mandoc/TODO - diff

Return to TODO CVS log

Up to [cvsweb.bsd.lv] / mandoc

Diff for /mandoc/TODO between version 1.133 and 1.308

-version 1.133, 2012/05/31 01:54:45
+version 1.308, 2020/11/12 20:46:28
 Line 3
 Line 3
 Line 3
  * $Id$
  ************************************************************************
+ Many issues are annotated for difficulty as follows:
+  - loc = locality of the issue
+     *    single file issue, affects file only, or very few
+     **   single module issue, affects several files of one module
+     ***  cross-module issue, significantly impacts multiple modules
+          and may require substantial changes to internal interfaces
+  - exist = difficulty of the existing code in this area
+     *    affected code is straightforward and easy to read and change
+     **   affected code is somewhat complex, but once you understand
+          the design, not particularly difficult to understand
+     ***  affected code uses a special, exceptionally tricky design
+  - algo = difficulty of the new algorithm to be written
+     *    the required logic and code is straightforward
+     **   the required logic is somewhat complex and needs a careful design
+     ***  the required logic is exceptionally tricky,
+          maybe an approach to solve that is not even known yet
+  - size = the amount of code to be written or changed
+     *    a small number of lines (at most 100, usually much less)
+     **   a considerable amount of code (several dozen to a few hundred)
+     ***  a large amount of code (many hundreds, maybe thousands)
+  - imp = importance of the issue
+     *    mostly for completeness
+     **   would be nice to have
+     ***  issue causes considerable inconvenience
+ Obviously, as the issues have not been solved yet, these annotations
+ are mere guesses, and some may be wrong.
  ************************************************************************
+ * assertion failures
+ ************************************************************************
+ - .if n .ce in the middle of .TS data
+   afl case f1/id:000103,sig:06,src:009024+009105,op:splice,rep:2 (jes@)
+   While roff_parseln() prevents .ce and similar requests in the middle
+   of a tbl, the guard is no longer effective when the .ce is wrapped
+   in a roff block, for example a conditional.  The resulting assertion
+   has never been seen in any real-world manual page.
+   This is too dangerous to fix before release because it requires
+   reorganizing the very delicate internals of roff_parseln(),
+   which risks causing more severe bugs.
+   loc * exist *** algo *** size * imp *
+ ************************************************************************
  * missing features
  ************************************************************************
  --- missing roff features ----------------------------------------------
- - roff.c should treat \n(.H>23 and \n(.V>19 in the pod2man(1)
-   preamble as true, see for example AUTHORS in MooseX::Getopt.3p
-   reported by Andreas Voegele <mail at andreasvoegele dot com>
-   Tue, 22 Nov 2011 15:34:47 +0100 on ports@
  - .ad (adjust margins)
    .ad l -- adjust left margin only (flush left)
    .ad r -- adjust right margin only (flush right)
-Line 22
+Line 61
 Line 22
 Line 61
    .na   -- temporarily disable adjustment without changing the mode
    .ad   -- re-enable adjustment without changing the mode
    Adjustment mode is ignored while in no-fill mode (.nf).
+   loc ***  exist ***  algo **  size **  imp **  (parser reorg would help)
- - .it (line traps) occur in mysql(1), yasm_arch(7)
+ - .fc (field control)
-   generated by DocBook XSL Stylesheets v1.71.1 <http://docbook.sf.net/>
+   found by naddy@ in xloadimage(1)
-   reported by brad@  Sat, 15 Jan 2011 15:48:18 -0500
+   loc **  exist ***  algo *  size *  imp *
  - .ns (no-space mode) occurs in xine-config(1)
+   when implementing this, also let .TH set it
    reported by brad@  Sat, 15 Jan 2011 15:45:23 -0500
+   loc ***  exist ***  algo ***  size **  imp *
- - xloadimage(1) wants .ti (temporary indent), rep by naddy@
+ - \w'' improve width measurements
-   reported again by bentley@ in nmh(1)  Mon, 23 Apr 2012 13:38:28 -0600
+   would not be very useful without an expression parser, see below
-   also uses .ce (center N lines) and .fc (field control)
+   needed for Tcl_NewStringObj(3) via wiz@  Wed, 5 Mar 2014 22:27:43 +0100
+   loc **  exist ***  algo ***  size *  imp ***
- - .ta (tab settings) occurs in ircbug(1) and probably gnats(1)
+ - .als only works for macros in mandoc, not for user-defined strings.
-   reported by brad@  Sat, 15 Jan 2011 15:50:51 -0500
+   Also, the "val" field in struct roffkv would have to be replaced
+   with a pointer to a reference-counted wrapper, and an alias
+   would have to point to the same wrapper as the original.
