TODO

   1 ************************************************************************
   2 * Official mandoc TODO.
   3 * $Id: TODO,v 1.89 2011/02/03 10:24:21 kristaps Exp $
   4 ************************************************************************
   5
   6 ************************************************************************
   7 * parser bugs
   8 ************************************************************************
   9
  10 - It should be noted that in mandoc, the TBL macros are interpreted as
  11   real roff macros (e.g., `.  TS' is ok), while the troff preprocessor
  12   only looks for `.TS' (static string).  This should be noted in the
  13   COMPATIBILITY section.  The same applies for nascent `.EQ' (as it is
  14   now, `.EN' is processed as a static string).
  15
  16 - .TP before .SH is still FATAL in man(7)
  17   reported by brad@  Sat, 15 Jan 2011 15:54:54 -0500
  18
  19 - The mdoc(7) parser should handle `\.' as punctuation:
  20     .Ql .if ${VAR} op something \.
  21   should produce
  22     `.if ${VAR} op something'.
  23   not
  24     `.if ${VAR} op something .'
  25   as seen in make(1).
  26   To escape the dot, one would use `\&.'.
  27
  28 - the roff parser doesn't tolerate additional characters between
  29   a macro and the \} terminating a conditional block, e.g.
  30   .if n \{
  31   .br \}
  32   reported by ulrich spoerlein  Tue, 19 Oct 2010 20:39:50 +0200
  33
  34 ************************************************************************
  35 * formatter bugs
  36 ************************************************************************
  37
  38 ************************************************************************
  39 * missing features
  40 ************************************************************************
  41
  42 --- missing roff features ----------------------------------------------
  43
  44 - .ad (adjust margins)
  45   .ad l -- adjust left margin only (flush left)
  46   .ad r -- adjust right margin only (flush right)
  47   .ad c -- center text on line
  48   .ad b -- adjust both margins (alias: .ad n)
  49   .na   -- temporarily disable adjustment without changing the mode
  50   .ad   -- re-enable adjustment without changing the mode
  51   Adjustment mode is ignored while in no-fill mode (.nf).
  52
  53 - .it (line traps) occur in mysql(1), yasm_arch(7)
  54   generated by DocBook XSL Stylesheets v1.71.1 <http://docbook.sf.net/>
  55   reported by brad@  Sat, 15 Jan 2011 15:48:18 -0500
  56
  57 - .ns (no-space mode) occurs in xine-config(1)
  58   reported by brad@  Sat, 15 Jan 2011 15:45:23 -0500
  59
  60 - xloadimage(1) wants .ti (temporary indent), rep by naddy@
  61
  62 - .ta (tab settings) occurs in ircbug(1) and probably gnats(1)
  63   reported by brad@  Sat, 15 Jan 2011 15:50:51 -0500
  64
  65 - pod2man expects `tr' to be implemented for \*(-- to work
  66
  67 - \c (interrupted text) occurs in chat(8)
  68
  69 --- missing mdoc features ----------------------------------------------
  70
  71 - fix bad block nesting involving multiple identical explicit blocks
  72   see the OpenBSD mdoc_macro.c 1.47 commit message
  73
  74 - .Bl -column .Xo support is missing
  75   ultimate goal:
  76   restore .Xr and .Dv to
  77   lib/libc/compat-43/sigvec.3
  78   lib/libc/gen/signal.3
  79   lib/libc/sys/sigaction.2
  80
  81 - edge case: decide how to deal with blk_full bad nesting, e.g.
  82   .Sh .Nm .Bk .Nm .Ek .Sh found by jmc@ in ssh-keygen(1)
  83   from jmc@  Wed, 14 Jul 2010 18:10:32 +0100
  84
  85 - \\ is now implemented correctly
  86   * when defining strings and macros using .ds and .de
  87   * when parsing roff(7) and man(7) macro arguments
  88   It does not yet work in mdoc(7) macro arguments
  89   because libmdoc does not yet use mandoc_getarg().
  90   Also check what happens in plain text, it must be identical to \e.
  91
  92 - .Bd -filled should not be the same as .Bd -ragged, but align both
  93   the left and right margin.  In groff, it is implemented in terms
  94   of .ad b, which we don't have either.  Found in cksum(1).
  95
  96 - implement blank `Bl -column', such as
  97   .Bl -column
  98   .It foo Ta bar
  99   .El
 100
 101 - explicitly disallow nested `Bl -column', which would clobber internal
 102   flags defined for struct mdoc_macro
 103
 104 - inside `.Bl -column' phrases, punctuation is handled like normal
 105   text, e.g. `.Bl -column .It Fl x . Ta ...' should give "-x -."
