东方不败 |labitbu

Gateway: The Transactions Update (v0.7.3) This version enables transaction locator indexing, adds a proper viewport, and simplified timeline. The first thing you'll notice is the interface is getting much sleeker. github.com/Blockamoto/gatewa… VIEWPORT The updated viewport now sits in the top-right, just as you would expect in a video editor inspired design. Not only does it show inscriptions, but also bitfeed.live mondrian blocks display inscriptions on squares and clicking transactions gives you a much better mempool.space style transaction flow. SIMPLIFIED TIMELINE The simplified expandable timeline uses one single scrubbable track and allows you to see the entire blockchain at a glance or zoom all the way into sats! You can still expand to see each indexable lane individually, but I've made simple the default. It's the difference between YouTube and Premiere Pro. But this update is supposed to be about Transactions! So let's get into what's different. THE TRANSACTIONS In the last version we enabled inscriptions, and I described the benefits of the Bitcoin Positional System for fetching inscriptions on-demand from Bitcoin peers with no index required. To do this though, you have to use the Bitcoin Positional System, which includes block in the basic schema to enable this. This is why we can fetch transactions, and thus inscriptions, without having a sync'd node. The schema tells us which block to ask a Bitcoin peer from, and then once we have it, we can locate the transactions and inscriptions and discard the block. I didn't enable the transaction index in that version, as I wanted to demonstrate how you could fetch and index inscriptions without a a sync'd node or with transaction locators indexed. But transaction locators are neccessary if you want to use inscription ID's without manually entering block. This version enables transaction locator indexing, which makes it possible to use inscription ID as normal. Of course, you have to run the index first! It's a much lighter index than the block index, but will take just as long as a normal full sync to fetch them all. If you only need them for inscription ID locating, you only have to sync from the very first inscription. This is a simple tick box in the inscriptions index window. You can do this when setting up your inscription index, or retroactively in that menu. THE INDEX SCHEMA I didn't talk much about the index schema in my last post, but I've made it extremely lean. When indexing inscriptions, it literally just stores the BPS locator so that it can fetch directly: inscription: <tx>i<index>.<block> The transaction locators index bridges the gap between inscription ID's and inscription locators. It simply creates pairs between the transaction ID and transaction index, nested within a block entry. So, to put simply, the inscription index uses BPS to say which block to look at, which transaction number to check, and the index of the inscription. The transaction locator index tells you which block a transaction hash is on, and its transaction index, so that an inscription ID knows where to look. The ironic thing is, if @rodarmor used his originally proposed schema for Ord, which is almost identical to BPS, there would be no need to bridge the gap with a transaction locator index. In that world, you could have just used that schema to locate any inscription. But without trying to build an on-demand system, there's no obvious reason why this earlier proposed schema would be superior. But this is what I'm running into now, and is why I introduced this schema independently of knowing that Casey actually proposed the exact mirror version of mine. Now, since we live in an inscription ID world, that does mean that existing hard-coded references like recursion, delegates, parent/child, etc. do suffer from this lack of being able to fetch on-demand directly from Bitcoin peers. This is why I want to propose the BPS schema to ord, because of the enormous benefits to an on-demand based system. Sure, we could simply create the recursive endpoints ourselves and make them compatible with Gateway and on-demand systems, but then we would be forking, splitting the Ord consensus. So, I'll first go the Ord route so we can at least align on this huge improvement, but I will probably implement it in Gateway in parrallel. I don't know, lot's to consider. But we'll get there. Anyways, the latest update is available on GitHub. It won't show through as an automatic update yet. I've marked it as pre-release. I want to say this was on purpose, but in reality I simply just didn't set it up to feed through to the updater this time around. You can download the update from GitHub, but should be back to normal automatic updates for v0.7.4. THE NEXT VERSION Now, I was planning on releasing Gateway peers next, but I'm toying with releasing further indexes first, so I'm not exactly sure what the next update will be focused on. Maybe I'll do a poll... Regardless, Here's the release: THE RELEASE github.com/Blockamoto/gatewa… Please enjoy! 🫡