BUGS known to infest THUD (at some point). Please send bug reports not covered in this file to . Include a recipe to reproduce the problem, including relevant files. The following is to remind the THUD maintainer what means what (ignorable): HEADER MEANING Id serial number of the bug Summary one-line description of the buggy behavior Status one of: unresolved, fix-applied, resolved Reported-Version released version of THUD that demonstrates this bug Reported-By either "ttn" or the someone's email address Reported-Date when reported (and hopefully entered into BUGS :-) Fix-Applied-Summary one-line description of the fix Fix-Applied-Date when status changed to fix-applied Fix-Applied-Version version at and after which no longer demonstrate the bug Resolved-Date when status changed to resolved Here is an example record: Id: 0 Summary: no BUGS file Status: resolved Reported-Version: 0.5 Reported-By: ttn Reported-Date: Fix-Applied-Summary: write a BUGS file Fix-Applied-Date: Fix-Applied-Version: 0.6 Resolved-Date: Self-knowledge requires attentive maintenance, dude! Id: 1 Summary: flop evaluation does not handle mutual dependencies properly Status: resolved Reported-By: ttn Fix-Applied-Summary: do two-stage evaluation Resolved-Date: 1998-08-08 Id: 2 Summary: zero QA for flop simulation Status: resolved Reported-By: ttn Fix-Applied-Summary: write some tests Resolved-Date: 1998-08-08 Now there is some QA, but not a lot. This bug is probably best re-expressed as "there is very little QA for flop simulation". Id: 3 Summary: no initialization to known state Status: resolved Reported-By: ttn Reported-Date: 1998-01-01 Fix-Applied-Summary: write `reset-internals' Fix-Applied-Date: 1998-08-08 Fix-Applied-Version: 0.9 Resolved-Date: 1998-08-08 The boot sequence calls `reset-internals'. Id: 4 Summary: permute does not jam inputs Status: resolved Reported-Version: 0.7 Reported-By: ttn Fix-Applied-Summary: rewrite `permute' Fix-Applied-Version: 0.8 Resolved-Date: 1998-08-08 Rewrote `permute' to use support for the `jam' eform property. Script samples/drivers/go now takes 45sec. :-/ Id: 5 Summary: hide-parens face not accepting input Status: resolved Reported-Version: 0.7 Fix-Applied-Summary: rewrite `hide-parens-face' Fix-Applied-Version: 0.8 Resolved-Date: 1998-08-08 Rewrote `hide-parens-face' to use GNU readline as well as `repl'. Id: 6 Summary: user script execution environment not well-defined Status: resolved Reported-Version: 0.7 Reported-By: ttn Fix-Applied-Summary: keep track of "current module" Fix-Applied-Version: 0.8 Resolved-Date: 1998-08-08 Users start in module `(thud user)' but can use `define-module' to move around. `thud-execute-user-script' keeps track of things using `save-module-excursion'. Id: 7 Summary: non-deterministic scheduling Status: unresolved Reported-Version: 0.7 Reported-By: ttn The solution to this is to investigate sched.scm... Id: 8 Summary: "help" not useful Status: resolved Reported-Version: 0.7 Reported-By: ttn Fix-Applied-Summary: write `help' Fix-Applied-Version: 0.8 Resolved-Date: 1998-08-08 The procedure `help' was rewritten to consult various places, usually the function docstring of public funcs as well as anything in the `(thud user)' module. Id: 9 Summary: guile-1.2 not supported Status: resolved Reported-Version: 0.8 Reported-By: ttn Fix-Applied-Summary: add some compatibility Fix-Applied-Date: 1998-08-08 Fix-Applied-Version: 0.9 Resolved-Date: 1998-08-08 In module `(thud boot)' a new variable `guile-version-1.2' flags the condition. In com.scm and face.scm, this var is used to add small compatibility workarounds (not so much a full compatibility layer). Id: 10 Summary: bit-slicing can only do "N:0" Status: unresolved Reported-By: ttn Bit-slicing is not sufficiently general. This exposes a weakness in the design in that bit-slicing introduces