+   .als to undefined does nothing; the alias is not created.
+   .rm'ing the original leaves the alias to point to the old value.
+   .de .als .de changes both, but
+   .de .als .rm .de only changes the new value, not the alias.
+   Found in groffer(1) version 1.19
+   Jan Stary 20 Apr 2019 20:16:54 +0200
+   loc *  exist **  algo **  size **  imp *
- - \c (interrupted text) should prevent the line break
+ - roff string condition comparisons fail when vars contain quotes:
-   even inside .Bd literal; that occurs in chat(8)
+   .ds s '
+   .if '\*s'' \&...
+   hard to fix because of the basic architecture (string replacement
+   happens before roff(7) syntax parsing)
+   Found in groffer(1) version 1.19
+   Jan Stary 20 Apr 2019 20:16:54 +0200
+   loc *  exist ***  algo ***  size **  imp *
- - using undefined strings or macros defines them to be empty
+ - mandoc replaces all ASCII control characters except tab and line feed
-   wl@  Mon, 14 Nov 2011 14:37:01 +0000
+   with '?' during input.  It would be better to replace them with
+   Unicode escapes in preconv_encode() or somewhere in the vicinity,
+   such that the already existing better replacement strings show
+   up in the output.  Emulating groff is not desirable: groff replaces
+x00, 0x0b, and 0x0d to 0x1f with the empty string (bad because
+   that's easy to overlook for the document author), 0x01 with '.'
+   (very confusing), and passes through 0x02 to 0x08, 0x0c, and 0x7f
+   raw (bad because that is insecure output).  Remember that 0x07 may
+   need special handling because it is sometimes used for certain
+   delimiters, so it may need handling *after* roff.c rather than before.
+   reminded by John Gardner 16 Jun 2020 14:26:28 +1000
+   Actually, more ASCII control characters than just 0x07 may need
+   later handling because they can for example be used in macro names.
+   So they may need handling after roff(7) processing.
+   pointed out by John Gardner 23 Jun 2020 18:28:08 +1000
+   more info from John Gardner 29 Jun 2020 19:54:04 +1000
+   loc **  exist **  algo **  size **  imp *
+ - many missing features used in old groff_char(7),
+   some can possibly be supported
+   kamil at netbsd 12 Nov 2020 17:27:09 +0100 + reply
  --- missing mdoc features ----------------------------------------------
- - fix bad block nesting involving multiple identical explicit blocks
+ - .Sh and .Ss should be parsed and partially callable, see groff_mdoc(7)
-   see the OpenBSD mdoc_macro.c 1.47 commit message
+   reed at reedmedia dot net Sat, 21 Dec 2019 17:13:07 -0600
+   loc **  exist **  algo **  size **  imp *
  - .Bl -column .Xo support is missing
    ultimate goal:
-Line 54
+Line 133
 Line 54
 Line 133
    lib/libc/compat-43/sigvec.3
    lib/libc/gen/signal.3
    lib/libc/sys/sigaction.2
+   loc *  exist ***  algo ***  size *  imp **
  - edge case: decide how to deal with blk_full bad nesting, e.g.
    .Sh .Nm .Bk .Nm .Ek .Sh found by jmc@ in ssh-keygen(1)
    from jmc@  Wed, 14 Jul 2010 18:10:32 +0100
+   loc *  exist ***  algo ***  size **  imp **
- - \\ is now implemented correctly
-   * when defining strings and macros using .ds and .de
-   * when parsing roff(7) and man(7) macro arguments
-   It does not yet work in mdoc(7) macro arguments
-   because libmdoc does not yet use mandoc_getarg().
-   Also check what happens in plain text, it must be identical to \e.
  - .Bd -filled should not be the same as .Bd -ragged, but align both
    the left and right margin.  In groff, it is implemented in terms
    of .ad b, which we don't have either.  Found in cksum(1).
+   loc ***  exist ***  algo **  size **  imp **  (parser reorg would help)
  - implement blank `Bl -column', such as
    .Bl -column
    .It foo Ta bar
    .El
+   loc *  exist ***  algo ***  size *  imp *
  - explicitly disallow nested `Bl -column', which would clobber internal
    flags defined for struct mdoc_macro
+   loc *  exist *  algo *  size *  imp **
  - In .Bl -column .It, the end of the line probably has to be regarded
    as an implicit .Ta, if there could be one, see the following mildly
-Line 87
+Line 164
 Line 87
 Line 164
      Default search path.
    reported by Michal Mazurek <akfaew at jasminek dot net>
    via jmc@ Thu, 7 Apr 2011 16:00:53 +0059
+   loc *  exist ***  algo **  size *  imp **
  - inside `.Bl -column' phrases, punctuation is handled like normal
    text, e.g. `.Bl -column .It Fl x . Ta ...' should give "-x -."
