docbook2mdoc: Import as wip/docbook2mdoc.
Update to 1.0.0 release.
ChangeLog:
$Id: ChangeLog,v 1.1 2019/04/17 18:50:38 schwarze Exp $
2019-04-17 18:45 tag VERSION_1_0_0
2019-04-17 18:45 schwarze
Deleted:
index.xml (1.14)
Added:
NEWS (1.1)
Changed:
Makefile (1.25), "Exp", lines: +10 -11
release 1.0.0
2019-04-16 21:21 schwarze
Changed:
docbook2mdoc.c (1.131), "Exp", lines: +2 -1
node.c (1.22), "Exp", lines: +2 -1
node.h (1.31), "Exp", lines: +2 -1
statistics.c (1.35), "Exp", lines: +3 -1
translate <productname> to .Sy
2019-04-16 21:19 schwarze
Changed:
parse.c (1.46), "Exp", lines: +3 -1
treat <sect3> and <sect4> like <section>
2019-04-16 14:58 schwarze
Changed:
node.c (1.21), "Exp", lines: +5 -5
node.h (1.30), "Exp", lines: +3 -2
parse.c (1.45), "Exp", lines: +15 -2
preserve leading whitespace in no-fill mode
2019-04-16 14:26 schwarze
Changed:
docbook2mdoc.c (1.130), "Exp", lines: +63 -42
macro.c (1.14), "Exp", lines: +20 -1
macro.h (1.5), "Exp", lines: +8 -1
Reorganize printing of .Pp macro lines, much improved formatting:
Refrain from manual tree inspection. Instead, let <para> and other
nodes keep paragraph state and automatically emit .Pp when needed
before printing text or macros.
2019-04-15 19:15 schwarze
Changed:
docbook2mdoc.c (1.129), "Exp", lines: +16 -1
node.c (1.20), "Exp", lines: +2 -1
node.h (1.29), "Exp", lines: +2 -1
Minimal translation of <xref> to .Sx. This might need refinement
in the future, but that won't be easy because the element is
horribly underspecified and the usage is practice is very
inconsistent.
2019-04-15 00:34 schwarze
Changed:
docbook2mdoc.c (1.128), "Exp", lines: +36 -1
node.c (1.19), "Exp", lines: +5 -1
node.h (1.28), "Exp", lines: +5 -1
statistics.c (1.34), "Exp", lines: +4 -1
The <olink> element is woefully underspecified, so the various
attributes may be used quite differently in different documents,
but let's try a basic translation anyway.
2019-04-14 23:59 schwarze
Changed:
macro.c (1.13), "Exp", lines: +6 -1
bugfix: closing quote was missing in macro_addarg()
2019-04-14 23:07 schwarze
Changed:
docbook2mdoc.c (1.127), "Exp", lines: +12 -3
statistics.c (1.33), "Exp", lines: +15 -1
ignore <quote> around <filename>
2019-04-14 22:37 schwarze
Changed:
docbook2mdoc.c (1.126), "Exp", lines: +19 -1
node.c (1.18), "Exp", lines: +4 -1
node.h (1.27), "Exp", lines: +4 -1
parse.c (1.44), "Exp", lines: +5 -1
statistics.c (1.32), "Exp", lines: +6 -1
Rudimentary implementation of <imagedata> by printing the filename.
Ignore the wrappers <caption>, <figure>, <imageobject>,
<mediaobject>.
2019-04-14 21:11 schwarze
Changed:
docbook2mdoc.c (1.125), "Exp", lines: +8 -1
node.c (1.17), "Exp", lines: +2 -2
statistics.c (1.31), "Exp", lines: +2 -1
The is no perfect way to render <footnote> in mdoc(7). For now,
use .Bo/.Bc as a full block.
2019-04-14 20:13 schwarze
Changed:
docbook2mdoc.c (1.124), "Exp", lines: +14 -1
node.c (1.16), "Exp", lines: +3 -1
node.h (1.26), "Exp", lines: +3 -1
statistics.c (1.30), "Exp", lines: +9 -1
render <superscript> and <subscript> with \(ha and _
2019-04-14 19:29 schwarze
Changed:
docbook2mdoc.c (1.123), "Exp", lines: +7 -3
statistics.c (1.29), "Exp", lines: +15 -1
do not emit .Em or .Fl right before another macro
2019-04-14 18:28 schwarze
Changed:
docbook2mdoc.c (1.122), "Exp", lines: +8 -4
statistics.c (1.28), "Exp", lines: +2 -1
avoid overquoting when <literal> is nested inside <quote>
2019-04-14 18:07 schwarze
Changed:
docbook2mdoc.c (1.121), "Exp", lines: +23 -14
node.c (1.15), "Exp", lines: +4 -8
node.h (1.25), "Exp", lines: +2 -6
parse.c (1.43), "Exp", lines: +6 -1
statistics.c (1.27), "Exp", lines: +39 -1
Minor improvements to the handling of some meta-information: *
Translate <simplesect> and <legalnotice> to .Ss or lower. * Do not
uppercase high-level <note> and similar elements. * Handle
<subtitle> like stand-alone <title>. * Handle <editor> by the
<author> handler function. * Make <copyright> print "Copyright
(C)". * Ignore some transparent nodes.
2019-04-14 16:26 schwarze
Changed:
statistics.c (1.26), "Exp", lines: +11 -3
Parse internal subset declarations in <!DOCTYPE> tags as XML; in
particular, they can contain XML comments. Similar to parts of
parse.c rev. 1.23, but simpler.
