BUGS

   1 BUGS
   2
   3 This file contains bugs we haven't fixed yet, and some we have.
   4 The format is very simple (one bug per page):
   5
   6 #   line                                   regexp
   7
   8 1   bug NUMBER -- ONE-LINE-DESCRIPTION     ^bug ([0-9]+) -- (.+)$
   9 2   reported-by: SOMEONE / DATE            ^reported-by: ([^/]) */ *(.+)$
  10 3   fixed: DATE (for guile VERSION)        ^fixed: (.+) .for guile (.+).$
  11 4+  OTHER-NOTES
  12
  13 DATE is formatted YYYY-MM-DD, or "not-yet".  Line numbering is zero-origin;
  14 line 0 is empty (newline always follows newpage (recognizable w/ "\f\n")).
  15 The suggested regexps can be used in a line-oriented parser.
  16
  17 \f
  18 bug 0 -- no BUGS file
  19 reported-by: ttn / 2001-09-25
  20 fixed: 2001-09-30 (1.5.x, 1.7.x)
  21
  22 bugs are good.
  23
  24 \f
  25 bug 1 -- (debug-disable 'debug) has no effect if first form in ~/.guile
  26 reported-by: Sven.Hartrumpf@fernuni-hagen.de / 2001-08-31
  27 fixed: not-yet
  28
  29 ttn sez:
  30 buggy behavior confirmed on both 1.5.x and 1.7.x.
  31
  32 on the other hand, when second or later, the expected effect (disabling use of
  33 debugging evaluator) seems to be reflected when viewing the debug options.  it
  34 could be that the bug is in display and not the actual setting.
  35
  36 \f
  37 bug 2 -- documentation mirroring not complete
  38 reported-by: lord@regexps.com / 2001-08
  39 fixed: not-yet
  40
  41 ttn sez:
  42 uncollated mirroring is already available:
  43 http://savannah.gnu.org/cgi-bin/viewcvs/guile/
  44 http://mail.gnu.org/pipermail/guile-devel/
  45
  46 \f
  47 bug 3 -- design decisions not recorded
  48 reported-by: lord@regexps.com / 2001-08
  49 fixed: not-yet
  50
  51 changes are made w/ little documentation on their motivation, design and
  52 consequences.  past initiatives and current maintainer discipline to support
  53 them need shoring up.
  54
  55 ttn sez:
  56 someone also mentioned that process and product are orthogonal; nastiness can
  57 result in/from either, and for different reasons.  i tend to believe this up
  58 until a certain point, when the process becomes self-aware and includes its own
  59 improvement in its considerations.  then, hopefully it becomes light.
  60
  61 to change people is difficult, but that's why we have guile...
  62
  63 \f
  64 bug 4 -- named-let transform evaluates initvalue in the wrong scope
  65 reported-by: agj@alum.mit.edu / 2001-09-21
  66 fixed: not-yet
  67
  68 currently:
  69 > (let name ((var init) ...) body ...)
  70 >
  71 > to
  72 >
  73 > (letrec ((name (lambda (var ...) body ...)))
  74 >   (name init ...))
  75
  76 should be:
  77 > ((letrec ((name (lambda (var ...) body ...))) name) init ...)
  78
  79 Dirk Herrmann sez:
  80 I will do it - no big deal (I hope).  But, I won't have time to do it
  81 before next week.  What about adding this code as a test case to the test
  82 suite?  Until it is fixed, it can be tagged as expected to fail.
  83
  84 \f
  85 bug 5 -- uniform vector read-hash "e" clashes w/ exactness specification
  86 reported-by: kwright@gis.net / 2001-10-07
  87 fixed: 2001-10-14 (1.5.x, 1.7.x)
  88
  89 Keith Wright sez:
  90 > Before getting into philosophy, I want to mention that while poking
  91 > around and rethinking I found the following Bug.
  92 >
  93 >     guile> (version)
  94 >     "1.5.2"
  95 >     guile> #E3
  96 >     3
  97 >     guile> #e3
  98 >     ERROR: read:uniform-vector list not found
  99 >     ABORT: (misc-error)
 100 >     guile> 3
 101 >
 102 > This violates both R5RS 7.1 ``Case is insignificant'' and
 103 >
 104 > > From: guile-1.5.2/doc/ref/guile.info-3 Line 568
 105 > >
 106 > >     The codes for indicating exactness (which can, incidentally, be
 107 > >  applied to all numerical values) are:
 108 > >
 109 > >   * `#e', `#E' -- the number is exact
 110
 111 2001-10-14  Mikael Djurfeldt  <mdj@linnaeus>
 112
 113         * read.c (scm_lreadr): When user-defined hash procedure returns
 114         SCM_UNSPECIFIED: Fall back to standard handling instead of raising
 115         an exception.  (This prevents parsing of uniform vectors from
 116         interfering with parsing of numbers.)
 117
 118 \f
 119 bug 6 -- exactness prefix doesn't do nothing gracefully
 120 reported-by: kwright@gis.net / 2001-10-09
 121 fixed: 2001-10-14 (1.5.x, 1.7.x) (see bug 5)
 122
 123 [this may be related to bug 5. --ttn]
 124
 125 Keith Wright sez:
 126 > The saga continues...
 127 >
 128 > > >     The codes for indicating exactness (which can, incidentally, be
 129 > > >  applied to all numerical values) are:
 130 > > >
 131 > > >   * `#e', `#E' -- the number is exact
 132 >
 133 >   guile> #E3  ==>  3
 134 >   guile> #E3.0
 135 >   ERROR: In procedure scm_lreadr:
 136 >   ERROR: unknown # object
 137 >
 138 > Better change the reference manual s/, incidentally,/'t/.
