missing_required_file -> ERROR -> feed_info.txt

[jollydogn]

Describe the bug

Hello, I am using v7.1.0, and despite not having a translations.txt file, it sees feed_info.txt as a required file. What can I do about this issue?

Steps/Code to Reproduce

Hello, I am using v7.1.0, and despite not having a translations.txt file, it sees feed_info.txt as a required file. What can I do about this issue?

Expected Results

Actual Results

I am waiting for it to pass validation

Screenshots

No response

Files used

No response

Validator version

v7.1.0,

Operating system

Linux

Java version

No response

Additional notes

No response

[skalexch]

hello @jollydogn , can you please provide the GTFS file you used for this validation?

[jollydogn]

hello @skalexch , I can’t send the file, but the files are like this.

[skalexch]

@jollydogn that's strange. I used a random GTFS from the MobilityDatabase. It worked correctly.
With feed_info.txt but no translations.txt -> Missing Required File
With feed_info.txt AND no translations.txt -> Missing Recommended File

The code that performs this validation is this piece. It looks correct (if no feed_info ->{if no translations -> missing recommended else -> missing required}).

@davidgamez do you see any possible other cause for this?
@jollydogn can you please try to test it on the online validator?

[jollydogn]

@skalexch Yes, it's not giving any errors right now. It's starting to give a very interesting error, but this time. Before I wrote this here, it was only giving me the feed_info.txt missing file error, and now it's giving me errors related to routes and networks. But I don't have any unnecessary or excessive .txt files here. I defined my networks and defined relationships in the route table, so why did it start giving such an error? The gtfs file I used is attached.

[jollydogn]

20250925_Güncel GTFS 2025.zip

[skalexch]

@jollydogn thank you for sharing! The spec logic says that route_networks.txt should not exist if network_id is specified in routes.txt.

The validator seems to do this check for both route_networks.txt (rule which the spec mentions) and networks.txt (rule doesn't exist in the spec), which is why the error is raised for you.

We will sync internally on why networks.txt is also considered in for this notice and we will address it.

Anyway, your GTFS is compliant with the spec. Nevertheless, to remove the error for now, you can delete networks.txt. And if you want the network_ids to be human-readable you can name them the same as their network names (e.g. network_id=1 is for Vapur, you can just replace all network_id=1 values with vapur in routes.txt).

[jollydogn]

@skalexch Thank you for the feedback. Implementing such a solution currently causes some issues in my system, so I hope a development will be made as soon as possible. Thank you for your interest.

[jollydogn]

I implemented it relationally, so writing a string expression there might be a bit costly for me.

[jollydogn]

@skalexch Also, it affects the 'fare' aspect. If I were to implement such a development now, so I will wait for the improvement in the validation.

[skalexch]

@jollydogn I missed one section of the spec. In fact, networks.txt is forbidden when network_id exists in routes.txt.

This is the network PR. Back then, the community agreed to keep this rule to not break backward compatibility (further context).

So for this case, the validator is actually following the spec. I will raise an editorial change to make sure the spec is consistent. For this case, you can just delete networks.txt.

Otherwise, if you want to implement fares you can still either use network_id in routes.txt or remove it and create the mapping route_networks.txt along networks.txt.

Also @jollydogn can you please share the first validator report where there was the missing feed info ERROR?