2019-04-14 14:59 schwarze
Changed:
statistics.c (1.25), "Exp", lines: +6 -2
Make <sbr> self-closing even without a trailing slash. This
logically merges parse.c rev. 1.23, even though the physical code
differs.
2019-04-14 14:00 schwarze
Changed:
docbook2mdoc.c (1.120), "Exp", lines: +1 -4
node.c (1.14), "Exp", lines: +1 -2
node.h (1.24), "Exp", lines: +1 -2
parse.c (1.42), "Exp", lines: +6 -1
statistics.c (1.24), "Exp", lines: +29 -1
Minor improvements for some inline elements: * New aliases <keycap>
and <property>. * Make <application> an alias for <command>. *
Ignore <keycode> and <returnvalue>.
2019-04-14 12:59 schwarze
Changed:
macro.c (1.12), "Exp", lines: +7 -4
Relax child macro escaping: * No escaping needed in ARG_UPPER mode.
* There are only four callable three-letter macros.
2019-04-14 12:38 schwarze
Changed:
README (1.7), "Exp", lines: +4 -3
docbook2mdoc.c (1.119), "Exp", lines: +146 -12
statistics.c (1.23), "Exp", lines: +12 -1
Move content from <bookinfo>, <articleinfo>, <refentryinfo>, and
<refmeta> that is not used in the prologue to the AUTHORS section.
2019-04-13 15:54 schwarze
Changed:
docbook2mdoc.c (1.118), "Exp", lines: +20 -1
node.c (1.13), "Exp", lines: +6 -2
node.h (1.23), "Exp", lines: +5 -1
statistics.c (1.22), "Exp", lines: +2 -1
rudimentary implementation of <systemitem>
2019-04-13 15:13 schwarze
Changed:
docbook2mdoc.c (1.117), "Exp", lines: +44 -42
Improve formatting of <arg> and <group>: * Respect the repeat
attribute even if there is a child macro. * Mark up all children
of <group>, not just the first. * Implement the repeat attribute
for <group>. * Do the proper was_impl dance for <group>.
2019-04-13 13:06 schwarze
Changed:
docbook2mdoc.c (1.116), "Exp", lines: +64 -54
node.c (1.12), "Exp", lines: +17 -1
node.h (1.22), "Exp", lines: +3 -1
parse.c (1.41), "Exp", lines: +2 -1
statistics.c (1.21), "Exp", lines: +7 -1
Improve the proplogue: * Use <pubdate> or <date> for .Dd. * Clean
up the way how <title> from <bookinfo> is used for .Nd. * Treat
<articleinfo> just like <bookinfo>.
2019-04-12 21:37 schwarze
Changed:
docbook2mdoc.c (1.115), "Exp", lines: +38 -11
parse.c (1.40), "Exp", lines: +2 -1
statistics.c (1.20), "Exp", lines: +11 -1
Author node polishing: * In the AUTHORS section, set -nosplit mode.
* In <author>, do not recurse for <email>: it is pointless because
it cannot reasonably contain child elements, but it can trigger
assertions. * For standalone <email>, do the standard was_impl
dance. * Treat <othercredit> just like <author>.
2019-04-12 19:14 schwarze
Added:
tree.c (1.1)
Changed:
Makefile (1.24), "Exp", lines: +3 -2
docbook2mdoc.1 (1.12), "Exp", lines: +32 -3
docbook2mdoc.c (1.114), "Exp", lines: +2 -2
format.h (1.2), "Exp", lines: +3 -2
main.c (1.6), "Exp", lines: +30 -5
node.c (1.11), "Exp", lines: +15 -3
node.h (1.21), "Exp", lines: +3 -1
Implement lint and tree dump output modes. Thanks to the
previously committed node property infrastructure in node.c, this
needs only 110 lines of code (including the license and the
documentation).
2019-04-12 16:40 schwarze
Changed:
README (1.6), "Exp", lines: +7 -6
docbook2mdoc.c (1.113), "Exp", lines: +13 -7
node.c (1.10), "Exp", lines: +150 -1
node.h (1.20), "Exp", lines: +24 -9
parse.c (1.39), "Exp", lines: +68 -178
Infrastructure to access node properties that can be used by both
the parser and the formatter. Use it to decide whether a text node
needs .Pf before a macro, whether the first word of a text node
needs to be split off after a macro, and whether a macro node can
suppress the SPACE flag.
2019-04-12 12:07 schwarze
Changed:
docbook2mdoc.c (1.112), "Exp", lines: +25 -12
Finally get rid of the "sv" variable in pnode_print(). Switch
<literal> and <quote> between explicit and implicit enclosures
depending on whether we are already inside an implicit context.
2019-04-12 11:37 schwarze
Changed:
docbook2mdoc.c (1.111), "Exp", lines: +6 -14
node.h (1.19), "Exp", lines: +1 -2
parse.c (1.38), "Exp", lines: +1 -4
There is no need to ever use the eqn(7) inline syntax ($$): Its
purpose it to inline eqn(7) source code in the *input* file. On
the *output* side, equations are inlined by default. To display
them, the .EQ block needs to be wrapped in .Bd. This patch gets
rid of another user of "sv".