-Line 95
+Line 173
 Line 95
 Line 173
    is not safe, e.g. `.Bl -column .It Pf a b .' gives "ab."
    but should give "ab ."
- - set a meaningful default if no `Bl' list type is assigned
- - have a blank `It' head for `Bl -tag' not puke
  - prohibit `Nm' from having non-text HEAD children
    (e.g., NetBSD mDNSShared/dns-sd.1)
    (mdoc_html.c and mdoc_term.c `Nm' handlers can be slightly simplified)
+ - support translated section names
+   e.g. x11/scrotwm scrotwm_es.1:21:2: error: NAME section must be first
+   that one uses NOMBRE because it is spanish...
+   deraadt tends to think that section-dependent macro behaviour
+   is a bad idea in the first place, so this may be irrelevant
+   loc **  exist **  algo **  size *  imp **
  - When there is free text in the SYNOPSIS and that free text contains
    the .Nm macro, groff somehow understands to treat the .Nm as an in-line
    macro, while mandoc treats it as a block macro and breaks the line.
-Line 110
+Line 191
 Line 110
 Line 191
    should be, needs investigation.
    uqs@  Thu, 2 Jun 2011 11:03:51 +0200
    uqs@  Thu, 2 Jun 2011 11:33:35 +0200
+   loc *  exist **  algo ***  size *  imp **
  --- missing man features -----------------------------------------------
- - groff an-ext.tmac macros (.UR, .UE) occur in xine(5)
+ - groff_www(7) .MTO and .URL
-   reported by brad@  Sat, 15 Jan 2011 15:45:23 -0500
+   These macros were used by the GNU grep(1) man page.
+   The groff_www(7) manual page itself uses them, too.
+   We should probably *not* add them to mandoc.
+   Just mentioning this here to keep track of the abuse.
+   Laura Morales <lauretas at mail dot com> 20 Apr 2018 07:33:02 +0200
+   loc **  exist *  algo *  size **  imp *
- - -T[x]html doesn't stipulate non-collapsing spaces in literal mode
  --- missing tbl features -----------------------------------------------
- - implement basic non-parametric .de to support e.g. sox(1)
+ - vertical centering in cells vertically spanned with ^
-   reported by naddy@ Sat, 16 Oct 2010 23:51:57 +0200
+   pali dot rohar at gmail dot com 16 Jul 2018 13:03:35 +0200
-   *** sox(1) still doesn't work, tbl(1) errors need investigation
+   loc *  exist ***  algo ***  size **  imp *
- - allow standalone `.' to be interpreted as an end-of-layout
+ - support mdoc(7) and man(7) macros inside tbl(7) code;
-   delimiter instead of being thrown away as a no-op roff line
+   probably requires the parser reorg and letting tbl(7)
-   reported by Yuri Pankov, Wed 18 May 2011 11:34:59 CEST
+   use roff_node such that macro sets can mix;
+   informed by bapt@ that FreeBSD needs this: 3 Jan 2015 23:32:23 +0100
+   loc ***  exist **  algo ***  size **  imp ***
+ - look at the POSIX manuals in the books/man-pages-posix port,
+   they use some unsupported tbl(7) features, mostly macros in tbl(7).
+   loc *  exist **  algo **  size **  imp ***
+ - look what Joerg Schilling manual pages use
+   Thu, 19 Mar 2015 18:31:48 +0100
+ --- missing eqn features -----------------------------------------------
+ - In a matrix, break the output line after each matrix line.
+   Found in the discussion at CDBUG 2015.  Suggested by Avi Weinstock.
+   This may not be the ideal solution after all: eqn(7) matrices
+   are lists of columns, so Avi's proposal would show each *column*
+   on its own *line*, which is likely to cause confusion.
+   A better solution, but much harder to implement, would be to
+   actually show the coordinates of column vectors on different
+   terminal output lines, using the clumnated output facilities
+   developed for .Bl -tag, .Bl -column, and also used for tbl(7).
+   loc *  exist *  algo **  size **  imp **
+ - The "size" keyword is parsed, but ignored by the formatter.
+   loc *  exist *  algo *  size *  imp *
+ - The spacing characters `~', `^', and tab are currently ignored,
+   see User's Guide (Second Edition) page 2 section 4.
+   loc *  exist *  algo **  size *  imp **
+ - Mark and lineup are parsed and ignored,
+   see User's Guide (Second Edition) page 5 section 15.
+   loc **  exist **  algo **  size **  imp **
+ - GNU eqn converts some operators to special characters, for example,
+   input HYPHEN-MINUS becomes output \(mi, unless it is part of a
+   quoted word.  mandoc(1) only does this when the operator is
+   surrounded by blanks, not when it is part of an unquoted word.
