Are you using Newman as a library, or via the CLI? - CLI
Did you encounter this recently, or has this bug always been there: Recently
Expected behaviour: Print more details about the error / hints at resolution
We were blocked for weeks because newman would fail with only the following:
Parse error
No details at all about what couldn't be parsed. I used htmlextra reporter to try to capture information about the response such as headers, body. This yielded absolutely nothing:
Finally we contacted our CDN+WAF, Imperva (formerly Incapsula). They confirmed that sometimes to detect bots they insert a "classification cookie" into their responses. Explanation: Browsers can handle the cookie, but Bots cannot. When they disabled this "classification cookie" feature, the Parse Error issue went away.
I'm creating this issue primarily in hopes that Imperva will work with Postman Labs & Co to resolve this issue, or at least so other people can see they aren't alone in if they encounter this issue (perhaps future me).
Ideally, the Parse Error should include more verbosity. For example, if the cookie was corrupted / invalid, it should report that rather than failing without any details.
(Imperva mentioned that they intentionally send a bad cookie that browsers discard but bots choke on).
newman -v
):We were blocked for weeks because newman would fail with only the following:
No details at all about what couldn't be parsed. I used htmlextra reporter to try to capture information about the response such as headers, body. This yielded absolutely nothing:
Finally we contacted our CDN+WAF, Imperva (formerly Incapsula). They confirmed that sometimes to detect bots they insert a "classification cookie" into their responses. Explanation: Browsers can handle the cookie, but Bots cannot. When they disabled this "classification cookie" feature, the
Parse Error
issue went away.See: https://www.imperva.com/blog/how-incapsula-client-classification-challenges-bots/
I'm creating this issue primarily in hopes that Imperva will work with Postman Labs & Co to resolve this issue, or at least so other people can see they aren't alone in if they encounter this issue (perhaps future me).
Ideally, the
Parse Error
should include more verbosity. For example, if the cookie was corrupted / invalid, it should report that rather than failing without any details.(Imperva mentioned that they intentionally send a bad cookie that browsers discard but bots choke on).