[Bug]: MessageEvent.visualize returns blank output for an empty TextContent block

Is there an existing issue for the same bug?

• I have searched existing issues and this is not a duplicate.

Bug Description

MessageEvent.visualize handles two semantically empty message representations inconsistently.

A message with no content blocks displays the existing placeholder:

Message(role="user", content=[])
# visualize.plain == "[no text content]"

However, a valid message containing one empty TextContent block produces blank output:

Message(
    role="user",
    content=[TextContent(text="")],
)
# visualize.plain == ""

DefaultConversationVisualizer subsequently skips the second event because its rendered text is empty. The event therefore disappears from conversation visualization instead of displaying [no text content].

Expected Behavior

Both empty representations should produce:

[no text content]

DefaultConversationVisualizer should create a visible event block for both representations. Non-empty text and image content should remain unchanged.

Actual Behavior

The two representations produce different results:

content=[]:
visualize.plain = "[no text content]"
event block created = True

content=[TextContent(text="")]:
visualize.plain = ""
event block created = False

The underlying event remains valid, but the default conversation visualizer omits it.

Steps to Reproduce

1. Check out release v1.44.1 at commit 9d143aac35c2dcec9cbb046ff9f35ac5eb072f6a.

2. Set up the development environment:

make build

3. Create tests/sdk/event/test_message_event_empty_visualization.py:

from openhands.sdk.conversation.visualizer.default import (
    DefaultConversationVisualizer,
)
from openhands.sdk.event.llm_convertible import MessageEvent
from openhands.sdk.llm import Message, TextContent


def make_event(content: list[TextContent]) -> MessageEvent:
    return MessageEvent(
        source="user",
        llm_message=Message(
            role="user",
            content=content,
        ),
    )


def test_empty_text_block_uses_no_text_placeholder():
    no_blocks = make_event([])
    empty_block = make_event([TextContent(text="")])
    visualizer = DefaultConversationVisualizer()

    observed = {
        "no_blocks_text": no_blocks.visualize.plain,
        "empty_block_text": empty_block.visualize.plain,
        "no_blocks_rendered": (
            visualizer._create_event_block(no_blocks) is not None
        ),
        "empty_block_rendered": (
            visualizer._create_event_block(empty_block) is not None
        ),
    }

    assert observed == {
        "no_blocks_text": "[no text content]",
        "empty_block_text": "[no text content]",
        "no_blocks_rendered": True,
        "empty_block_rendered": True,
    }

4. Run:

uv run pytest \
    tests/sdk/event/test_message_event_empty_visualization.py \
    -q

5. Observe that the test fails because empty_block_text is "" and empty_block_rendered is False.

Acceptance Criteria

• MessageEvent.visualize renders [no text content] for content=[].
• It renders the same placeholder for [TextContent(text="")].
• DefaultConversationVisualizer does not discard the empty-block event.
• Non-empty text and image visualization remain unchanged.
• Regression coverage includes both empty representations.

Installation Method

Source checkout using make build (uv sync --dev)

If you selected "Other", please specify

Not applicable

SDK Version

1.44.1, main@9d143aac35c2dcec9cbb046ff9f35ac5eb072f6a

Version Confirmation

• I have confirmed this bug exists on the LATEST version of OpenHands SDK

Python Version

3.13.2

Model Name (if applicable)

Not applicable; reproduced by directly invoking deterministic event visualization code.

Operating System

MacOS

Logs and Error Messages

content=[]:
visualize.plain = '[no text content]'
event block created = True

content=[TextContent(text='')]:
visualize.plain = ''
event block created = False

The received assertion value is equivalent to:

{
    "no_blocks_text": "[no text content]",
    "empty_block_text": "",
    "no_blocks_rendered": True,
    "empty_block_rendered": False,
}

Minimal Code Sample

from openhands.sdk.event.llm_convertible import MessageEvent
from openhands.sdk.llm import Message, TextContent


event = MessageEvent(
    source="user",
    llm_message=Message(
        role="user",
        content=[TextContent(text="")],
    ),
)

assert event.visualize.plain == "[no text content]"

Screenshots and Additional Context

No screenshot is required; this is a deterministic unit-level reproduction using public SDK models.

The root cause is that content_to_str returns [""] for the empty block. Although its only string is empty, the list itself is truthy:

text_parts = content_to_str(self.llm_message.content)

if text_parts:
    full_content = "".join(text_parts)
    content.append(full_content)
else:
    content.append("[no text content]")

Joining [""] produces "", so the placeholder branch is skipped. The default visualizer then removes the event:

content = event.visualize

if not content.plain.strip():
    return None

The repository already treats [TextContent(text="")] as blank content in the native test test_message_tool_calls_strip_blank_list_content. This is therefore a supported model shape rather than an invalid handcrafted value.

A possible fix is to decide whether to use the placeholder after joining the text parts:

full_content = "".join(text_parts)

if full_content:
    content.append(full_content)
else:
    content.append("[no text content]")

The relevant implementation is unchanged through main@704cbe6015e3d59cabe04632175d99df2d448999. Targeted issue and pull-request searches found no same-root report.

Start at MessageEvent.visualize and content_to_str, then run tests/sdk/event/test_message_event_empty_visualization.py with the provided uv command. Check the existing native test test_message_tool_calls_strip_blank_list_content for related blank-content behavior; done means both empty representations render [no text content], remain visible, and non-empty text and image output is unchanged.