepts only letters, digits, underscore, colon, hyphen, and dot.

14. Added &apos; and &quot; to the tables, as these are supposed to be
    recognized by all XML processors.

15. Added recognition of SDOP_SHARE (can be overridden by -S).

16. Added support for BPH (U+0082, "break permitted here") and NBH (U+0083, "no
    break here").

17. When SDoP's PostScript output was converted to a PDF, the index pointers
    for any prefaces were not working. I also added a PDF index item for the
    title page, when present.

18. Stop SDoP from crashing if an entity is unexpectedly its own partner,
    for example, <index/>.


Version 0.52
------------

 1. Because inserted files sometimes have bad bounding box values, and also
    for general aesthetic reasons, I've added a feature to <imageobject> that
    allows the image to be moved on the page. To do this, the "role" parameter
    has been abused, so as not to break the DocBook DTD.

 2. Added table_indent as a global processing instruction.

 3. The character U+2027 (a centred dot) was not being recognized; it
    corresponds to the PostScript character "periodcentered".

 4. The character U+0145, though correctly handled, was generating a spurious
    "unrecognized character" error.

 5. Added support for align="char" to <tgroup> and <colspec>. The defaults for
    "char" and "charoff" are "." and "50" respectively. If the given character
    is not found in a cell's text, SDoP behaves as if it were present at the
    end of the string. If "char" alignment would cause overflow or underflow,
    an error is generated and right or left alignment (respectively) is used
    instead.

 6. Added support for "align", "char", and "charoff" to <entry>.

 7. SDoP borrows some code from PCRE for checking whether a Unicode character
    is a letter. This is used when hyphenating. The way PCRE does this has been
    updated to be more efficient; the revised code has now been installed in
    SDoP. Efficiency isn't the only point: the main table is auto-generated
    from the Unicode tables for PCRE. The old way of generating is no longer
    maintained, so it's best to move to the new scheme.

 8. Added support for <?sdop extra_leading="n"?>.

 9. When a non-standardly encoded font (e.g. the PMW-Music font) was used,
    characters that are special in standard encoding (e.g. 160, the "hard
    space") were being treated specially, which should not be the case. All
    characters in such fonts are now just treated as data characters.

10. If a non-standardly encoded font did not contain a width for the '-'
    character, SDoP could crash.

11. If a non-standardly encoded font did not contain the character used for
    substituting for unknown characters, a hard internal error was given. Now,
    instead of trying to use the normal unknown character substitution, the
    first defined character in the special font is used, and a warning is
    output to say so.

12. SDoP could loop if a <literallayout> section was sufficiently long that it
    continued over two page breaks, and there was a forced page break at its
    end.

13. Added support for <?sdop index_sort_omit="...."?>.

14. Added support for <?sdop toc_printed_sections=".."?>, which limits the
    contents of the printed TOC, as opposed to the PDF bookmarks TOC. Setting
    toc_sections also sets toc_printed_sections.

15. Vertical stretching was not being applied to space specified before items;
    it was only being applied to the "after" spacing.

16. Allowed a little vertical stretching in indexes.

17. Ensure that at least one secondary index line appears in the same column as
    its primary - that is, avoid having a primary last in a column if it is
    followed by one or more secondaries.

18. When a list of secondary index entries flows over to a new column, arrange
    for a "(continued)" primary item to head the column.

19. The default column separation for indexes was not quite right.


Version 0.51
------------

 1. When a font contained a character with a negative width (the PMW-Music font
    has such characters), SDoP was generating a spurious "unknown character"
    error.

 2. An attempt to specify a single font (e.g. a section title) as bolditalic
    failed.

 3. Previously, all section titles, however deeply nested, used the same title
    font. Now the first uses the "section font" and any nested ones use the
    "subsection" font, which defaults to 11 point.

 4. When the table column overflow option was set to "overprint", spurious
    warnings could be given when an overflowing entry was the last one
    specified in a row. The code now also checks for overflowing into blank
    columns (previously overflowing into a blank column could trigger the
    warning if the justification was appropriate).