 106
 107 - inside `.Bl -column' phrases, TERMP_IGNDELIM handling by `Pf'
 108   is not safe, e.g. `.Bl -column .It Pf a b .' gives "ab."
 109   but should give "ab ."
 110
 111 - set a meaningful default if no `Bl' list type is assigned
 112
 113 - have a blank `It' head for `Bl -tag' not puke
 114
 115 - prohibit `Nm' from having non-text HEAD children
 116   (e.g., NetBSD mDNSShared/dns-sd.1)
 117   (mdoc_html.c and mdoc_term.c `Nm' handlers can be slightly simplified)
 118
 119 --- missing man features -----------------------------------------------
 120
 121 - bashbug(1) complains "line scope broken" after
 122   .SM
 123   .B something
 124   should either just work or be a warning
 125   reported by naddy@
 126
 127 - groff an-ext.tmac macros (.UR, .UE) occur in xine(5)
 128   reported by brad@  Sat, 15 Jan 2011 15:45:23 -0500
 129
 130 --- missing tbl features -----------------------------------------------
 131
 132 - implement basic non-parametric .de to support e.g. sox(1)
 133   reported by naddy@ Sat, 16 Oct 2010 23:51:57 +0200
 134   *** sox(1) still doesn't work, tbl(1) errors need investigation
 135
 136 --- missing misc features ----------------------------------------------
 137
 138 - clean up escape sequence handling, creating three classes:
 139   (1) fully implemented, or parsed and ignored without loss of content
 140   (2) unimplemented, potentially causing loss of content
 141       or serious mangling of formatting (e.g. \n) -> ERROR
 142       see textproc/mgdiff(1) for nice examples
 143   (3) undefined, just output the character -> perhaps WARNING
 144
 145 - The \t escape sequence is the same as a literal tab, see for example
 146   the ASCII table in hexdump(1) where
 147     .Bl -column \&000_nu \&001_so \&002_st \&003_et \&004_eo
 148     .It \&000\ nul\t001\ soh\t002\ stx\t003\ etx\t004\ eot\t005\ enq
 149   produces
 150     000 nul  001 soh  002 stx  003 etx  004 eot  005 enq
 151   and the example in oldrdist(1)
 152
 153 - look at pages generated from reStructeredText, e.g. devel/mercurial hg(1)
 154   These are a weird mixture of man(7) and custom autogenerated low-level
 155   roff stuff.  Figure out to what extent we can cope.
 156   For details, see http://docutils.sourceforge.net/rst.html
 157   noted by stsp@  Sat, 24 Apr 2010 09:17:55 +0200
 158   reminded by nicm@  Mon, 3 May 2010 09:52:41 +0100
 159
 160 - check compatibility with Plan9:
 161   http://swtch.com/usr/local/plan9/tmac/tmac.an
 162   http://swtch.com/plan9port/man/man7/man.html
 163   "Anthony J. Bentley" <anthonyjbentley@gmail.com> 28 Dec 2010 21:58:40 -0700
 164
 165 ************************************************************************
 166 * formatting issues: ugly output
 167 ************************************************************************
 168
 169 - a column list with blank `Ta' cells triggers a spurrious
 170   start-with-whitespace printing of a newline
 171
 172 - double quotes inside double quotes are escaped by doubling them
 173   implement this in mdoc(7), too
 174   so far, we only have it in roff(7) and man(7)
 175   reminded by millert@  Thu, 09 Dec 2010 17:29:52 -0500
 176
 177 - perl(1) SYNOPSIS looks bad; reported by deraadt@
 178   1) man(7) seems to need SYNOPSIS .Nm blocks, too
 179
 180 - In .Bl -column,
 181   .It Em Authentication<tab>Key Length
 182   ought to render "Key Length" with emphasis, too,
 183   see OpenBSD iked.conf(5).
 184
 185 - empty phrases in .Bl column produce too few blanks
 186   try e.g. .Bl -column It Ta Ta
 187   reported by millert Fri, 02 Apr 2010 16:13:46 -0400
 188
 189 - .%T can have trailing punctuation.  Currently, it puts the trailing
 190   punctuation into a trailing MDOC_TEXT element inside its own scope.
 191   That element should rather be outside its scope, such that the
 192   punctuation does not get underlines.  This is not trivial to
 193   implement because .%T then needs some features of in_line_eoln() -
 194   slurp all arguments into one single text element - and one feature
 195   of in_line() - put trailing punctuation out of scope.