execution forms that are gross and inefficient. A redesign may be required, but let's hope not. This issue may never be resolved. Id: 11 Summary: loading THUD does not boot THUD Status: unresolved Reported-By: ttn Loading thud.in should be sufficient for running scripts. The script should be required to use `define-module', but there should be no other restrictions. (Currently, scripts must be loaded via "-x" or "x", which call `thud-execute-user-script'. This potentially makes it incompatible with persistent librarians like GUSH.) Id: 12 Summary: there is no way for TH code to jam signals Status: resolved Reported-Version: 0.11 Reported-By: ttn Fix-Applied-Summary: expose `permute'-style jamming to the user Fix-Applied-Date: 1999-07-26 Fix-Applied-Version: 0.16 Resolved-Date: 1999-07-26 The jamming support used by `permute' should be bubbled to the user. Id: 13 Summary: at the hide-parens prompt, "set!" by itself crashes THUD Status: resolved Reported-Version: 0.15 Reported-By: ttn Fix-Applied-Summary: get rid of hide-parens face Fix-Applied-Date: 1999-07-26 Fix-Applied-Version: 0.16 Resolved-Date: 1999-07-26 Hmm, face.scm needs to be formalized. In general, a comint and help module should be developed. Fixed: hide-parens-face was completely trashed in favor of terse-cmd-face. Id: 14 Summary: installation does not handle multiple guiles gracefully Status: unresolved Reported-By: ttn If there is a /usr/bin/guile and a /usr/local/bin/guile, the PATH is consulted to find the one used for @GUILE@ substitution. There should be an explicit configure option for this. Also, it would be a good idea to suggest de-installing old guile instances, perhaps in README. Id: 15 Summary: "thud -i homedir" produces garbled output Status: fix-applied Reported-Version: 0.14 Reported-By: ttn Fix-Applied-Summary: get rid of "thud -i homedir" Fix-Applied-Date: 2008-02-11 Fix-Applied-Version: 0.25 Both "homedir" and "hooddir" are obseleted by module catalogs. Id: 16 Summary: thud_4.html nested table is garbled Status: resolved Reported-Version: 0.14 Reported-By: ttn Fix-Applied-Summary: fix typo Fix-Applied-Date: 1999-06-06 Fix-Applied-Version: 0.15 Resolved-Date: 1999-06-06 Hmm, old makeinfo cannot handle leading spaces in the table definition (even if the table is nested, sigh). Perhaps newer ones can, now... Id: 17 Summary: at the hide-parens prompt, "q" crashes THUD Status: resolved Reported-Version: 0.15 Reported-By: ttn Fix-Applied-Summary: get rid of hide-parens face Fix-Applied-Date: 1999-07-26 Fix-Applied-Version: 0.16 Resolved-Date: 1999-07-26 Related to bug #13. Id: 18 Summary: using procedure `compile-to' causes an error Status: resolved Reported-Version: 0.15 Reported-By: ttn Fix-Applied-Summary: fix syntax error Fix-Applied-Date: 1999-07-26 Fix-Applied-Version: 0.16 Resolved-Date: 1999-07-26 The docstring had an extra double-quote character in it (syntax error, argh!!!). Id: 19 Summary: cannot do interactive testing Status: unresolved Reported-By: ttn Proabably a good idea to look into "greg" (guile regression) again, which has this capability built in. Until then, we cannot address bugs #12, #13, or #17, for example. Id: 20 Summary: `eval-print' only partially obeys `scope' Status: unresolved Reported-By: ttn Hmmm, what to say here? Id: 21 Summary: `deposit' does not obey `scope' at all Status: unresolved Reported-By: ttn `deposit' uses `eval-print'. Fixing #20 will fix this one. Id: 22 Summary: terse command parser does not actually consult proc formals Status: unresolved Reported-By: ttn This results in, eg, terse command "add" not accepting multiple args whereas that is possible from a scheme program. Id: 23 Summary: 1024-bit recursive adder overflows stack on link phase Status: unresolved Reported-By: ali.fawaz@tsqware.fr Reported-Date: 2001-04-06 BUGS database ends here.