TODO

   1 ************************************************************************
   2 * Official mandoc TODO.
   3 * $Id: TODO,v 1.125 2011/11/17 17:41:07 schwarze Exp $
   4 ************************************************************************
   5
   6 ************************************************************************
   7 * parser bugs
   8 ************************************************************************
   9
  10 - ".\}" on its own line gets translated to bare ".\&"
  11   which forces pset() into man(7)
  12   and then triggers an unknown macro error
  13   reported by naddy@  Sun, 3 Jul 2011 21:52:24 +0200
  14
  15 - .It is parsed in general, except in .Bl -diag
  16   deraadt@  Mon, 07 Nov 2011 11:10:52 -0700
  17
  18 ************************************************************************
  19 * formatter bugs
  20 ************************************************************************
  21
  22 - tbl(7): Horizontal and vertical lines are formatted badly:
  23   With the box option, there is too much white space at the end of cells.
  24   Horizontal lines from "=" lines are a bit too long.
  25   yuri dot pankov at gmail dot com  Thu, 14 Apr 2011 05:45:26 +0400
  26
  27 ************************************************************************
  28 * missing features
  29 ************************************************************************
  30
  31 --- missing roff features ----------------------------------------------
  32
  33 - .if n \{
  34   .br\}
  35   should cause an extra space to be raised.
  36
  37 - .ad (adjust margins)
  38   .ad l -- adjust left margin only (flush left)
  39   .ad r -- adjust right margin only (flush right)
  40   .ad c -- center text on line
  41   .ad b -- adjust both margins (alias: .ad n)
  42   .na   -- temporarily disable adjustment without changing the mode
  43   .ad   -- re-enable adjustment without changing the mode
  44   Adjustment mode is ignored while in no-fill mode (.nf).
  45
  46 - .it (line traps) occur in mysql(1), yasm_arch(7)
  47   generated by DocBook XSL Stylesheets v1.71.1 <http://docbook.sf.net/>
  48   reported by brad@  Sat, 15 Jan 2011 15:48:18 -0500
  49
  50 - .ns (no-space mode) occurs in xine-config(1)
  51   reported by brad@  Sat, 15 Jan 2011 15:45:23 -0500
  52
  53 - xloadimage(1) wants .ti (temporary indent), rep by naddy@
  54
  55 - .ta (tab settings) occurs in ircbug(1) and probably gnats(1)
  56   reported by brad@  Sat, 15 Jan 2011 15:50:51 -0500
  57
  58 - \c (interrupted text) occurs in chat(8)
  59
  60 - using undefined strings or macros defines them to be empty
  61   wl@  Mon, 14 Nov 2011 14:37:01 +0000
  62
  63 --- missing mdoc features ----------------------------------------------
  64
  65 - fix bad block nesting involving multiple identical explicit blocks
  66   see the OpenBSD mdoc_macro.c 1.47 commit message
  67
  68 - .Bl -column .Xo support is missing
  69   ultimate goal:
  70   restore .Xr and .Dv to
  71   lib/libc/compat-43/sigvec.3
  72   lib/libc/gen/signal.3
  73   lib/libc/sys/sigaction.2
  74
  75 - edge case: decide how to deal with blk_full bad nesting, e.g.
  76   .Sh .Nm .Bk .Nm .Ek .Sh found by jmc@ in ssh-keygen(1)
  77   from jmc@  Wed, 14 Jul 2010 18:10:32 +0100
  78
  79 - \\ is now implemented correctly
  80   * when defining strings and macros using .ds and .de
  81   * when parsing roff(7) and man(7) macro arguments
  82   It does not yet work in mdoc(7) macro arguments
  83   because libmdoc does not yet use mandoc_getarg().
  84   Also check what happens in plain text, it must be identical to \e.
  85
  86 - .Bd -filled should not be the same as .Bd -ragged, but align both
  87   the left and right margin.  In groff, it is implemented in terms
  88   of .ad b, which we don't have either.  Found in cksum(1).