 139 >
 140 > (That's a joke son.  The Scheme language requires #e in front
 141 > of any number to at least do nothing successfully.)
 142
 143 \f
 144 bug 7 -- readline configuration anomolies not handled gracefully
 145 reported-by: kwright@gis.net / 2001-10-17
 146 fixed: not-yet
 147
 148 \f
 149 bug 8 -- install requires makeinfo
 150 reported-by: kwright@gis.net / 2001-10-17
 151 fixed: not-yet
 152
 153 Keith Wright sez:
 154 > I got guile-1.5.4 and installed it on one machine with no problem.
 155 > On another I did ./configure;make as before and it died with:
 156 >
 157 > > ...
 158 > > on.doc vports.doc weaks.doc        symbols-deprecated.doc regex-posix.doc ramap.doc unif.doc filesys.doc posix.doc net_db.doc socket.doc | ./guile-snarf-docs-texi > guile.texi \
 159 > > || { rm guile.texi; false; }
 160 > > rm -f guile-procedures.txt
 161 > > makeinfo --force -o guile-procedures.txt guile.texi || test -f guile-procedures.txt
 162 > > /bin/sh: makeinfo: command not found
 163 > > make[1]: *** [guile-procedures.txt] Error 1
 164 > > make[1]: Leaving directory `/home/kwright/build/guile-1.5.4/libguile'
 165 > > make: *** [all-recursive] Error 1
 166 >
 167 > I get the impression that it is supposed to build without 'makeinfo',
 168 > if this is not so it should be mentioned in the README or INSTALL.
 169
 170 \f
 171 bug 9 -- sibling threads not infinite looping
 172 reported-by: giraud@tif.inria.fr / 2001-09-24
 173 fixed: not-yet
 174
 175 Manuel Giraud sez:
 176 > I've configure guile-1.5.4 '--with-threads' and '--with-modules' on a
 177 > "classical" Linux box and used this code that crash silently :
 178 >
 179 > (use-modules (ice-9 threads))
 180 >
 181 > (begin-thread
 182 >  (let loop ()
 183 >    (display "a")
 184 >    (newline)
 185 >    (loop)))
 186 >
 187 > (begin-thread
 188 >  (let loop ()
 189 >    (display "b")
 190 >    (newline)
 191 >    (loop)))
 192
 193 mdj sez:
 194 + 2001-10-06  Mikael Djurfeldt  <mdj@linnaeus>
 195 +
 196 +       These changes fixes a race condition in the Guile coop - pthread
 197 +       compatibility code.
 198 +
 199 +       * coop.c (mother_awake_p): New variable.
 200 +       (coop_create): Set mother_awake_p before creating or signalling
 201 +       mother; wait until mother is going to sleep before returning.
 202 +       (mother): Reset mother_awake_p before going to sleep.
 203 +
 204
 205 ttn sez:
 206 crash still occurs w/ "guile -s bug.scm" (cvs checkout on 2001/11/06).
 207 however, "guile -l bug.scm" works as expected.
 208
 209 \f
 210 bug 10 -- embedded libltdl masks system version unconditionally
 211 reported-by: grib@linuxdevel.com / 2001-11-04
 212 fixed: not-yet
 213
 214 Bill Gribble sez:
 215 > The guideline for when to use the convenience version of libltdl is
 216 > this:
 217 >
 218 > > In general you can safely use the convenience library in
 219 > > programs which don't depend on other libraries that might
 220 > > use libltdl too.
 221 >
 222 > So guile's use of the convenience library embodies an assumption about
 223 > programs which use guile: that they will not use libltdl except through
 224 > guile's wrapper (dynamic-link and friends).
 225 >
 226 > It seems like this is a bad assumption for guile to make, since it is
 227 > intended to be used as an extension language for many different kinds of
 228 > programs and not to put a priori restrictions on what they can do.
 229 > Especially when those restrictions are not documented.
 230 >
 231 > I would suggest that prior to the guile-1.6 release the use of the
 232 > convenience ltdl library be removed, instead depending on an installed
 233 > version of libltdl.
 234
 235 \f
 236 bug 11 -- (ice-9 optargs) #:rest arg polluted by keys/values
 237 reported-by: ttn / 2001-11-09
 238 fixed: not-yet
 239
 240 ttn sez:
 241 > the following code displays
 242 >
 243 >   (#<procedure good x> ())
 244 >   (#<procedure good x> (1))
 245 >   (#<procedure good x> ((1 2 3)))
 246 >   (#<procedure blah x> JAY KAY (#:j JAY #:k KAY))
 247 >   (#<procedure blah x> JAY KAY (#:j JAY #:k KAY 1))
 248 >   (#<procedure blah x> JAY KAY (#:j JAY #:k KAY (1 2 3)))
 249 >
 250 > but `x' should be the same for both `good' and `blah':
 251 >
 252 >   (use-modules (ice-9 optargs) (ice-9 rdelim))
 253 >
 254 >   (define* (good #:rest x)
 255 >     (write-line (list good x)))
 256 >
 257 >   (good)
 258 >   (good 1)
 259 >   (good '(1 2 3))
 260 >
 261 >   (define* (blah #:key (j 'j) (k 'k) #:rest x)
 262 >     (write-line (list blah j k x)))
 263 >
 264 >   (blah #:j 'JAY #:k 'KAY)
 265 >   (blah #:j 'JAY #:k 'KAY 1)
 266 >   (blah #:j 'JAY #:k 'KAY '(1 2 3))
 267
 268 \f
 269 [BUGS ends here]