+   Also, check whether there are more such cases (e.g., +?).
+   reported by bentley@  20 Jun 2017 02:04:29 -0600
+   loc *  exist **  algo **  size *  imp *
+ - Primes, opprime, and '
+   bentley@  Thu, 13 Jul 2017 23:14:20 -0600
  --- missing misc features ----------------------------------------------
+ - -T man does not handle eqn(7) and tbl(7)
+   Stephen Gregoratto 16 Feb 2020 01:28:07 +1100
+   loc **  exist **  algo **  size ***  imp **
+ - man -ks 1,8 route; kn@ Jul 13, 2018 orally
+ - italic correction (\/) in PostScript mode
+   Werner LEMBERG on groff at gnu dot org  Sun, 10 Nov 2013 12:47:46
+   loc **  exist **  algo *  size *  imp *
+ - change the default PAGER to more -Es and use the pager
+   even for apropos title line output; req by bapt@
+   loc *  exist *  algo *  size *  imp ***
  - clean up escape sequence handling, creating three classes:
    (1) fully implemented, or parsed and ignored without loss of content
    (2) unimplemented, potentially causing loss of content
        or serious mangling of formatting (e.g. \n) -> ERROR
        see textproc/mgdiff(1) for nice examples
    (3) undefined, just output the character -> perhaps WARNING
+   loc ***  exist **  algo **  size **  imp *** (parser reorg helps)
- - The \t escape sequence is the same as a literal tab, see for example
+ - kettenis wants base roff, ms, and me  Fri, 1 Jan 2010 22:13:15 +0100 (CET)
-   the ASCII table in hexdump(1) where
+   loc **  exist **  algo **  size ***  imp *
-     .Bl -column \&000_nu \&001_so \&002_st \&003_et \&004_eo
-     .It \&000\ nul\t001\ soh\t002\ stx\t003\ etx\t004\ eot\t005\ enq
-   produces
-nul  001 soh  002 stx  003 etx  004 eot  005 enq
-   and the example in oldrdist(1)
+ --- compatibility checks -----------------------------------------------
+ - is .Bk implemented correctly in modern groff?
+   sobrado@  Tue, 19 Apr 2011 22:12:55 +0200
+ - compare output to Heirloom roff, Solaris roff, and
+   http://repo.or.cz/w/neatroff.git  http://litcave.rudi.ir/
+ - look at AT&T DWB http://www2.research.att.com/sw/download
+   Carsten Kunze <carsten dot kunze at arcor dot de> has patches
+   Mon, 4 Aug 2014 17:01:28 +0200
+   ported version: https://github.com/n-t-roff/DWB3.3
+   Carsten Kunze  Wed, 22 Apr 2015 11:21:43 +0200
  - look at pages generated from reStructeredText, e.g. devel/mercurial hg(1)
    These are a weird mixture of man(7) and custom autogenerated low-level
    roff stuff.  Figure out to what extent we can cope.
-Line 152
+Line 305
 Line 152
 Line 305
    noted by stsp@  Sat, 24 Apr 2010 09:17:55 +0200
    reminded by nicm@  Mon, 3 May 2010 09:52:41 +0100
+ - look at pages generated from ronn(1) github.com/rtomayko/ronn
+   (based on markdown)
+ - look at pages generated from Texinfo source by yat2m, e.g. security/gnupg
+   First impression is not that bad.
+ - look at pages generated by pandoc; see
+   https://github.com/jgm/pandoc/blob/master/src/Text/Pandoc/Writers/Man.hs
+   porting planned by kili@  Thu, 19 Jun 2014 19:46:28 +0200
  - check compatibility with Plan9:
    http://swtch.com/usr/local/plan9/tmac/tmac.an
    http://swtch.com/plan9port/man/man7/man.html
    "Anthony J. Bentley" <anthonyjbentley@gmail.com> 28 Dec 2010 21:58:40 -0700
+ - check compatibility with COHERENT troff:
+   http://www.nesssoftware.com/home/mwc/source.php
+ - check compatibility with the man(7) formatter
+   https://raw.githubusercontent.com/rofl0r/hardcore-utils/master/man.c
+ - check compatibility with
+   http://ikiwiki.info/plugins/contrib/mandoc/
+   https://github.com/schmonz/ikiwiki/compare/mandoc
+   Amitai Schlair  Mon, 19 May 2014 14:05:53 -0400
+ - check compatibility with
+   https://git.sr.ht/~sircmpwn/scdoc
+ - check features of the Slackware man.conf(5) format
+   Carsten Kunze  Wed, 11 Mar 2015 17:57:24 +0100
  ************************************************************************
  * formatting issues: ugly output
  ************************************************************************
- - a column list with blank `Ta' cells triggers a spurrious
+ - revisit empty in-line macros
+   look at the difference between "Em x Em ." and "Sq x Em ."
