Planenmo test with <has_line line="..."> fails when stdout is bigger than 2**14 characters long.

When we run tools the stdout is cut if its length is greater than 2**14 characters.

This behavior causes error in the `has_line` test, as it seems to rely on the "size-reduced" stdout instead of the "whole" stdout.

Success and fail examples
---
Here are two examples of xml tools.
This one works fine because the output will be 1200 chars long:
```xml

python

```

But in this one, the test fails because the stdout is cut in the area shown above:
```xml

python

```

Output is cut here:
```txt
02727
02728
02729
02
..
9
03270
03271
03272
```

So, the line "03000" cannot be found, even if it is present in the logs.

As we can see, the first part of the output is 6*2729+3 (=16377) characters long. Which is only seven characters away from to 2**14 (=16384).

I suppose other tests like `has_text`, `not_has_text`, `has_text_matching` - and so on - will be prone to this problem.

EDIT: `has_text` is error prone too.

tests
---
to test these files; I ran the following script:

```sh
VENV=.venv
TEST_DIR=/tmp/test_tool

if [ ! -d ".venv" ];then
python3 -m virtualenv ${VENV} ;
fi

. ${VENV}/bin/activate

if [ "$(which planemo)" = "" ];then
pip install planemo ;
fi

if [ ! -d "${TEST_DIR}" ];then
planemo conda_install --conda_prefix ${TEST_DIR} . ;
fi

mkdir -p test-data
touch test-data/input.txt

planemo lint --fail_level error ./*.xml || exit 255

planemo test \
--install_galaxy \
--conda_dependency_resolution \
--conda_prefix ${TEST_DIR} \
--no_cleanup \
;
```

Start by running the provided shell reproduction with `planemo test` and the two XML tool examples. Trace the `has_line` and `has_text` checks against the truncated stdout, then verify that assertions can inspect the whole output while any display truncation remains. Done means the `03000` line and longer-text checks pass beyond 2**14 characters.

Written by the indexing model from the issue text.