[Axiom-developer] _*STANDARD_-OUTPUT_* and similar

[Mike Thomas]

Hi Tim.

I spent some time this morning on Windows Axiom with particular reference to
the sometimes on, sometimes off problem with compiling AHYP.spad:

------------------------------------------------------------------------
   initializing NRLIB AHYP for ArcHyperbolicFunctionCategory
The syntax of the command is incorrect.
Error: equal: x is a NULL pointer
Fast links are on: do (si::use-fast-links nil) for debugging
Error signalled by RETURN.
Broken at APPLY.  Type :H for Help.
BOOT>>
------------------------------------------------------------------------

There are a couple of problems here:

1. interpsys thinks that the syntax of some command or other is incorrect,
and

2. the error system ends up with a NULL pointer in the GCL C code.

I managed to isolate problem 2 to a simple case: evaluating 2*y at the
interpsys prompt, which in my incomplete system leads immediately to an
attempt to load the as yet uncompiled polynomial domain code in POLY.o,
hence an error and then the NULL pointer.

I then found that in "g-error.boot.pamphlet" the function  "sayErrorly()"
calls "saturnSayErrorly()" which contains the output stream variable
_*STANDARD_-OUTPUT_*.

Altering that line as follows restores order to the error reporting:

---------------------------------------------------------------------
$ cvs diff src/interp/g-error.boot.pamphlet
Index: src/interp/g-error.boot.pamphlet
===================================================================
RCS file: /cvsroot/axiom/axiom/src/interp/g-error.boot.pamphlet,v
retrieving revision 1.3
diff -r1.3 g-error.boot.pamphlet
175c175
<   _*STANDARD_-OUTPUT_* : fluid := $texOutputStream
---
>   *STANDARD-OUTPUT* : fluid := $texOutputStream
---------------------------------------------------------------------

However I then found that similar strange variations on various IO related
variables and functions exist in the interpreter code as shown further down.

Are these constructs intentional (perhaps some kind of TeXism which my TeX
system is not handling properly for example) or an unfortunate editing
error?

Cheers

Mike Thomas.


$ fgrep -r -i "_*STANDARD" c:/cvs/head/axiom/src/interp
c:/cvs/head/axiom/src/interp/compat.boot.pamphlet:
read_-line(_*STANDARD_-INPUT
_*)
c:/cvs/head/axiom/src/interp/int-top.boot.pamphlet:    a:=
serverReadLine(_*STAN
DARD_-INPUT_*)
c:/cvs/head/axiom/src/interp/msgdb.boot.pamphlet:  _*STANDARD_-OUTPUT_* :
fluid
:= $texOutputStream
c:/cvs/head/axiom/src/interp/msgdb.boot.pamphlet:  _*STANDARD_-OUTPUT_* :
fluid
:= $texOutputStream
c:/cvs/head/axiom/src/interp/msgdb.boot.pamphlet:  _*STANDARD_-OUTPUT_* :
fluid
:= $texOutputStream

address@hidden /c/cvs/head/fptools
$ fgrep -r -i "read_-line" c:/cvs/head/axiom/src/interp
c:/cvs/head/axiom/src/interp/compat.boot.pamphlet:  s => read_-line(first s)
c:/cvs/head/axiom/src/interp/compat.boot.pamphlet:
read_-line(_*STANDARD_-INPUT
_*)
c:/cvs/head/axiom/src/interp/cstream.boot.pamphlet:        a:=shoeread_-line
s
c:/cvs/head/axiom/src/interp/database.boot.pamphlet:--++  while (not PLACEP
(x:=
 READ_-LINE stream)) repeat
c:/cvs/head/axiom/src/interp/g-error.boot.pamphlet:  READ_-LINE
_*TERMINAL_-IO_*

c:/cvs/head/axiom/src/interp/i-syscmd.boot.pamphlet:    line :=
read_-line(files
tream,false)
c:/cvs/head/axiom/src/interp/msgdb.boot.pamphlet:  ans := READ_-LINE
conStream
c:/cvs/head/axiom/src/interp/server.boot.pamphlet:    READ_-LINE(stream)
c:/cvs/head/axiom/src/interp/server.boot.pamphlet:      l :=
READ_-LINE(stream)
c:/cvs/head/axiom/src/interp/server.boot.pamphlet:        parseAndInterpret
READ
_-LINE(CURINSTREAM) )))
c:/cvs/head/axiom/src/interp/word.boot.pamphlet:    while (not PLACEP (x:=
READ_
-LINE stream)) repeat