+   Carsten Kunze  Fri, 12 Dec 2014 00:15:41 +0100
+   loc *** exist *** algo *** size * imp **
+ - a column list with blank `Ta' cells triggers a spurious
    start-with-whitespace printing of a newline
- - double quotes inside double quotes are escaped by doubling them
+ - In .Bl -column, .It a<tab>"b<tab>c"
-   implement this in mdoc(7), too
+   shows the quotes in groff, but not in mandoc
-   so far, we only have it in roff(7) and man(7)
+   loc * exist *** algo ** size * imp **
-   reminded by millert@  Thu, 09 Dec 2010 17:29:52 -0500
- - perl(1) SYNOPSIS looks bad; reported by deraadt@
-) man(7) seems to need SYNOPSIS .Nm blocks, too
  - In .Bl -column,
    .It Em Authentication<tab>Key Length
    ought to render "Key Length" with emphasis, too,
    see OpenBSD iked.conf(5).
    reported again Nicolas Joly via wiz@ Wed, 12 Oct 2011 00:20:00 +0200
+   loc *  exist ***  algo ***  size **  imp ***
  - empty phrases in .Bl column produce too few blanks
    try e.g. .Bl -column It Ta Ta
    reported by millert Fri, 02 Apr 2010 16:13:46 -0400
+   loc *  exist ***  algo ***  size *  imp **
  - .%T can have trailing punctuation.  Currently, it puts the trailing
    punctuation into a trailing MDOC_TEXT element inside its own scope.
-Line 190
+Line 373
 Line 190
 Line 373
    slurp all arguments into one single text element - and one feature
    of in_line() - put trailing punctuation out of scope.
    Found in mount_nfs(8) and exports(5), search for "Appendix".
+   loc **  exist **  algo ***  size *  imp **
+ - Trailing punctuation after .%T triggers EOS spacing, at least
+   outside .Rs (eek!).  Simply setting ARGSFL_DELIM for .%T is not
+   the right solution, it sends mandoc into an endless loop.
+   reported by Nicolas Joly  Sat, 17 Nov 2012 11:49:54 +0100
+   loc *  exist **  algo **  size *  imp **
+ - global variables in the SYNOPSIS of section 3 pages
+   .Vt vs .Vt/.Va vs .Ft/.Va vs .Ft/.Fa ...
+   from kristaps@  Tue, 08 Jun 2010 11:13:32 +0200
+ - implicit whitespace around inline equations
+   example code:  where '$times$' denotes matrix multiplication
+   must not have an HTML line break, nor a blank, before <math>
+   partial solution: html.c {"math", HTML_NLINSIDE | HTML_INDENT},
+   bentley@  Thu, 13 Jul 2017 19:00:59 -0600
  - in enclosures, mandoc sometimes fancies a bogus end of sentence
    reminded by jmc@  Thu, 23 Sep 2010 18:13:39 +0059
+   loc *  exist **  algo ***  size *  imp ***
+ - a line starting with "\fB something" counts as starting with whitespace
+   and triggers a line break; found in audio/normalize-mp3(1)
+   This will become easier once escape sequences are represented
+   by syntax tree nodes.
+   loc **  exist *  algo **  size *  imp **
+ - formatting /usr/local/man/man1/latex2man.1 with groff and mandoc
+   reveals lots of bugs both in groff and mandoc...
+   reported by bentley@  Wed, 22 May 2013 23:49:30 -0600
+ --- PostScript and PDF issues ------------------------------------------
+ - PDF output doesn't use a monospaced font for .Bd -literal
+   Example: "mandoc -Tpdf afterboot.8 > output.pdf && pdfviewer output.pdf".
+   Search the text "Routing tables".
+   Also check what PostScript mode does when fixing this.
+   reported by juanfra@ Wed, 04 Jun 2014 21:44:58 +0200
+   instructions from juanfra@  Wed, 11 Jun 2014 02:21:01 +0200
+     add a new <</Type /Font>> block to the PDF files with /BaseFont /Courier
+     and change the /Name from /F0 to the new font (/F5 (?)).