Version 0.50
------------

 1. Added support (or at least recognition) for these elements from Simplified
    DocBook:

      <abbrev>             No new processing is actually involved.
      <abstract>           No new processing is actually involved.
      <acronym>            No new processing is actually involved.
      <article>
      <articleinfo>
      <attribution>
      <audiodata>          Completely ignored
      <audioobject>        Completely ignored
      <authorblurb>        Completely ignored
      <caption>
      <citetitle>          Treated as <emphasis>
      <computeroutput>     Treated as <literal>
      <corpauthor>         Sets &author_corpauthor;
      <edition>            Sets &book_edition;
      <editor>             Sets various &editor_xxx; entities
      <email>              Italic by default
      <epigraph>           Works like <blockquote> but no title
      <example>            Parameters similar to <figure> and <table>
      <footnoteref>        But only when on the same page as the footnote
      <inlinemediaobject>  Treated as a synonym for <mediaobject>
      <issuenum>           Sets &book_issuenum;
      <jobtitle>
      <keyword>            Completely ignored
      <keywordset>         Completely ignored
      <legalnotice>        Use <?sdop insert="legalnotice"?> to insert
      <lineannotation>     Uses a small italic font.
      <link>               Completely ignored
      <note>               Works like <blockquote>, but no <attribution>
      <objectinfo>         Completely ignored
      <othercredit>        Sets various &othercredit_xxx; entities
      <pubdate>            Sets &book_pubdate;
      <publishername>      Sets &book_publishername;
      <releaseinfo>        Sets &book_releaseinfo;
      <replaceable>        Italic by default
      <revdescription>     Use <?sdop insert="revdescription"?> to insert
      <revremark>          Completely ignored
      <sectioninfo>        Completely ignored
      <sidebar>            Works like <blockquote>, but no <attribution>
      <subject>            Completely ignored
      <subjectset>         Completely ignored
      <subjectterm>        Completely ignored
      <systemitem>         Completely ignored
      <trademark>          Completely ignored
      <userinput>          Monospaced by default
      <videodata>          Completely ignored
      <videoobject>        Completely ignored
      <volumenum>          Sets &book_volumenum;

    Note that &book_revnumber; gets the value of the last <revnumber> to be
    encountered if there is more than one.

 2. Added support for <subtitle> on chapters (previously it was only acted on
    for book titles), which makes it also work for articles. Added
    font_chapter_subtitle to vary the font used.

 3. Nobody had noticed that the PostScript output contained comments like
    "%%Page 3 5" rather than the correct form "%%Page: 3 5". Fixed.

 4. Added support for <emphasis role="roman"> to make it possible to revert to
    roman type in the middle of non-roman (which might have been turned on by
    some other element, e.g. <citetitle>).

 5. Added support for <honorific>, <lineage>, and <othername> inside <author>,
    though all it does is set up special entities. They are not used in the
    default title templates.

 6. Added support for <emphasis role="smallfont"> to implement a smaller font
    that can be used for anything, but in particular, for "legalnotice" type
    stuff on title verso pages. The small fonts are also used for
    <lineannotation>s.

 7. Added support for <title> in <blockquote> and processing instructions
    font_blockquote_title and blockquote_title_justify.

 8. Added <?sdop blockquote_ruled="yes"?> to put horizontal rules above and
    below block quotes.


Version 0.06
------------

 1. If a chapter or section title contained a special entity, and head or foot
    lines were set up to contain the chapter or section title, SDoP did not
    recognize the entities in the head/foot lines, and gave an "unrecognized"
    error.

 2. If an automatically inserted hyphen followed an "fi" ligature, or any other
    character whose code point is greater than 255, the hyphen character was
    incorrectly output.

 3. Most Unicode characters greater than 255 in chapter or section titles are
    converted to '?' in the PDFMARK data that is included in the PostScript so
    that a PDF table of contents can be created. Some new replacements have
    been added: opening and closing quotes are converted to ASCII versions, and
    en and em dashes are converted to hyphens.