 196   Found in mount_nfs(8) and exports(5), search for "Appendix".
 197
 198 - in enclosures, mandoc sometimes fancies a bogus end of sentence
 199   reminded by jmc@  Thu, 23 Sep 2010 18:13:39 +0059
 200
 201 ************************************************************************
 202 * formatting issues: gratuitious differences
 203 ************************************************************************
 204
 205 - .Rv (and probably .Ex) print different text if an `Nm' has been named
 206   or not (run a manual without `Nm blah' to see this).  I'm not sure
 207   that this exists in the wild, but it's still an error.
 208
 209 - In .Bl -bullet, the groff bullet is "+\b+\bo\bo", the mandoc bullet
 210   is just "o\bo".
 211   see for example OpenBSD ksh(1)
 212
 213 - The characters "|" and "\*(Ba" should never be bold,
 214   not even in the middle of a word, e.g. ".Cm b\*(Bac" in
 215   "mknod [-m mode] name b|c major minor"
 216   in OpenBSD ksh(1)
 217
 218 - A bogus .Pp between two .It must not produce a double blank line,
 219   see between -R and -r in OpenBSD rm(1), before "update" in mount(8),
 220   or in DIAGNOSTICS in init(8), or before "is always true" in ksh(1).
 221   The same happens with .Pp just before .El, see bgpd.conf(5).
 222   Also have `It' complain if `Pp' is invoked at certain times (not
 223   -compact?).
 224
 225 - If the *first* line after .It is .Pp, break the line right after
 226   the tag, do not pad with space characters before breaking.
 227   See the description of the a, c, and i commands in sed(1).
 228
 229 - If the first line after .It is .D1, do not assert a blank line
 230   in between, see for example tmux(1).
 231   reported by nicm@  13 Jan 2011 00:18:57 +0000
 232
 233 - .Nx 1.0a
 234   should be "NetBSD 1.0A", not "NetBSD 1.0a",
 235   see OpenBSD ccdconfig(8).
 236
 237 - In .Bl -tag, if a tag exceeds the right margin and must be continued
 238   on the next line, it must be indented by -width, not width+1;
 239   see "rule block|pass" in OpenBSD ifconfig(8).
 240
 241 - When the -width string contains macros, the macros must be rendered
 242   before measuring the width, for example
 243     .Bl -tag -width ".Dv message"
 244   in magic(5), located in src/usr.bin/file, is the same
 245   as -width 7n, not -width 11n.
 246
 247 - The \& zero-width character counts as output.
 248   That is, when it is alone on a line between two .Pp,
 249   we want three blank lines, not two as in mandoc.
 250
 251 - When .Fn arguments exceed one output line, all but the first
 252   should be indented, see e.g. rpc(3);
 253   reported by jmc@ on discuss@  Fri, 29 Oct 2010 13:48:33 +0100
 254
 255 - .Ns should work when called at the end of an input line, see
 256   the following code in vi(1):
 257     .It Xo
 258     .Op Ar line
 259     .Cm a Ns Op Cm ppend Ns
 260     .Op Cm !\&
 261     .Xc
 262     The input text is appended after the specified line.
 263
 264 - Header lines of excessive length:
 265   Port OpenBSD man_term.c rev. 1.25 to mdoc_term.c
 266   and document it in mdoc(7) and man(7) COMPATIBILITY
 267   found while talking to Chris Bennett
 268
 269 - In man(7), the sequence
 270     .HP
 271     one line of regular text
 272     .SH
 273   should not produce two blank lines before the .SH,
 274   see for example named-checkconf(8).
 275
 276 - In man(7), the sequence
 277     regular text
 278     .RS
 279     .IP
 280   should have a blank line between the text and the beginning of the
 281   indented paragraph, see for example sudo(1).
 282
 283 - At least sometimes, .sp is ignored right after .SH; see named.conf(5).
 284
 285 - trailing whitespace must be ignored even when followed by a font escape,
 286   see for example
 287     makes
 288     \fBdig \fR
 289     operate in batch mode
 290   in dig(1).
 291
 292 ************************************************************************
 293 * error reporting issues
 294 ************************************************************************
 295
 296 ************************************************************************
 297 * performance issues
 298 ************************************************************************
 299
 300 Several areas can be cleaned up to make mandoc even faster.  These are
 301
 302 - improve hashing mechanism for macros (quite important: performance)
 303
 304 - improve hashing mechanism for characters (not as important)
 305
 306 - the PDF file is HUGE: this can be reduced by using relative offsets
 307
 308 ************************************************************************
 309 * structural issues
 310 ************************************************************************
 311
 312 - Find better ways to prevent endless loops
 313   in roff(7) macro and string expansion.
 314