  89
  90 - implement blank `Bl -column', such as
  91   .Bl -column
  92   .It foo Ta bar
  93   .El
  94
  95 - explicitly disallow nested `Bl -column', which would clobber internal
  96   flags defined for struct mdoc_macro
  97
  98 - In .Bl -column .It, the end of the line probably has to be regarded
  99   as an implicit .Ta, if there could be one, see the following mildly
 100   ugly code from login.conf(5):
 101     .Bl -column minpasswordlen program xetcxmotd
 102     .It path Ta path Ta value of Dv _PATH_DEFPATH
 103     .br
 104     Default search path.
 105   reported by Michal Mazurek <akfaew at jasminek dot net>
 106   via jmc@ Thu, 7 Apr 2011 16:00:53 +0059
 107
 108 - inside `.Bl -column' phrases, punctuation is handled like normal
 109   text, e.g. `.Bl -column .It Fl x . Ta ...' should give "-x -."
 110
 111 - inside `.Bl -column' phrases, TERMP_IGNDELIM handling by `Pf'
 112   is not safe, e.g. `.Bl -column .It Pf a b .' gives "ab."
 113   but should give "ab ."
 114
 115 - set a meaningful default if no `Bl' list type is assigned
 116
 117 - have a blank `It' head for `Bl -tag' not puke
 118
 119 - prohibit `Nm' from having non-text HEAD children
 120   (e.g., NetBSD mDNSShared/dns-sd.1)
 121   (mdoc_html.c and mdoc_term.c `Nm' handlers can be slightly simplified)
 122
 123 - When there is free text in the SYNOPSIS and that free text contains
 124   the .Nm macro, groff somehow understands to treat the .Nm as an in-line
 125   macro, while mandoc treats it as a block macro and breaks the line.
 126   No idea how the logic for distinguishing in-line and block instances
 127   should be, needs investigation.
 128   uqs@  Thu, 2 Jun 2011 11:03:51 +0200
 129   uqs@  Thu, 2 Jun 2011 11:33:35 +0200
 130
 131 --- missing man features -----------------------------------------------
 132
 133 - groff an-ext.tmac macros (.UR, .UE) occur in xine(5)
 134   reported by brad@  Sat, 15 Jan 2011 15:45:23 -0500
 135
 136 - -T[x]html doesn't stipulate non-collapsing spaces in literal mode
 137
 138 --- missing tbl features -----------------------------------------------
 139
 140 - implement basic non-parametric .de to support e.g. sox(1)
 141   reported by naddy@ Sat, 16 Oct 2010 23:51:57 +0200
 142   *** sox(1) still doesn't work, tbl(1) errors need investigation
 143
 144 - allow standalone `.' to be interpreted as an end-of-layout
 145   delimiter instead of being thrown away as a no-op roff line
 146   reported by Yuri Pankov, Wed 18 May 2011 11:34:59 CEST
 147
 148 --- missing misc features ----------------------------------------------
 149
 150 - clean up escape sequence handling, creating three classes:
 151   (1) fully implemented, or parsed and ignored without loss of content
 152   (2) unimplemented, potentially causing loss of content
 153       or serious mangling of formatting (e.g. \n) -> ERROR
 154       see textproc/mgdiff(1) for nice examples
 155   (3) undefined, just output the character -> perhaps WARNING
 156
 157 - The \t escape sequence is the same as a literal tab, see for example
 158   the ASCII table in hexdump(1) where
 159     .Bl -column \&000_nu \&001_so \&002_st \&003_et \&004_eo
 160     .It \&000\ nul\t001\ soh\t002\ stx\t003\ etx\t004\ eot\t005\ enq
 161   produces
 162     000 nul  001 soh  002 stx  003 etx  004 eot  005 enq
 163   and the example in oldrdist(1)
 164
 165 - look at pages generated from reStructeredText, e.g. devel/mercurial hg(1)
 166   These are a weird mixture of man(7) and custom autogenerated low-level
 167   roff stuff.  Figure out to what extent we can cope.
