pablomartin4btc

Updated changes: - Updated the code with a new workaround to avoid failures on unit tests (`httpserver_tests`). Thanks @stickies-v for your prompt response and your support on quick workarounds on...

> but I think the commit message [should be updated](https://github.com/bitcoin/bitcoin/pull/27253#pullrequestreview-1348929370) to better describe the bug, the fix, and the behaviour change introduced Yes, I have that pending to do asap.

> > It would be better to have two separate commits - one with the non-functional optimization and one with the fix. > Agreed, that could probably be improved. I...

Updated changes: - Moving the uri parsing validation from the http_request_cb to the HTTPRequest constructor, while doing this realised that it's better to validate the result of `evhttp_request_get_uri()` also at...

Regarding CI failures: - `./test/functional/wallet_create_tx.py` passes locally, actioned manually re-run and it passes too. - checking the fuzzer one `cirrus-ci-build/ci/scratch/qa-assets/fuzz_seed_corpus/http_request" failed with exit code 77` - `fuzz: httpserver.cpp:618: void HTTPRequest::WriteReply(int,...

Working on functional failure`interface_rest`on the invalid uri case, something didn't like from latest changes. I think I've found the issue! Both, .py and fuzz were getting the same failure `Assertion...

> Try to set it to `false` since we are calling `ReadBody()` to read and process the result. I think we need to re-think this fuzz test (`http_request.cpp`), as due...

> I think that is fine because if the input is such that `HTTPRequest` constructor throws, then we need not to test those because they will never happen in the...

Updated changes: - Fixed fuzz test failing. > Since the replySent parameter of the HTTPRequest constructor is used only from the tests, I would suggest to remove it and adjust...

> Would be nice to split the actual fix from the other changes. I guess it would be easier for other reviewers too. I'm working on it. > I would...