Diff tool failed

[Mac-V-Sog]

App Version

3.1.1

API Provider

Google Gemini

Model Used

Gemini 2.5 Pro

Actual vs. Expected Behavior

I have a 570k context window and start noticing that the diff tool fails all the time now. Roo reverts to write_to_file which does actually help, but obviously not optimised.

This did not happen when I first started the project, but possibly due to large context and/or project.

Detailed Steps to Reproduce

1. Have 500k+ context
2. Make any changes to or modifications.

Relevant API Request Output

Additional Context

No response

[snowleopardd]

same, been getting this a lot with the new version.

[shushu789]

The same thing happens mainly in contexts larger than 300k. I'm using Gemini 2.5 pro.

[hannesrudolph]

The same thing happens mainly in contexts larger than 300k. I'm using Gemini 2.5 pro.

yes Iit does seem to be on larger files or larger contexts.. hmm.. i wonder if the output gemini is putting out degrades at a higher context so its ability to follow the rules needed to make the tool calls falters. @KJ7LNW thoughts?

@Mac-V-Sog the new version has a more agressive multiblock diff process which makes multiple edits in a larger file instead of writing the whole file out or making a number of single diffs.

[hannesrudolph]

BTW @Mac-V-Sog if it is happening when the context is larger that is most certainly because gemini is not handling its tool calls well at higher contexts. Thats a gemini problem then and has nothign to do with Roo.. I think.. that being said.. I am not ruling out Roo being the cause. We need to keep an eye on this. Please update with anything else you discover. Thank you.

[shushu789]

@hannesrudolph I encountered this many times today. Let me correct you. This problem occurs not only in larger context windows, but also in context windows below 100k. However, it is likely to be solved automatically after multiple retries. I am not sure whether it is related to the Gemini model or roo, because I only have the Gemini API.

[hannesrudolph]

Thanks for the update. @shushu789

[Mac-V-Sog]

BTW @Mac-V-Sog if it is happening when the context is larger that is most certainly because gemini is not handling its tool calls well at higher contexts. Thats a gemini problem then and has nothign to do with Roo.. I think.. that being said.. I am not ruling out Roo being the cause. We need to keep an eye on this. Please update with anything else you discover. Thank you.

Thanks - I will do some more testing this evening and report back. Presumably Google are monitoring all the APIs too as part of this 'free use' of Gemini. I have already seen some people commenting on how Gemini is unable to recall information despite the 1mn context window. We must note that it is experimental as most things in this space. That being said, it has done a good job of pushing through the errors and made progress with my app... Roo Code has been much more successful than Cline.

[JabolDev]

BTW @Mac-V-Sog if it is happening when the context is larger that is most certainly because gemini is not handling its tool calls well at higher contexts. Thats a gemini problem then and has nothign to do with Roo.. I think.. that being said.. I am not ruling out Roo being the cause. We need to keep an eye on this. Please update with anything else you discover. Thank you.

I dont want to sound rude, but the solution here is to refactor the whole system instruction, this is you guys supposed to be in top priority instead of adding new feats,

Current system instruction design is very poor, it works to claude, but most of the llm including gemini struggles and confuse. 2.0 flash is crap in diffs but not 2.5 i dont have a single issue with 2.5.

If total system instruction is higher than 500 lines and too many redundant unclear, not concise instruction, dont expect it to work well specially to gemini.

Suggested principle for instruction.
Concise and direct.

If you guys dont want to risk of degrading claude for new instruction, create a system where theres a unique instruction per model.

[mandelbox]

+1. I'm struggling to get Gemini to work with diffs at all, even sub 100k context windows. I have disabled diffs for now. Given how capable Gemini 2.5 Pro is, I suspect some system prompt tuning would do the trick.

[KJ7LNW]

Everyone on this ticket: In order to troubleshoot this, I need the following when errors occur:

1. click the button in blue and save the file:

2. scroll to the and of the file and find the <error> response:

<error>
...
</error>

3. scroll to the end of the file and find the apply_diff section that failed. For example:

<apply_diff>
<path>story.txt</path>
<diff>
<<<<<<< SEARCH
:start_line:2
:end_line:4
-------
Wind whispered through the broken panes.
=======
Wind whispered through the broken glass.
>>>>>>> REPLACE
...
</diff>
</apply_diff>

4. Do not modify the content

Attach both XML samples to your response.

I know this may contain confidential information, and you can decline to share it if you wish, but if you modify the content it may be difficult for me to fully understand edge cases. If you do change the content for any reason please tell me so that I am aware that the information maybe incomplete.

5. reproducible?

If you have a reproducer please provide as much information as you can. I would love to replicate this exactly and reliably, so it can be solved either with force-feedback to the model or system instructions.