 168   For details, see http://docutils.sourceforge.net/rst.html
 169   noted by stsp@  Sat, 24 Apr 2010 09:17:55 +0200
 170   reminded by nicm@  Mon, 3 May 2010 09:52:41 +0100
 171
 172 - check compatibility with Plan9:
 173   http://swtch.com/usr/local/plan9/tmac/tmac.an
 174   http://swtch.com/plan9port/man/man7/man.html
 175   "Anthony J. Bentley" <anthonyjbentley@gmail.com> 28 Dec 2010 21:58:40 -0700
 176
 177 ************************************************************************
 178 * formatting issues: ugly output
 179 ************************************************************************
 180
 181 - a column list with blank `Ta' cells triggers a spurrious
 182   start-with-whitespace printing of a newline
 183
 184 - double quotes inside double quotes are escaped by doubling them
 185   implement this in mdoc(7), too
 186   so far, we only have it in roff(7) and man(7)
 187   reminded by millert@  Thu, 09 Dec 2010 17:29:52 -0500
 188
 189 - perl(1) SYNOPSIS looks bad; reported by deraadt@
 190   1) man(7) seems to need SYNOPSIS .Nm blocks, too
 191
 192 - In .Bl -column,
 193   .It Em Authentication<tab>Key Length
 194   ought to render "Key Length" with emphasis, too,
 195   see OpenBSD iked.conf(5).
 196   reported again Nicolas Joly via wiz@ Wed, 12 Oct 2011 00:20:00 +0200
 197
 198 - empty phrases in .Bl column produce too few blanks
 199   try e.g. .Bl -column It Ta Ta
 200   reported by millert Fri, 02 Apr 2010 16:13:46 -0400
 201
 202 - .%T can have trailing punctuation.  Currently, it puts the trailing
 203   punctuation into a trailing MDOC_TEXT element inside its own scope.
 204   That element should rather be outside its scope, such that the
 205   punctuation does not get underlines.  This is not trivial to
 206   implement because .%T then needs some features of in_line_eoln() -
 207   slurp all arguments into one single text element - and one feature
 208   of in_line() - put trailing punctuation out of scope.
 209   Found in mount_nfs(8) and exports(5), search for "Appendix".
 210
 211 - in enclosures, mandoc sometimes fancies a bogus end of sentence
 212   reminded by jmc@  Thu, 23 Sep 2010 18:13:39 +0059
 213
 214 ************************************************************************
 215 * formatting issues: gratuitous differences
 216 ************************************************************************
 217
 218 - .Rv (and probably .Ex) print different text if an `Nm' has been named
 219   or not (run a manual without `Nm blah' to see this).  I'm not sure
 220   that this exists in the wild, but it's still an error.
 221
 222 - In .Bl -bullet, the groff bullet is "+\b+\bo\bo", the mandoc bullet
 223   is just "o\bo".
 224   see for example OpenBSD ksh(1)
 225
 226 - The characters "|" and "\*(Ba" should never be bold,
 227   not even in the middle of a word, e.g. ".Cm b\*(Bac" in
 228   "mknod [-m mode] name b|c major minor"
 229   in OpenBSD ksh(1)
 230
 231 - A bogus .Pp between two .It must not produce a double blank line,
 232   see between -R and -r in OpenBSD rm(1), before "update" in mount(8),
 233   or in DIAGNOSTICS in init(8), or before "is always true" in ksh(1).
 234   The same happens with .Pp just before .El, see bgpd.conf(5).
 235   Also have `It' complain if `Pp' is invoked at certain times (not
 236   -compact?).
 237
 238 - .Pp between two .It in .Bl -column should produce one,
 239   not two blank lines, see e.g. login.conf(5).
 240   reported by jmc@  Sun, 17 Apr 2011 14:04:58 +0059
 241
 242 - If the *first* line after .It is .Pp, break the line right after
 243   the tag, do not pad with space characters before breaking.
 244   See the description of the a, c, and i commands in sed(1).
 245
 246 - If the first line after .It is .D1, do not assert a blank line
 247   in between, see for example tmux(1).
 248   reported by nicm@  13 Jan 2011 00:18:57 +0000
 249
 250 - .Nx 1.0a
 251   should be "NetBSD 1.0A", not "NetBSD 1.0a",
 252   see OpenBSD ccdconfig(8).
