> 
> On Thu, Mar 1, 2018 at 11:47 PM, Waldek Hebisch
> <hebi...@math.uni.wroc.pl> wrote:
> > oldk1331 wrote:
> >>
> >> Hi Waldek, you haven't commented on the following patches.
> >> Also, please merge the documentation fix patches as long as
> >> you see fit.
> >>
> >> > fix display of 'box':
> >> >
> >> > https://github.com/oldk1331/fricas/commit/2c02ad4ea19d7869327ed708a5e8ff93fd6bfc29.patch
> >
> > This is a workaround, but AFAICS OutputForm is correct and we should
> > render AGGLST in such a way that multiplication does not vanish.
> 
> Currently 'box' only accepts one argument, so there is no point to
> wrap it inside AGGLST.

1) AGGLST is a legal output form so it should render correctly.
2) Wrapping or not 'box' in AGGLST is a semantic decision.  We should
   not make such decisions to work around rendering problems.

> 
> >> >
> >> > fix PostScript format image output:
> >> >
> >> > https://github.com/oldk1331/fricas/commit/7f0c3c5c8de3b28acfed3d0818af23af5e29398a.patch
> >
> > ATM I do not see what you are fixing.  Changing dynamic allocation
> > to fixed length arrays in C code looks like a step backwards.
> 
> You can not save image as PS format, this patch fixes that:
> 
> (1) -> write(draw(x,x=1..2),"/tmp/a","postscript")
>    Compiling function %B with type DoubleFloat -> DoubleFloat
>    Graph data being transmitted to the viewport manager...
>    FriCAS2D data being transmitted to the viewport manager...
> view2D: write2d.c:192: writeViewport: Assertion
> `fricas_sprintf_actual_size >= 0 && fricas_sprintf_actual_size <
> fricas_sprintf_buffer_size' failed.
> 
>    (1)  "/tmp/a"

OK, I see.
 
> >> >
> >> > trace.boot : call 'clearClams' after 'clearConstructorCaches':
> >> >
> >> > https://github.com/oldk1331/fricas/commit/1ae84818ee9fdfa98dfb6586907cd71abca8ffdf.patch
> >
> > It is not clear for me if we should do this.
> 
> In your reply to "[BUG] strange behavior caused by ')trace'" in 2017 Dec 7,
> you showed your support for this change.

Well, partially.  The real bug is in different place and the
patch just masks most obvious breakage.  The patch may
be part of full solution, but ATM we just loose a testcase and
create false confidence in users: currently tracing of Expression
like domains only works when domains are freshly reinitialized
after clear operations.

> >> > don't print equation number in TexFormat when $IOindex is NIL:
> >> >
> >> > https://github.com/oldk1331/fricas/commit/9ae65815db85f24a5b79113d30d69a021a309904.patch
> >
> > Again, it is not clear if we shoud do this.  More precisely, it
> > is not clear if $IOindex should be NIL, and if it is what it
> > means.
> >
> 
> You can set '$IOindex' to NIL to prevent equation numbering:
> 
> (1) -> )set message prompt plain
> -> 1
> 
>    (1)  1
>                                                         Type: PositiveInteger

Note that this uses variable '$inputPromptType'.

> -> )boot $IOindex := nil
> 
> (EVAL-WHEN (EVAL LOAD) (SETQ |$IOindex| NIL))
> Value = NIL
> -> 1
> 
>    1
>                                                         Type: PositiveInteger

So this is redundant.

-- 
                              Waldek Hebisch

-- 
You received this message because you are subscribed to the Google Groups 
"FriCAS - computer algebra system" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to fricas-devel+unsubscr...@googlegroups.com.
To post to this group, send email to fricas-devel@googlegroups.com.
Visit this group at https://groups.google.com/group/fricas-devel.
For more options, visit https://groups.google.com/d/optout.

Reply via email to