Sean Owen

I think that's inevitable, as it now refuses to consider weakly detected points. Maybe I can tweak the logic a little to catch edge cases, not sure.

I don't know that this is a bug. Generally you'd put one event in a QR code. I don't know how a device would act on several at once. Still,...

Yes, the right place. I'm sure it's syntactically correct and therefore makes some sense to parse it. After all it's not just phones that use the decoding logic. But I...

You're right, I don't believe it actually respects those values. The DataMatrix encoder was adopted from another project and works somewhat differently. If you're interested in looking at plumbing through...

Is this related to what the oss-fuzz project is finding? I see https://github.com/google/oss-fuzz/blob/90dc42fb873d18c515056d1577cd1ec5c291c4e5/projects/zxing/MultiFormatEncodeFuzzer.java but that doesn't seem to have a main method to run. Or is it possible to see...

Oh I see this is just https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=37794&q=label%3AProj-zxing Yeah, this one I'm having a harder time figuring out. It's a real issue I'm pretty sure, even if a corner case.

Looks like it's not public, yeah. It could well be the same. The oss-fuzz project is running fuzzing automatically now and is generating a number of issues like this. This...

Doesn't seem like the empty string - there is a whole barcode here in the log output. It failed to decode something that was encoded

OK then a problem with the encoding? I think this test is a round-trip one

Right, I mean it is probably not encoding the empty string, but something that results in this incorrect encoding - seems like it doesn't make sense that there are 'no'...