+   re-reported by tb@ Mon, 16 Mar 2015 16:47:21 +0100
+   loc *  exist **  algo **  size *  imp **
+ --- HTML issues --------------------------------------------------------
+ - get rid of the last handful of style= attributes such that
+   Content-Security-Policy: can be enabled without unsafe-inline
+   suggested by bentley@  Nov 10, 2019 at 06:02:49AM -0700
+   loc *  exist *  algo *  size *  imp **
+ - .Bf at the beginning of a paragraph inserts a bogus 1ex horizontal
+   space, see for example random(3).  Introduced in
+   http://mdocml.bsd.lv/cgi-bin/cvsweb/mdoc_html.c.diff?r1=1.91&r2=1.92
+   reported by deraadt@ Mon, 28 Sep 2015 20:14:13 -0600 (MDT)
+   loc **  exist **  algo **  size *  imp *
+ - jsg on icb, Nov 3, 2014:
+   try to guess Xr in man(7) for hyperlinking
+   and render them with <a class="Xr" href=...>
+   https://github.com/Debian/debiman/issues/15
+   loc *  exist *  algo **  size **  imp **
+ - space characters can end up in href= attributes, for example coming
+   from the first .Xr argument (where they make no sense, but still);
+   does this affect other characters, other source macros...?
+   Jackson Pauls  29 Aug 2017 16:56:27 +0100
+ - The tables used to render the three-part page headers actually force
+   the width of the <body> to the max-width given for <html>.
+   Not yet sure how to fix that...
+   Observed by an Anonymous Coward on undeadly.org:
+   http://undeadly.org/cgi?action=article&sid=20140925064244&pid=1
+   loc *  exist *  algo **  size *  imp ***
+ - generate <img> tags in HTML
+   idea from florian@  Tue, 7 Apr 2015 00:26:28 +0000
+   may be possible to implement with .Lk img://something.png alt_text
+ - check https://github.com/trentm/mdocml
  ************************************************************************
  * formatting issues: gratuitous differences
  ************************************************************************
- - .Rv (and probably .Ex) print different text if an `Nm' has been named
+ - .Fn reopens a new scope after punctuation in mandoc,
-   or not (run a manual without `Nm blah' to see this).  I'm not sure
+   but closes its scope for good in groff.
-   that this exists in the wild, but it's still an error.
+   Do we want to change mandoc or groff?
+   Steffen Nurpmeso  Sat, 08 Nov 2014 13:34:59 +0100
+   loc *  exist **  algo **  size *  imp **
- - In .Bl -bullet, the groff bullet is "+\b+\bo\bo", the mandoc bullet
+ - In .Bl -enum -width 0n, groff continues one the same line after
-   is just "o\bo".
+   the number, mandoc breaks the line.
-   see for example OpenBSD ksh(1)
+   mail to kristaps@  Mon, 20 Jul 2009 02:21:39 +0200
+   loc *  exist **  algo **  size *  imp **
- - The characters "|" and "\*(Ba" should never be bold,
-   not even in the middle of a word, e.g. ".Cm b\*(Bac" in
-   "mknod [-m mode] name b|c major minor"
-   in OpenBSD ksh(1)
- - A bogus .Pp between two .It must not produce a double blank line,
-   see between -R and -r in OpenBSD rm(1), before "update" in mount(8),
-   or in DIAGNOSTICS in init(8), or before "is always true" in ksh(1).
-   The same happens with .Pp just before .El, see bgpd.conf(5).
-   Also have `It' complain if `Pp' is invoked at certain times (not
-   -compact?).
  - .Pp between two .It in .Bl -column should produce one,
    not two blank lines, see e.g. login.conf(5).
    reported by jmc@  Sun, 17 Apr 2011 14:04:58 +0059
    reported again by sthen@  Wed, 18 Jan 2012 02:09:39 +0000 (UTC)
+   loc *  exist ***  algo **  size *  imp **
  - If the *first* line after .It is .Pp, break the line right after
    the tag, do not pad with space characters before breaking.
    See the description of the a, c, and i commands in sed(1).
+   loc *  exist **  algo **  size *  imp **
  - If the first line after .It is .D1, do not assert a blank line
    in between, see for example tmux(1).
    reported by nicm@  13 Jan 2011 00:18:57 +0000
+   loc *  exist **  algo **  size *  imp **
- - .Nx 1.0a
+ - Trailing punctuation after .It should trigger EOS spacing.
-   should be "NetBSD 1.0A", not "NetBSD 1.0a",
+   reported by Nicolas Joly  Sat, 17 Nov 2012 11:49:54 +0100
-   see OpenBSD ccdconfig(8).
+   Probably, this should be fixed somewhere in termp_it_pre(), not sure.
+   loc *  exist **  algo **  size *  imp **
- - In .Bl -tag, if a tag exceeds the right margin and must be continued
-   on the next line, it must be indented by -width, not width+1;
-   see "rule block|pass" in OpenBSD ifconfig(8).