2019-04-12 11:19 schwarze
Changed:
docbook2mdoc.c (1.110), "Exp", lines: +24 -19
Clean up the <refnamediv> mess to get rid of the worst "sv" user:
Explicit loop over <refname> elements in <refnamediv>. Treat stray
<refname> elements just like <command>.
2019-04-12 10:34 schwarze
Changed:
docbook2mdoc.c (1.109), "Exp", lines: +17 -7
Improve <link linkend="..."> formatting. Take markup of the
content into account. Do not break a partial implicit macro line,
if one is open. Do not let the parentheses extend too far.
2019-04-12 09:39 schwarze
Changed:
docbook2mdoc.c (1.108), "Exp", lines: +21 -10
Fix an assertion failure when content inside a term causes an
output line break. The solution isn't perfect: content from the
<term> still leaks into the .It body, but fixing *that* is a much
more complex task than merely fixing the assertion.
2019-04-12 09:01 schwarze
Changed:
macro.c (1.11), "Exp", lines: +3 -3
another case of incomplete escaping: element space dot
2019-04-12 08:48 schwarze
Changed:
docbook2mdoc.c (1.107), "Exp", lines: +17 -36
macro.c (1.10), "Exp", lines: +10 -3
Move escaping of control characters and backslashes on text lines
to print_text() such that it works for all text lines.
In pnode_printtext(), use macro_addarg() or print_text() to
get the required escaping. On the other hand, there is no need to
handle linefeed characters because these can no longer occur in
text nodes.
Stephen Gregoratto <dev at sgregoratto dot me> reported
that escaping was incomplete in some cases.
2019-04-12 07:53 schwarze
Changed:
parse.c (1.37), "Exp", lines: +34 -8
If a text node precedes a non-text node without intervening
whitespace, split out the last word for use with .Pf.
2019-04-12 07:05 schwarze
Changed:
docbook2mdoc.c (1.106), "Exp", lines: +4 -4
node.h (1.18), "Exp", lines: +2 -2
parse.c (1.36), "Exp", lines: +4 -2
statistics.c (1.19), "Exp", lines: +3 -1
handle <computeroutput> and <markup>; patch from Stephen Gregoratto
<dev at sgregoratto dot me> who found that both are frequently used
in the doclifter manpage
2019-04-12 06:46 schwarze
Changed:
parse.c (1.35), "Exp", lines: +65 -21
If a text node follows a non-text node without intervening
whitespace, keep the text in it as short as possible.
2019-04-12 04:39 schwarze
Changed:
node.c (1.9), "Exp", lines: +15 -1
node.h (1.17), "Exp", lines: +2 -1
parse.c (1.34), "Exp", lines: +4 -19
New function pnode_alloc() to reduce code duplication. No
functional change.
2019-04-12 04:17 schwarze
Changed:
docbook2mdoc.c (1.105), "Exp", lines: +8 -2
parse.c (1.33), "Exp", lines: +4 -2
Do not include literal linefeed characters in text nodes; instead,
let the parser split them into two text nodes.
For now, let the formatter start text nodes preceded by
whitespace on new output lines, which is a crude way of preserving
input line breaks and making sure output text lines are limited to
reasonable length.
2019-04-12 03:38 schwarze
Changed:
docbook2mdoc.c (1.104), "Exp", lines: +4 -2
Outside <funcprototype>, format <paramdef> and <parameter> in-line
rather than stand-alone. While many potential assertion failures
remain, this patch fixes the last remaining assertion that actually
occurs in Xenocara.
2019-04-11 04:23 schwarze
Changed:
README (1.5), "Exp", lines: +2 -4
docbook2mdoc.c (1.103), "Exp", lines: +80 -47
macro.c (1.9), "Exp", lines: +18 -76
macro.h (1.4), "Exp", lines: +6 -3
node.c (1.8), "Exp", lines: +2 -2
node.h (1.16), "Exp", lines: +1 -3
parse.c (1.32), "Exp", lines: +21 -19
Introduce FMT_* formatter flags to control what can be added to the
current macro line. Move trailing punctuation handling to the text
formatter, allowing elimination of the function macro_closepunct()
and of the "real" and "bsz" members from struct node. Substantial
functional improvements, yet minus 25 lines of code.
2019-04-10 14:37 schwarze
Changed:
docbook2mdoc.c (1.102), "Exp", lines: +9 -9
Last instance of conistent "struct pnode *nc, *nn" variable naming.
While the previous commit caused no binary change, this one does,
but only because one variable serving two different purposes is
split into two. Still no functional change.
2019-04-10 14:34 schwarze
Changed:
parse.c (1.31), "Exp", lines: +24 -24
Last instance of consistent "struct parse *p" function argument
naming. While the previous commit caused no binary change, this
one does, but only because the argument is used in an assert(3).
Still no functional change.
2019-04-10 14:22 schwarze
Changed:
docbook2mdoc.c (1.101), "Exp", lines: +282 -282
macro.c (1.8), "Exp", lines: +8 -8
node.c (1.7), "Exp", lines: +41 -41
parse.c (1.30), "Exp", lines: +103 -103
Consitently use for function arguments: struct parse *p, struct
format *f, struct pnode *n, struct pattr *a
Consistently use struct pnode *nc, *nn, *np for child,
next, and previous nodes, respectively.
I admit this makes commit history a bit harder to inspect,
but i think seeing at once what variables mean is worth it.
