mittelmark / pantcl Goto Github PK

I have the feeling that with 0.9.9 version, for code/sqlite evaluation, you always need :

• either eval=true
• or environmental FILTEREVAL = true

I prefer this way, but that means your documentation is not up-to-date ( eval=true by default in your documentation)

Here the input:

https://raw.githubusercontent.com/mittelmark/pantcl/main/tests/sample.rst

and here the output for example as a PDF file:

sample-rst.pdf

pantcl version V0.9.0

When using pandoc --filter pantcl.tapp with a text like 'foo @text bar', i have an error :

===== ERROR LOG ========
Path element 9: "t" not found
while executing
"::rl_json::json get $jsonData blocks {*}$tkey"
(procedure "walk" line 30)
invoked from within
"walk $nkey $ind"
(procedure "walk" line 21)
invoked from within
"walk $nkey $ind"
(procedure "walk" line 21)
invoked from within
"walk $ckey $ind "
(procedure "walk" line 40)
invoked from within
"walk [list $i c $j]"
(procedure "main" line 59)
invoked from within
"main $jsonData"
invoked from within
"puts -nonewline [main $jsonData]"
(file "/home/myself/.bin/Xwriter/plugin/pantcl.tapp" line 1507)
Error running filter /home/myself/.bin/Xwriter/plugin/pantcl.tapp:
Filter returned error status 1

When using sqlite/cmd filters, you create files in the current folder.
As far as i understood, it's only temporary files that should be written in /tmp or $HOME/.cache/pantcl/ if possible
The 'file=...' should be reserved only if user want to redirect output to a permanent file, but i have the feeling you cannot make code evaluation without
if 'file=...' is absolutely necessary, may be you should authorize absolute path as 'file=/tmp/temp.c' ( I got an error when i tried: couldn't execute "./tmp/temp.c")

Note : i had problem with /tmp and ubuntu last time i tried ubuntu with my application. But .cache can be a good place for small files

The filters works on small document but introduce a high processing time with medium/high-sized document.
I have a difference of nearly 15 seconds with/without filters with an average laptop
This extra processing time append even if the document contains no code block

Is it possible to reduce this processing time with :

• Disable some filters ?
• Change the parsing strategy to apply your filters only on Code block ?

I have introduced your filters in my application : https://github.com/htopkill/Xwriter

Error with filter-pipe on this document: https://github.com/htopkill/test-only file : docs.md

=== ERROR LOG ===
Error running filter /home/myself/.bin/Xwriter/plugin/pantcl.tapp:
Filter returned error status 1
[WARNING] Could not parse YAML metadata at line 3794 column 1: Unexpected '`'
couldn't execute "R": no such file or directory
while executing
"open "|R -q --interactive --vanilla 2>@1" r+"
(procedure "filter-pipe" line 26)
invoked from within
"filter-pipe $c $d"
(procedure "main" line 78)
invoked from within
"main $jsonData"
invoked from within
"puts -nonewline [main $jsonData] "
(file "/home/myself/.bin/Xwriter/plugin/pantcl.tapp" line 1509)

Below is a gzipped version of pantcl version 0.9.12 for Linux 64bit
pantcl.bin.gz