Re: [Texmacs-dev] rewrite-inactive

[Sam Liddicott]

I can't manage to insert an inactive* like this:

<assign|sa2|<macro|x|<inactive*|<arg|x>>>>

And I'm still working out how to strip the inactive* when it exists in the main tree enclosing the macro argument, but it looks like it is no longer there but already taken effect.

By hacky take-to-pieces macros like this:

<assign|showme|<macro|x|[<get-label|<arg|x>>(<length|<arg|x>>)=<arg|x>]>>

<assign|sa|<\xmacro|x>

  <assign|z|<look-up|<quasi|<tuple|<unquote*|<quote-arg|x>>>>|0>><showme|<look-up|<value|z>|0>><assign|zz|<look-up|<value|z>|2>>[<showme|<look-up|<value|zz>|1>>]<assign|zzz|<look-up|<value|zz>|1>>

  <showme|<look-up|<value|zzz>|1>>

</xmacro>>

I can confirm that:

<sa|<unquote*|...

passes the already re-written tree.

Maybe I can temporarily re-define inline-tag and such like so that it can be recovered.

darn: <with|inline-tag|<value|identity>|... caused CPU to reach 100%

Sam

On Fri, May 25, 2012 at 12:24 PM, Sam Liddicott <address@hidden> wrote:

Well... I leaned enough to do:

<with|mode|<quote|src>... and then crash texmacs again

./Edit/Interface/edit_interface.cpp:35: string MODE_LANGUAGE(string): Assertion `false' failed.

I got to that because this:

<sa|<inactive*|<assign|ll|<macro|x|<with|ornament-color|#eeeeee|ornament-sunny-color|#000000|ornament-shadow-color|#000000|ornament-borderx|2l|ornament-hpadding|2ex|ornament-vpadding|2ex|<ornament|<surround||<htab|0spc>|<indent-both|1cm|1cm|<arg|x>>>>>>>>>

when sa is defined as:

<assign|sa|<macro|x|<quasiquote|<compound|<merge|get-label|>|<unquote*|<arg|x>>>>>>

emits:

<compound|get-label|mode|<quote|src>|<assign|ll|<macro...

So I presume that <unquote* must be <with|mode|<quote|src>|...

What I'm trying to define is a macro which takes an argument and renders it in source view AND evaluates it.

Sam

On Fri, May 25, 2012 at 9:20 AM, Sam Liddicott <address@hidden> wrote:

It's failing for me because (!is_func (t[0], ARG)) is failing in env_exec.cpp

I don't understand what is_func is doing:

inline bool is_func (tree t, tree_label l) { return (t.rep->op==l) && (N(t)!=0); }

it seems to mean that ...

and the first argument to rewrite-inactive has to have children

But is suggests a relationship between the first argument, and ARG (whatever ARG is)

On Fri, May 25, 2012 at 8:46 AM, Sam Liddicott <address@hidden> wrote:

Does anyone have an example of rewrite-inactive called from a texmacs document which doesn't result in: "invalid rewrite-inactive" ?

It is mentioned here: http://www.texmacs.org/tmweb/manual/webman-styling.en.html

These all fail

<rewrite-inactive|<length|<tuple|1|2|3>>|once>

<assign|t|once><rewrite-inactive|<length|<tuple|1|2|3>>|<value|t>>

<assign|t|once><rewrite-inactive|<length|<tuple|1|2|3>>|t>

Thanks

Sam