2019-04-09 16:57 schwarze
Changed:
docbook2mdoc.1 (1.11), "Exp", lines: +48 -21
update content in many respects
2019-04-09 15:23 schwarze
Changed:
docbook2mdoc.1 (1.10), "Exp", lines: +55 -4
main.c (1.5), "Exp", lines: +6 -6
node.h (1.15), "Exp", lines: +5 -4
parse.c (1.29), "Exp", lines: +30 -37
switch to mandoc(1)-style EXIT STATUS and DIAGNOSTICS
2019-04-09 13:35 schwarze
Changed:
parse.c (1.28), "Exp", lines: +4 -2
Fix a logic bug causing tree corruption: When parsing an internal
subset declaration, text does not extend beyond closing square
brackets.
2019-04-09 12:23 schwarze
Changed:
parse.c (1.27), "Exp", lines: +2 -1
Fix a logic bug and a memory leak: After parsing an attribute
value, clear the attribute flag, such that, if another attribute
value follows, it gets ignored rather than clobbering the first
value.
2019-04-09 01:39 schwarze
Changed:
node.c (1.6), "Exp", lines: +2 -1
node.h (1.14), "Exp", lines: +3 -1
parse.c (1.26), "Exp", lines: +16 -2
implement file inclusion via <xi:include>
2019-04-08 23:40 schwarze
Changed:
parse.c (1.25), "Exp", lines: +10 -3
implement file inclusion via the ENTITY SYSTEM mechanism
2019-04-08 22:47 schwarze
Changed:
main.c (1.4), "Exp", lines: +23 -30
parse.c (1.24), "Exp", lines: +82 -27
Make the function parse_file() fit for recursion: * Save and
restore reporting data around the read loop. * Open the file
inside the function, not before calling it. * On the top level,
change directory on a best-effort basis. * Finalize the parse tree
only on the top level. No new functionality yet.
2019-04-08 14:37 schwarze
Changed:
node.c (1.5), "Exp", lines: +5 -1
node.h (1.13), "Exp", lines: +7 -1
parse.c (1.23), "Exp", lines: +133 -32
Handle DOCTYPE declarations containing ENTITY definitions. Also
make <sbr> self-closing even without a trailing slash.
2019-04-07 19:33 schwarze
Changed:
docbook2mdoc.c (1.100), "Exp", lines: +14 -5
node.h (1.12), "Exp", lines: +2 -2
parse.c (1.22), "Exp", lines: +5 -2
statistics.c (1.18), "Exp", lines: +6 -1
handle <appendix>, <article>, <book>, and <legalnotice> similar to
<section>
2019-04-07 18:51 schwarze
Changed:
docbook2mdoc.c (1.99), "Exp", lines: +6 -3
parse.c (1.21), "Exp", lines: +6 -1
statistics.c (1.17), "Exp", lines: +10 -1
handle glossaries just like variablelists
2019-04-07 17:55 schwarze
Changed:
docbook2mdoc.c (1.98), "Exp", lines: +5 -1
parse.c (1.20), "Exp", lines: +2 -1
statistics.c (1.16), "Exp", lines: +2 -1
render <blockquote> as .Bd -ragged -offset indent
2019-04-07 17:42 schwarze
Changed:
docbook2mdoc.c (1.97), "Exp", lines: +9 -5
do not access null pointers when the document element is <title>
2019-04-07 17:39 schwarze
Changed:
parse.c (1.19), "Exp", lines: +3 -3
Discard attributes of ignored elements. As a side benefit, this
avoids a null pointer access if the document element is ignored and
has attributes.
2019-04-07 17:00 schwarze
Changed:
docbook2mdoc.c (1.96), "Exp", lines: +39 -7
node.c (1.4), "Exp", lines: +5 -2
node.h (1.11), "Exp", lines: +4 -1
parse.c (1.18), "Exp", lines: +2 -2
statistics.c (1.15), "Exp", lines: +3 -1
better handling of <link> and <ulink> elements including endterm=,
linkend=, url=, and xlink:href= attributes
2019-04-07 15:06 schwarze
Changed:
docbook2mdoc.c (1.95), "Exp", lines: +7 -1
macro.c (1.7), "Exp", lines: +2 -1
some nodes never continue an existing macro line
2019-04-07 14:49 schwarze
Changed:
docbook2mdoc.c (1.94), "Exp", lines: +15 -1
parse.c (1.17), "Exp", lines: +45 -3
handle preceding a macro without intervening whitespace
2019-04-07 13:16 schwarze
Changed:
docbook2mdoc.c (1.93), "Exp", lines: +56 -51
Move printing of text nodes out of pnode_print() into a dedicated
function, like for all other node types that require a substantial
amount of code. No functional change.
2019-04-06 22:37 schwarze
Changed:
docbook2mdoc.c (1.92), "Exp", lines: +5 -3
macro.c (1.6), "Exp", lines: +62 -11
macro.h (1.3), "Exp", lines: +2 -1
node.h (1.10), "Exp", lines: +2 -1
parse.c (1.16), "Exp", lines: +34 -28
Store the information whether a node is preceded by whitespace into
the node tree. Use that information in the formatter to suppress
the insertion of whitespace in text-text, macro-text, and
macro-macro node sequences.
Text-macro sequences are not yet handled. They are more
complicated because they require emitting a .Pf macro as part of a
text node depending on the spacing properties of the *following*
macro node.