 253
 254 - In .Bl -tag, if a tag exceeds the right margin and must be continued
 255   on the next line, it must be indented by -width, not width+1;
 256   see "rule block|pass" in OpenBSD ifconfig(8).
 257
 258 - When the -width string contains macros, the macros must be rendered
 259   before measuring the width, for example
 260     .Bl -tag -width ".Dv message"
 261   in magic(5), located in src/usr.bin/file, is the same
 262   as -width 7n, not -width 11n.
 263
 264 - The \& zero-width character counts as output.
 265   That is, when it is alone on a line between two .Pp,
 266   we want three blank lines, not two as in mandoc.
 267
 268 - When .Fn arguments exceed one output line, all but the first
 269   should be indented, see e.g. rpc(3);
 270   reported by jmc@ on discuss@  Fri, 29 Oct 2010 13:48:33 +0100
 271   reported again by Nicolas Joly via wiz@  Sun, 18 Sep 2011 18:24:40 +0200
 272   Also, we don't want to break the line within the argument of:
 273   .Fa "chtype tl"
 274
 275 - .Ns should work when called at the end of an input line, see
 276   the following code in vi(1):
 277     .It Xo
 278     .Op Ar line
 279     .Cm a Ns Op Cm ppend Ns
 280     .Op Cm !\&
 281     .Xc
 282     The input text is appended after the specified line.
 283
 284 - Header lines of excessive length:
 285   Port OpenBSD man_term.c rev. 1.25 to mdoc_term.c
 286   and document it in mdoc(7) and man(7) COMPATIBILITY
 287   found while talking to Chris Bennett
 288
 289 - In man(7), the sequence
 290     .HP
 291     one line of regular text
 292     .SH
 293   should not produce two blank lines before the .SH,
 294   see for example named-checkconf(8).
 295
 296 - In man(7), the sequence
 297     .SH HEADER
 298     <blank line>
 299     .PP
 300     regular text
 301   should not produce any blank lines between the header and the text,
 302   see for example rsync(1).
 303   Reported by naddy@  Mon, 28 Mar 2011 20:45:42 +0200
 304
 305 - In man(7), the sequence
 306     regular text
 307     .IP
 308     .IP "tag"
 309     indented text
 310   should produce one, not four blank lines between the regular text
 311   and the tag, see for example rsync(1).
 312   Likewise,
 313     regular text
 314     .IP
 315     indented text
 316   should produce one, not two blank lines in between, and
 317     regular text
 318     .IP
 319     .RS
 320     .IP tag
 321     indented text
 322   should produce one, not three blank lines.
 323   Reported by naddy@  Mon, 28 Mar 2011 20:45:42 +0200
 324
 325 - trailing whitespace must be ignored even when followed by a font escape,
 326   see for example
 327     makes
 328     \fBdig \fR
 329     operate in batch mode
 330   in dig(1).
 331
 332 ************************************************************************
 333 * error reporting issues
 334 ************************************************************************
 335
 336 ************************************************************************
 337 * performance issues
 338 ************************************************************************
 339
 340 Several areas can be cleaned up to make mandoc even faster.  These are
 341
 342 - improve hashing mechanism for macros (quite important: performance)
 343
 344 - improve hashing mechanism for characters (not as important)
 345
 346 - the PDF file is HUGE: this can be reduced by using relative offsets
 347
 348 - instead of re-initialising the roff predefined-strings set before each
 349   parse, create a read-only version the first time and copy it
 350
 351 ************************************************************************
 352 * structural issues
 353 ************************************************************************
 354
 355 - We use the input line number at several places to distinguish
 356   same-line from different-line input.  That plainly doesn't work
 357   with user-defined macros, leading to random breakage.
 358
 359 - Find better ways to prevent endless loops
 360   in roff(7) macro and string expansion.
 361
 362 - Finish cleanup of date handling.
 363   Decide which formats should be recognized where.
 364   Update both mdoc(7) and man(7) documentation.
 365   Triggered by  Tim van der Molen  Tue, 22 Feb 2011 20:30:45 +0100