"Windows" new line required for some banks about ofxgo HOT 12 CLOSED

No worries - life happens.

Thanks for the reply - my opinion is that we shouldn't let lack of an authoritative source stop us from moving forward (though I can still hope for one!).

from ofxgo.

That's directly from USAA, I'm pretty sure.

Ah, okay. The name "OFXAdapter" made me think you might be using something in the middle to try to make it behave.

I'd welcome a PR for this. My initial thoughts are that it would require both a change to the XML library (see https://github.com/aclindsa/xml/blob/207ec7fb74201d07016cbd7064cb6331dde91eba/marshal.go#L962) and where ofxgo writes the SGML header: https://github.com/aclindsa/xml/blob/207ec7fb74201d07016cbd7064cb6331dde91eba/marshal.go#L962. There may be a few other places sprinkled around that I haven't discovered yet that you'll find during testing!

I think ideally which line ending to use would be a runtime-configurable parameter, and not a hard-coded thing (in both the xml and ofxgo libraries).

from ofxgo.

@JohnStarich Well-deserved pokes, to be sure. Sorry about the delay - unfortunately you lost the latency lottery and sent these when I'm busier than I've been in a while. I have not forgotten about them, though, and hope to review them this week!

from ofxgo.

I think I'm okay with this in principle (though I'd like to do a little sanity check of my own that a few common banks continue to work after the change).

Is the "OFXAdapter: Failed to parse request..." message you're getting a straight response from USAA, or are you using some other piece of software to intermediate?

from ofxgo.

Ok cool. Yeah, makes sense! 2 of my credit cards and both of my USAA accounts all work, if that helps 👍 Probably needs some more testing than just my accounts, but I don't have very many to try, unfortunately 😅

That's directly from USAA, I'm pretty sure. That line appears above the OFX headers but after the HTTP headers in the raw response. (I had a similar, but different message when I passed closing tags in my other issue #18)

from ofxgo.

Sounds good 👍 thanks for the tips!
I like the config parameter idea. It seems like \r\n line-endings are compatible with more institutions, so I think it would make a good default value, then this parameter could let you switch to unix line-endings if necessary?

I'll be out of town next week, but I'll probably give it a shot the week after 🌴

from ofxgo.

It seems like \r\n line-endings are compatible with more institutions

Is your basis for this that the two projects you mentioned above do it that way, or have you found other info? I think the projects are probably sufficient evidence, but I'm curious to learn more if you came across other sources...

from ofxgo.

Sorry about my delay in getting back to you!

I based that on those projects I mentioned, and they appear to have a little traction based on their GitHub stats. It isn't really much to go on, I haven't done much research in this area 😉
Based on my brief searches (best one seems to be "ofx" carriage return), this issue does seem to pop up on forums and in a few libraries

from ofxgo.

First PR is ready to go! aclindsa/xml#2

from ofxgo.

friendly poke 👉 @aclindsa

from ofxgo.

Oh no! 😆

No rush, just wanted to make sure you weren’t asleep at the wheel! I’ve got work-arounds for now, anyway 😉

from ofxgo.

Second part is up now too, for extra fun 💥 #20

from ofxgo.