2019-04-06 13:45 schwarze
Changed:
parse.c (1.15), "Exp", lines: +2 -1
statistics.c (1.14), "Exp", lines: +2 -1
treat <simpara> just like <para>; simplified version of a patch
from Stephen Gregoratto <dev at sgregoratto dot me>
2019-04-06 13:30 schwarze
Changed:
README (1.4), "Exp", lines: +78 -8
update the instructions for adding elements; triggered by a smaller
patch from Stephen Gregoratto <dev at sgregoratto dot me>
2019-04-05 14:37 schwarze
Changed:
parse.c (1.14), "Exp", lines: +193 -184
Split parse_file() into parse_file() to fill the parse buffer and
parse_string() to handle it, to allow parsing from alternative
sources in a subsequent step. Only advance line and column number
when parsing from the main input file. Represent parsing state as
an enum rather than with multiple flags.
2019-04-03 18:52 schwarze
Changed:
docbook2mdoc.c (1.91), "Exp", lines: +21 -2
statistics.c (1.13), "Exp", lines: +3 -1
rudimentary implementation of <simplelist>
2019-04-03 17:53 schwarze
Changed:
docbook2mdoc.c (1.90), "Exp", lines: +7 -15
node.h (1.9), "Exp", lines: +2 -4
parse.c (1.13), "Exp", lines: +6 -4
statistics.c (1.12), "Exp", lines: +14 -1
Various in-line elements: Translate <errorname> to .Er. Translate
<sgmltag> to .Ic. Translate <literal> to .Ql, avoiding .Li. Map
<code> and <userinput> to <literal>. Map <structname> to <type>.
Map <structfield> to <parameter>.
2019-04-03 16:52 schwarze
Changed:
docbook2mdoc.c (1.89), "Exp", lines: +1 -4
statistics.c (1.11), "Exp", lines: +13 -1
do not abort(3) when there is more than one <refmeta>
2019-04-03 16:08 schwarze
Changed:
parse.c (1.12), "Exp", lines: +2 -1
statistics.c (1.10), "Exp", lines: +6 -1
introduce <symbol> as an alias for <constant>
2019-04-03 15:59 schwarze
Changed:
docbook2mdoc.c (1.88), "Exp", lines: +3 -1
treat <synopsis> as a literal display
2019-04-03 15:27 schwarze
Changed:
statistics.c (1.9), "Exp", lines: +16 -9
mark more relations involving parameters, indexterms, and sections
as handled
2019-04-03 15:24 schwarze
Changed:
docbook2mdoc.c (1.87), "Exp", lines: +2 -1
translate <paramdef> to .Fa even outside <funcprototype>
2019-04-03 15:05 schwarze
Changed:
docbook2mdoc.c (1.86), "Exp", lines: +13 -21
statistics.c (1.8), "Exp", lines: +7 -1
Make <funcdef> more robust: If the first child is text, use it for
.Ft, then use all remaining children for .Fo.
Make <funcprototype> more robust: Use all children but the
first <funcdef> for .Fa.
2019-04-03 15:04 schwarze
Changed:
macro.c (1.5), "Exp", lines: +10 -8
Make macro_addnode() more flexible: Iteratively descend singleton
children to find a singleton text node.
2019-04-03 14:02 schwarze
Changed:
docbook2mdoc.c (1.85), "Exp", lines: +2 -5
statistics.c (1.7), "Exp", lines: +8 -1
Do not handle <varlistentry> in pnode_print() because it generates
.It which must not occur outside .Bl.
2019-04-03 13:42 schwarze
Changed:
docbook2mdoc.c (1.84), "Exp", lines: +2 -3
macro.c (1.4), "Exp", lines: +3 -2
Improve robustness by allowing macro_close() in any state. This is
relevant because pnode_print() may change state.
2019-04-03 12:29 schwarze
Changed:
docbook2mdoc.c (1.83), "Exp", lines: +36 -12
statistics.c (1.6), "Exp", lines: +20 -1
print tables containing two columns as .Bl -tag
2019-04-03 11:46 schwarze
Changed:
docbook2mdoc.c (1.82), "Exp", lines: +38 -12
node.c (1.3), "Exp", lines: +2 -1
node.h (1.8), "Exp", lines: +2 -2
parse.c (1.11), "Exp", lines: +2 -2
Prints tables containing only one column as .Bl -bullet -compact.
The number of columns is taken from the "cols" attribute.
No neat to treat <informaltable> separately from <table>;
the only difference is whether or not it has a title.
Treat <table> as transparent and handle <tgroup> instead.
The advantages are that <title> gets a generic handler which also
works in other contexts and that other children of <table> are now
covered as well.
2019-04-03 11:23 schwarze
Changed:
parse.c (1.10), "Exp", lines: +6 -4
statistics.c (1.5), "Exp", lines: +6 -4
handle attributes in single quotes
2019-04-03 09:49 schwarze
Changed:
statistics.c (1.4), "Exp", lines: +55 -2
optional parent-child histogram for one relation
2019-04-03 08:39 schwarze
Changed:
statistics.c (1.3), "Exp", lines: +14 -1
merge parse.c rev. 1.8: skip XML comments containing greater-than
characters
2019-04-03 08:36 schwarze
Changed:
Makefile (1.23), "Exp", lines: +3 -3
link with -g for easier debugging and clean statistics.core
2019-04-02 17:06 schwarze
Changed:
docbook2mdoc.c (1.81), "Exp", lines: +1 -2
handle trailing delimiters after <citerefentry>/.Xr; bug reported
by Stephen Gregoratto <dev at sgregoratto dot me>
2019-04-02 16:24 schwarze
Changed:
docbook2mdoc.c (1.80), "Exp", lines: +11 -1
use the idiom ".An Name Aq Mt email" for author email addresses;
issue reported by Stephen Gregoratto <dev at sgregoratto dot me>
2019-04-02 15:53 schwarze
Changed:
docbook2mdoc.c (1.79), "Exp", lines: +8 -1
macro.c (1.3), "Exp", lines: +5 -4
node.h (1.7), "Exp", lines: +2 -1
parse.c (1.9), "Exp", lines: +120 -7
Translate XML character entity references to roff character escape
sequences. Missing feature reported by Stephen Gregoratto <dev at
sgregoratto dot me>.
