Ryan

@dabasov I thought so too My original PR did this: https://github.com/0chain/0chain/pull/1548/files but Peter and Dayi disagreed See this conversation https://0chain.slack.com/archives/C01U94VF56Y/p1659679807160019?thread_ts=1659559029.315439&channel=C01U94VF56Y&message_ts=1659679807.160019 This is where this PR comes from

@sculptex I understand the concern but this is already a massive undertaking (100+ endpoints and counting) https://docs.google.com/spreadsheets/d/1beGvye1RcGFpfoTJJy1_ltNQO_7ujJPT5Sy4OLYQqUY/edit?usp=sharing It's not feasible to introduce a duplicate endpoint and flags for each and...

@sculptex i get you, but that's a lot more work/maintenance than it looks when you're talking about 100 paths good point, i missed that, will fix now

I totally understand what you mean but like I said there are 100+ endpoints that need renamed, that's 100+ new branches in the code, and 200+ new unit tests (deprecated...

`"It wouldn't be 100+ new branches, the file the code snippet was from had 15 or so paths in the one file."` good point, maybe more like 20 branches, but...

It's a standard pattern for migrating versioned endpoints - v1 endpoints remain where they are, no new endpoints are added to v1 - Any new endpoints are added to v2...

Totally and i'm very glad you have chipped in, this stuff is important Yes this needs fixed before mainnet, but with the v2 approach we can migrate and deprecate even...

unit tests shouldn't rely on an external service, can calls to Postgres mot be mocked instead?

@peterlimg just checking if this is still a valid issue, we still see flaky behaviour/periodic test failures

completely agreed @kushthedude - I'd like to see each request logged with a unique transaction id and return this to the user on error https://0chain.slack.com/archives/C028ZCJT6LU/p1628543067137800