  - When the -width string contains macros, the macros must be rendered
    before measuring the width, for example
      .Bl -tag -width ".Dv message"
-Line 246
+Line 498
 Line 246
 Line 498
    as -width 7n, not -width 11n.
    The same applies to .Bl -column column widths;
    reported again by Nicolas Joly Thu, 1 Mar 2012 13:41:26 +0100 via wiz@ 5 Mar
+   reported again by Franco Fichtner Fri, 27 Sep 2013 21:02:28 +0200
+   reported again by Bruce Evans Fri, 17 Feb 2017 21:22:44 +0100 via bapt@
+   loc ***  exist ***  algo ***  size **  imp ***
+   An easy partial fix would be to just skip the first word if it starts
+   with a dot, including any following white space, when measuring.
+   loc *  exist *  algo *  size *  imp ***
  - The \& zero-width character counts as output.
    That is, when it is alone on a line between two .Pp,
    we want three blank lines, not two as in mandoc.
+   loc **  exist **  algo **  size *  imp **
- - When .Fn arguments exceed one output line, all but the first
+ - Sequences of multiple man(7) paragraphs (.PP, .IP) interspersed
-   should be indented, see e.g. rpc(3);
+   with .ps and .nf/.fi produce execessive blank lines, see libJudy
-   reported by jmc@ on discuss@  Fri, 29 Oct 2010 13:48:33 +0100
+   and graphics/dcmtk.  The parser reorg may help with this.
-   reported again by Nicolas Joly via wiz@  Sun, 18 Sep 2011 18:24:40 +0200
-   Also, we don't want to break the line within the argument of:
-   .Fa "chtype tl"
- - .Ns should work when called at the end of an input line, see
-   the following code in vi(1):
-     .It Xo
-     .Op Ar line
-     .Cm a Ns Op Cm ppend Ns
-     .Op Cm !\&
-     .Xc
-     The input text is appended after the specified line.
- - Header lines of excessive length:
-   Port OpenBSD man_term.c rev. 1.25 to mdoc_term.c
-   and document it in mdoc(7) and man(7) COMPATIBILITY
-   found while talking to Chris Bennett
- - In man(7), the sequence
-     .HP
-     one line of regular text
-     .SH
-   should not produce two blank lines before the .SH,
-   see for example named-checkconf(8).
- - In man(7), the sequence
-     .SH HEADER
-     <blank line>
-     .PP
-     regular text
-   should not produce any blank lines between the header and the text,
-   see for example rsync(1).
-   Reported by naddy@  Mon, 28 Mar 2011 20:45:42 +0200
- - In man(7), the sequence
-     regular text
-     .IP
-     .IP "tag"
-     indented text
-   should produce one, not four blank lines between the regular text
-   and the tag, see for example rsync(1).
-   Likewise,
-     regular text
-     .IP
-     indented text
-   should produce one, not two blank lines in between, and
-     regular text
-     .IP
-     .RS
-     .IP tag
-     indented text
-   should produce one, not three blank lines.
-   Reported by naddy@  Mon, 28 Mar 2011 20:45:42 +0200
  - trailing whitespace must be ignored even when followed by a font escape,
    see for example
      makes
      \fBdig \fR
      operate in batch mode
    in dig(1).
+   loc **  exist **  algo **  size *  imp **
  ************************************************************************
- * error reporting issues
+ * warning issues
  ************************************************************************
- - .TP directly followed by .RS gives an assertion.
+ - shorten/simplify error messages for usage errors
+   To: deraadt@ 25 Oct 2020 23:37:01 +0100
+   loc **  exist *  algo *  size **  imp ***
+ - warn about duplicate .Sh/.Ss heads
+   gre(4): Rename duplicate sections 20 Apr 2018 15:27:33 +0200
+   loc *  exist *  algo *  size *  imp **
+ - style message about macros inside .Bd -literal and .Dl, in particular
+   font changing macros like .Cm, .Ar, .Fa (from the mdoclint TODO)
+ - style message about mismatches between the section number in the
+   file name (if it is known) and the section number in .Dt
+   (from the mdoclint TODO)
+ - style message about NULL without .Dv (from the mdoclint TODO)
+ - style message about error constants without .Er (from the mdoclint TODO)
+ - warn when .Sh or .Ss contain other macros
+   Steffen Nurpmeso, savannah.gnu.org/bugs/index.php?45034
+   loc *  exist *  algo *  size *  imp **
+ - style message about violations of the convention
+   .An name Aq Mt localpart@domain in AUTHORS (from the mdoclint TODO)
+ - warn about attempts to call non-callable macros
+   Steffen Nurpmeso  Tue, 11 Nov 2014 22:55:16 +0100
+   Note that formatting is inconsistent in groff.