Remaining known issues: * Whitespace handling isn't perfect
yet. * Numeric character references aren't handled yet. * The
list of entities is still very incomplete. * When it grows longer,
we may have to switch to binary search. * Local entities declared
in the DTD are not yet handled.
2019-04-02 13:11 schwarze
Changed:
parse.c (1.8), "Exp", lines: +25 -1
skip XML comments even if they contain greater-than characters;
issue reported by Stephen Gregoratto <dev at sgregoratto dot me>
2019-03-29 18:09 schwarze
Changed:
statistics.c (1.2), "Exp", lines: +44 -12
allow excluding relations that are already fully implemented
2019-03-29 15:55 schwarze
Added:
statistics.c (1.1)
Changed:
Makefile (1.22), "Exp", lines: +4 -0
Add a utility for docbook2mdoc developers to collect element usage
and parenting statistics, to help decide which nodes should be most
urgently worked on.
2019-03-28 20:41 schwarze
Changed:
docbook2mdoc.c (1.78), "Exp", lines: +64 -21
macro.c (1.2), "Exp", lines: +44 -10
macro.h (1.2), "Exp", lines: +4 -1
node.h (1.6), "Exp", lines: +2 -1
parse.c (1.7), "Exp", lines: +5 -4
Implement a formatter for <author> elements, handling <contrib>,
<personname>, <firstname>, <othername>, <surname>, as well as
arbitrary children properly.
This required minor work on the formatting infrastructure:
Improve macro_addnode() such that it also handles text nodes. Add
a companion function print_textnode(). Let print_text() optionally
work without ARG_SPACE.
Triggered by a report from Stephen Gregoratto <dev at
sgregoratto dot me> that and how GTK documentation uses <contrib>.
2019-03-28 15:05 schwarze
Changed:
parse.c (1.6), "Exp", lines: +42 -33
Simple warn_msg() and error_msg() functions to avoid repetitive
code. While here, drop the warning about unknown attributes.
Those are just too abundant and hardly matter.
2019-03-28 12:21 schwarze
Changed:
Makefile (1.21), "Exp", lines: +1 -1
index.xml (1.13), "Exp", lines: +4 -6
main.c (1.3), "Exp", lines: +3 -3
node.c (1.2), "Exp", lines: +3 -3
node.h (1.5), "Exp", lines: +2 -1
parse.c (1.5), "Exp", lines: +322 -100
The expat library aborts parsing as soon as it encounters invalid
input, and the basic design of the library practically precludes
fixing it. However, whether the input is well-formed XML or not is
totally irrelevant, and in fact, i have seen real-world documents
from X.org that expat rejects as not well-formed. Kristaps reports
the same from OpenGL.
We really want to parse *ANYTHING* whatsoever without ever
throwing a fatal error - after all, the point is to convert legacy
documents to a better format, and nitpicking about the syntax
merely alienates users (including myself).
Consequently, ditch expat and write a parser from scratch,
optimized for robustness on invalid input.
Oh, and by the way, it only requires 200 lines of code,
compared to 15,000 lines in expat - an economy of 98.5% at the sime
time as being much more useful in practice.
2019-03-26 22:39 schwarze
Changed:
docbook2mdoc.c (1.77), "Exp", lines: +1 -3
node.h (1.4), "Exp", lines: +4 -6
parse.c (1.4), "Exp", lines: +40 -13
Provide a way to exclude elements including their children from the
tree and use that for <anchor>, <indexterm>, <primary>, and
<secondary>.
2019-03-26 21:52 schwarze
Changed:
docbook2mdoc.c (1.76), "Exp", lines: +1 -4
node.h (1.3), "Exp", lines: +1 -8
parse.c (1.3), "Exp", lines: +8 -8
mark the first seven elements as ignored, shortening enum nodeid,
and in one case even pnode_print()
2019-03-26 21:47 schwarze
Changed:
Makefile (1.20), "Exp", lines: +2 -1
add forgotten macro.{c,h} dependency rules
2019-03-26 20:54 schwarze
Changed:
node.h (1.2), "Exp", lines: +3 -2
parse.c (1.2), "Exp", lines: +31 -11
Provide an easy way to parse an XML element without generating a
node, either ignoring it outright or emitting a warning if -W was
specified. Use this to handle <xi:include> more cleanly, fixing
two FIXMEs.
2019-03-26 20:06 schwarze
Changed:
main.c (1.2), "Exp", lines: +8 -2
Parsing errors are too easy to miss because they typically emit a
one-line error message followed by lengthy formatted output of what
was parsed before the error occurred.
Make parsing errors more conspicious by taking two steps:
1. If there was a parsing error, print a blank line to
standard output before starting formatted output, for better
separation.