Version 0.05
------------

 1. Treat a paragraph consisting of just a single zero-width space character
    (&#x200b;) specially by removing the line of data. This causes the vertical
    white space to be a bit less than the case when an actually blank line is
    output. The latter can be caused, for example, by using a hard space
    (&nbsp;).


Version 0.04
------------

 1. When a title was followed by a mediaobject (image) and the pair of them
    could not fit on a page, sdop looped because it always tries to keep a
    title together with what follows. Now it outputs the title on its own; if
    the media object itself won't fit, that will be caught elsewhere.

 2. A chapter without a title that contained only a mediaobject and no other
    text was not forcing a new page.

 3. Added auth-enti-cation to the hyphenation dictionary.

 4. Add --help to hytest and buildhy.

 5. When a background color was set for a page, the PostScript path was not
    being deleted; if another drawing was on the page, it got messed up.

 6. The distribution was missing the *-t and *-p data files, and the Makefile
    wasn't installing them.

 7. The typo "while isspace(*p)" wasn't noticed because the code was tested
    on a system where isspace() is a macro that ends up in parentheses. The
    missing parens have been added.


Version 0.03
------------

 1. Added page_full_length, page_line_width, page_foot_line_width, and
    page_head_line_width.

 2. Added many processing instructions for changing details of specific fonts:
    scale_typesize_base and lots of font_xxx...

 3. Missing substrings for toc_line_chapter_strings were not being set empty.

 4. Added paper_size and margin_{bottom,left_recto,left_verso} processing
    instructions.

 5. Re-jigged the font handling to avoid binding different virtual font numbers
    with the same characteristics.

 6. Added background_rgb to set background colour.

 7. Fixed a bug with the indentation of <term>s in <variablelist>s. By default
    they would end up 2 points to the left the main text (not always
    noticeable). Invented a separate set of layout parameters for the titles of
    <variablelists>s.

 8. Added many processing instructions for changing the indentation of many
    different types of paragraph.


Version 0.02
------------

 1. Added -p and -pf to select specific output pages.

 2. Added processing instruction chapter_head_skip="no" to skip head for a
    chapter starting page.

 3. Added a whole lot more character entity definitions.

 4. Added processing instructions to change the three font familes.

 5. Added index_headings processing instruction.

 6. Added support for <figure>.

 7. Added support for <title> in <table>s.

 8. Added support for <sectn>, but always as a synonym for <section>.

 9. Added support for <programlisting> as a synonym for <screen>.

10. Added support for <![CDATA[...]]> sections.

11. Added support for <subscript> and <superscript>.

12. If an unsupported character's code was less than 384, no warning message
    was given, and the substitute character's width was incorrect.

13. Added -qc, -wc, and processing parameter warn_unsupported_characters.

14. SDoP wasn't properly skipping processing instructions for other processors
    if they extended over more than one line.


Version 0.01
------------

 1. When a line that ended with </literal> was joined onto the next line, the
    inserted space was in the monospaced font and not in the ordinary font,
    thus often looking far too wide.

 2. Tidy up error messages for files that fail to open.

 3. Do not output a table of contents if there is nothing to go in it.

 4. A hyphen at the start of a line or table cell was being taken as a possible
    hyphenation point by mistake.

 5. Added support for one or more <preface> elements.

 6. The toc_sections processing instruction is now local rather than global, so
    it can be changed while processing. In parcticular, a different setting can
    be used within a preface.

 7. Added separate header and footer templates for the preface and the TOC, and
    processing instructions to vary the left/centre/right variables. Page
    numbers in the TOC are now in italic roman numerals by default.

 8. Added support for appendices and colophons.

 9. Added support for footnotes.

10. Added support for <mediaobject> with <textobject> and <imageobject> in
    EPS or JPEG formats.

11. Some refactoring to make the handling of newlines within the markup more
    straightforward.

12. Added support for exotic fonts via

      <emphasis role="exotic-font-name/size/leading">

    If size and leading are omitted, the values from the current font are used.
    Size can be given as 0 to mean "the current size" if just additional
    leading is needed. Leading space happens above the line. Fonts with
    standard encoding are re-encoded as Unicode; for others, only characters in
    the range 0-255 are supported, in their native encoding.

** End **
