On 8/15/2018 03:07, Hans Hagen wrote:
> On 8/15/2018 5:31 AM, Aditya Mahajan wrote:
>> On Tue, 14 Aug 2018, Rik Kabel wrote:
>>
>>> I really think, though, that a buffers.strip function would be the 
>>> way to go, instead of futzing after the damage is done.
>>
>> In lua, buffers.getcontent('name') gives the contents of the buffer 
>> labelled 'name'. You can post-process this in whatever way you want. 
>> For example, you could write a lua function userdata.strip that 
>> strips the leading and ending black spaces from a string, and then call
>>
>> context(userdata.strip(buffers.getcontent('name)))
> local s = [[
>
> foo
>
> bar
>
>
> bar
>
> ]]
>
> print(string.fullstrip(s))
>
>
So, it looks likethe following is closer to working for me in a simple 
test harness, but not in my full-blown work:

    \dontleavehmode{\llap{StripA~}}\ctxlua{context(utilities.strings.striplines(string.gsub(buffers.getcontent("TestBuffer"),"^\\relax",""),"prune"))}

(or similar with string.fullstrip).

However, it fails when the buffer starts with an unescaped left square 
bracket. That is probably why the \relax is inserted, and I can see that 
it is not a simple job to handle this. So, I will accept that I must 
either append [] or get rid of the open layout for these cases.

I suspect that the issue in the larger project has to do with quoting 
for the RE ("^\\relax") since compilation fails with:

    %% \stopAttribution ...getcontent("Attribution"),"^\\
    %% relax","")))}\stopparagrap...

Any pointers on such quoting would be appreciated.

-- 
Rik