2. After the formatted output, if there was a parsing
error, print the following to standard error: a blank line, a
message that output is incomplete, and another blank line.
2019-03-26 19:17 schwarze
Added:
macro.c (1.1)
macro.h (1.1)
Changed:
Makefile (1.19), "Exp", lines: +3 -3
docbook2mdoc.c (1.75), "Exp", lines: +2 -202
The file docbook2mdoc.c is still large, so split out the macro line
formatter, which is quite self-contained.
2019-03-26 18:32 schwarze
Deleted:
extern.h (1.34)
Added:
format.h (1.1)
main.c (1.1)
node.c (1.1)
node.h (1.1)
parse.c (1.1)
parse.h (1.1)
Changed:
Makefile (1.18), "Exp", lines: +15 -6
docbook2mdoc.c (1.74), "Exp", lines: +56 -645
The program docbook2mdoc(1) has become large enough that splitting
it into a number of logical components makes sense: node tree,
parser, formatter, each with interface and implementation, and the
main program. That way, it becomes easier to see what interacts
with what, and what is independent of what.
2019-03-25 23:14 schwarze
Changed:
docbook2mdoc.c (1.73), "Exp", lines: +17 -33
extern.h (1.33), "Exp", lines: +2 -2
Various parser simplifications and improvements.
Delete the redundant member "node" of struct parse. The
same is already available from cur->node.
No need to check in xml_char() whether an element is open.
Text outside the document element results in expat errors "not
well-formed (invalid token)" or "junk after document element" and
the function xml_char() is not called.
No need to check in xml_elem_end() whether an element is
open. Bogus closing tags result in expat errors "not well-formed
(invalid token)" or "mismatched tag" and the function
xml_elem_end() is not called.
In xml_elem_start(): no point in skipping the element name
check for the document element; and the error error "multiple
refentries" can no longer happen due to earlier cleanups.
2019-03-25 17:28 schwarze
Changed:
docbook2mdoc.c (1.72), "Exp", lines: +128 -139
For macro_addarg(), macro_addnode(), and macro_nodeline(), provide
a flag ARG_SINGLE to request quoting of strings containing
whitespace. Use it for .Dt, .Fa, .Fo, .Nm, and .Xr.
Do not \&-escape macros in quoted strings. Escape quote
characters in macro arguments. NUL-terminate the content of text
nodes. Eliminate the awkward "bufappend()" global output buffer.
Simplify and improve handling of <paramdef>.
Substantial functional improvements, easier to read, yet
minus 20 LOC.
2019-03-24 23:48 schwarze
Changed:
docbook2mdoc.c (1.71), "Exp", lines: +37 -44
To avoid use after free, use TAILQ_FOREACH_SAFE(3) rather than
TAILQ_FOREACH(3) when deleting list elements during the iteration.
Factor out some repeated code into a new function
pnode_printtitle().
Where pnode_print() calls per-element pnode_print*()
functions, call exactly one function per element and do everything
that is required inside, making the huge function pnode_print()
slightly smaller and the various pnode_print*() more
self-contained. In particular, call pnode_unlinksub() as close as
possible to the place where the processing justifying the deletion
was done.
2019-03-24 21:00 schwarze
Changed:
docbook2mdoc.c (1.70), "Exp", lines: +24 -9
Avoid reckless use of low-level stdio output functions like
putchar(3), puts(3), and fputs(3) in high-level formatting code.
For clarity and robustness, be explicit whether we are printing to
a text line with print_text() or to a macro line with
macro_addarg().
While here, fix NODE_REFNAME formatting which i broke
previously.
For now, leave MathML to eqn(7) translation alone.
Usually, that happens in .EQ blocks, i.e. writing text lines. But
it can also happen in inline ($$) context on macro lines. That is
certainly very fragile and will often fail especially for
non-trivial formulae, but switching to print_text() would break it
even more. So postpone repairs until i come round to work on
equation formatting for real.
2019-03-24 16:45 schwarze
Changed:
docbook2mdoc.c (1.69), "Exp", lines: +260 -348
As usual, whitespace handling is the heart and soul of roff(7)
processing, so overhaul it. Keep output line state and provide a
set of output functions to open and close macro lines and to add
arguments to them. Mostly avoid asserting output line state
because elements can occur in any order and we should just deal
with whatever comes in.
This diff shortens the code by 60 lines, improves
robustness and readability, and fixes some bugs.
2019-03-23 11:13 schwarze
Changed:
docbook2mdoc.c (1.68), "Exp", lines: +6 -1
translate <citetitle> to .%T
2019-03-23 10:53 schwarze
Changed:
docbook2mdoc.c (1.67), "Exp", lines: +6 -1
translate <keysym> to .Sy
2019-03-23 10:22 schwarze
Changed:
docbook2mdoc.c (1.66), "Exp", lines: +4 -4
the handlers for NODE_LITERALLAYOUT and NODE_PROGRAMLISTING were
exchanged
2019-03-22 19:44 schwarze
Changed:
docbook2mdoc.c (1.65), "Exp", lines: +8 -1
extern.h (1.32), "Exp", lines: +3 -1
support <personname> and <email>; based on a patch from Stephen
Gregoratto <dev at sgregoratto dot me>
2019-03-22 19:07 schwarze
Changed:
docbook2mdoc.c (1.64), "Exp", lines: +124 -138
KNF: eliminate Yoda notation, and also drop a few redundant
assertions
2019-03-22 18:02 schwarze
Changed:
docbook2mdoc.c (1.63), "Exp", lines: +15 -15
* Fix the return value from main(). * Do not use the pointless
constants EXIT_SUCCESS and EXIT_FAILURE. * KNF: remove parentheses
from return statements.