+   .Fn Po prints "Po()", .Ar Sh prints "file ..." and no "Sh".
+   Relatively hard because the relevant code is scattered
+   all over mdoc_macro.c and all subtly different.
+   loc **  exist **  algo **  size **  imp **
+ - warn about punctuation - e.g. ',' and ';' - at the beginning
+   of a text line, if it is likely intended to follow the preceding
+   output without intervening whitespace, in particular after a
+   macro line (from the mdoclint TODO)
+ - makewhatis -p complains about language subdirectories:
+   /usr/local/man//ru: Unknown directory part
  ************************************************************************
- * performance issues
+ * documentation issues
  ************************************************************************
- Several areas can be cleaned up to make mandoc even faster.  These are
+ - mark macros as: page structure domain, manual domain, general text domain
+   is this useful?
- - improve hashing mechanism for macros (quite important: performance)
+ - mention /usr/share/misc/mdoc.template in mdoc(7)?
- - improve hashing mechanism for characters (not as important)
+ - Is all the content from http://www.std.com/obi/BSD/doc/usd/28.tbl/tbl
+   covered in tbl(7)?
+ ************************************************************************
+ * performance issues
+ ************************************************************************
  - the PDF file is HUGE: this can be reduced by using relative offsets
- - instead of re-initialising the roff predefined-strings set before each
-   parse, create a read-only version the first time and copy it
  ************************************************************************
  * structural issues
  ************************************************************************
- - We use the input line number at several places to distinguish
+ - POSIX says in the documentation of sysconf(3) that PATH_MAX
-   same-line from different-line input.  That plainly doesn't work
+   is allowed to be so large that it is a bad idea to use it
-   with user-defined macros, leading to random breakage.
+   for sizing static buffers.  So use dynamic buffers throughout.
+   See the file test-PATH_MAX.c for details.
+   Found by Aaron M. Ucko in the GNU Hurd via Bdale Garbee,
+   https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=829624
+ - Is it possible to further simplify ENDBODY_SPACE?
  - Find better ways to prevent endless loops
    in roff(7) macro and string expansion.
- - Finish cleanup of date handling.
+ - make buffers for parsing functions const
-   Decide which formats should be recognized where.
+   christos@ via wiz@  Fri, 18 Dec 2015 17:10:01 +0100
-   Update both mdoc(7) and man(7) documentation.
-   Triggered by  Tim van der Molen  Tue, 22 Feb 2011 20:30:45 +0100
+ - struct mparse refactoring
+   Steffen Nurpmeso  Thu, 04 Sep 2014 12:50:00 +0200
+ ************************************************************************
+ * CGI issues
+ ************************************************************************
+  - Inspect httpd(8) logs on man.openbsd.org and consider
+    whether logging can be improved, where bad syntax comes from,
+    and what needs to be done to get rid of COMPAT_OLDURI.
+  - Enable HTTP compression by detecting gzip encoding and filtering
+    output through libz.
+  - Privilege separation (see OpenSSH).
+  - Enable caching support via HTTP 304 and If-Modified-Since.
+ ************************************************************************
+ * to improve in the groff_mdoc(7) macros
+ ************************************************************************
+ - delete OS release verification from .Dx, .Fx, .Nx, .Ox etc.
+   https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=629161
+   also Branden Robinson 18 Dec 2019 00:59:52 +1100
+ - Can the distinction between .Vt and .Va be made stricter,
+   recommending .Vt extern char * Ns Va optarg ; ?
+   What about the block macro properties of .Vt in the SYNOPSIS?
+   zeurkous 25 Dec 2019 08:48:36 +0100
+ - .Cd # arch1, arch2 in section 4 pages:
+   find better way to indicate multiple architectures, maybe:
+   allow .Dt vgafb 4 "macppc sparc64"
+   already shown as "Device Drivers Manual (macppc sparc64)"
+   for apropos, make that "vgafb(4) - macppc # sparc64" instead of "- all"
+   groff can be made to show multiple arches, too, but it is
+   tedious to do the string parsing in roff code...
+   jmc@ 23 Apr 2018 07:24:52 +0100 [man for vgafb(4)...]
+   loc **  exist **  algo *  size *  imp ***
+ - use uname(1) to set doc-default-operating-system at install time
+   tobimensch  Mon, 1 Dec 2014 00:25:07 +0100
+ - apostrophe (39), circumflex (94), grave (96), tilde (126)
+   in manuals: \(aq, \(ha, \`, \(ti
+   Re: [Groff] ASCII Minus Sign in man Pages.
+   bentley@ 26 Apr 2017 10:02:06 -0600
+   Do we need to fix existing manuals?
+   Do we need to fix the definition of the mdoc(7) language?

CVSweb