2019-03-22 17:42 schwarze
Changed:
docbook2mdoc.c (1.62), "Exp", lines: +126 -130
KNF: remove parentheses from switch cases
2019-03-22 17:28 schwarze
Changed:
docbook2mdoc.c (1.61), "Exp", lines: +16 -8
no output for PARA right inside ENTRY, and fix the same in LISTITEM
2019-03-22 17:01 schwarze
Changed:
docbook2mdoc.c (1.60), "Exp", lines: +25 -5
extern.h (1.31), "Exp", lines: +3 -1
handle class="monospaced" in NODE_LITERALLAYOUT
2019-03-22 16:55 schwarze
Changed:
docbook2mdoc.c (1.59), "Exp", lines: +26 -34
factor out new function pnode_getattr_raw()
2019-03-22 16:50 schwarze
Changed:
docbook2mdoc.c (1.58), "Exp", lines: +13 -1
extern.h (1.30), "Exp", lines: +2 -1
translate NODE_LINK to .Sx
2019-03-22 16:47 schwarze
Changed:
docbook2mdoc.c (1.57), "Exp", lines: +3 -1
translate NODE_FIRSTTERM to .Em
2019-03-22 16:42 schwarze
Changed:
docbook2mdoc.c (1.56), "Exp", lines: +146 -181
extern.h (1.29), "Exp", lines: +1 -9
Work towards reducing the number of node IDs.
Allow mapping several node names to the same node ID and
use that for chapter, part, refsect*, sect* -> NODE_SECTION. Do
not require the root to be of a specific type. Delete the useless
NODE_IGNTEXT.
2019-03-22 16:35 schwarze
Changed:
docbook2mdoc.c (1.55), "Exp", lines: +3 -1
ignore NODE_INDEXTERM for now
2019-03-22 16:32 schwarze
Changed:
docbook2mdoc.c (1.54), "Exp", lines: +37 -10
new function pnode_printpara() to avoid printing redundant .Pp, for
example after section headers
2019-03-22 16:25 schwarze
Changed:
docbook2mdoc.c (1.53), "Exp", lines: +2 -2
NODE_ITEMIZEDLIST is .Bl -bullet, not .Bl -enum
2019-03-22 16:21 schwarze
Changed:
docbook2mdoc.c (1.52), "Exp", lines: +67 -47
automatic assignment of header levels
2019-03-22 16:14 schwarze
Changed:
docbook2mdoc.c (1.51), "Exp", lines: +13 -8
get the .Dt name from the id attribute of the root node
2019-03-22 16:10 schwarze
Changed:
docbook2mdoc.c (1.50), "Exp", lines: +93 -4
extern.h (1.28), "Exp", lines: +52 -0
many new nodes found below /usr/xenocara/proto/xorgproto/specs/
2019-03-22 15:54 schwarze
Deleted:
rules.c (1.27)
Changed:
Makefile (1.17), "Exp", lines: +5 -5
docbook2mdoc.c (1.49), "Exp", lines: +1 -20
extern.h (1.27), "Exp", lines: +0 -12
Delete the rest of rules.c.
Just like validation of element nesting is pointless,
validating attributes makes no sense either.
2019-03-22 15:38 schwarze
Changed:
README (1.3), "Exp", lines: +1 -20
docbook2mdoc.c (1.48), "Exp", lines: +2 -14
extern.h (1.26), "Exp", lines: +0 -1
rules.c (1.26), "Exp", lines: +0 -1331
Delete the isparent() validation function.
The DocBook language is totally ill-designed, a gigantic
accretion of arbitrary elements without any kind of discernible
design or cohesion, with an absurdly large number of arbitrary and
pointless rules of what is allowed to nest inside which other
elements. So attempting to validate DocBook input against the
totally crazy specification makes no sense whatsoever, and even
less so because we are certainly not encouraging anybody to write
new or maintain existing DocBook documents. The whole point of
having a DocBook parser is to be able to parse legacy documents and
convert them to a sane language, so it is utterly irrelevant
whether the input is considered valid or invalid by some idiotic
standard.
This deletion allows substantial simplification of the code
and will massively speed up development in the future.
Kristaps@ says that i can move forward with development
without asking for individual OKs.
2019-03-17 13:53 schwarze
Changed:
docbook2mdoc.1 (1.9), "Exp", lines: +5 -5
minor improvements from Stephen Gregoratto <dev at sgregoratto dot
me>
2019-03-08 15:18 schwarze
Changed:
docbook2mdoc.c (1.47), "Exp", lines: +3 -3
for the null character '\0', use the standard shorthand NUL, not
"nil"
2019-03-08 15:09 schwarze
Changed:
docbook2mdoc.c (1.46), "Exp", lines: +13 -13
fix systematically wrong (int) casts in character classification
functions
2019-03-08 10:13 schwarze
Changed:
README (1.2), "Exp", lines: +10 -4
update links to the DocBook website and fix a typo in a function
name
2019-03-08 10:04 schwarze
Changed:
docbook2mdoc.c (1.45), "Exp", lines: +6 -5
improve error handling and usage in main()
UnifiedSplitRaw
