lbryio / grooming Goto Github PK

lbryio/lbry-sdk#762

lbryio/lbryum#155

After repro, update issue with next steps

after investigating, file issues on whats involved in closing lbryio/lbry-sdk#762 (and size them, etc)

Afterwards, move lbryio/lbry-desktop#762 into Groomed

review ghost documentation, try implementing a ghost blog, and see if there are patterns that can be applied to spee.ch

Prototype a version of the speech package and an implementation where the the client and server bundles are built at the implementation level

lbryio/lbry-desktop#829

For lbryio/lbry-sdk#1121
This would allow the API calls to be made before the completion of startup. And the functions requiring specific component will be registered when the component has been loaded. This will be probably be realized using decorators.

this iteration

• research what things need to be moved and such (3 pts)
• dht (5 pts)
• file manager (5 pts)

next iteration

• rationalize component stuff
• peer protocol factory (3)
• reflector protocol factory (3)
• update tests (5)
• actually do what the title says (7)

next-er iteration

• write tests(I HATE THEM SO FAR!!! but I've been told it gets better) (5+5)
• knock out the last thing with component (1)

• reviewing prs
• triaging issues
• engaging community devs

After that, add more info to lbryio/lbry-sdk#956

lbryio/lbry-sdk#803

For lbryio/lbry-sdk#1121
Copying from here

Try to batch calls which query network/blockchain or make them non-blocking by running them after the required parts of daemon have started up. For instance checking up if all the blobs are present in reflector can either be batched or done after the daemon startup

lbryio/lbry-sdk#803

https://github.com/lbryio/internal-issues/issues/95

After: break down #854

set up the new customizeable package on staging.spee.ch

• finish everything in i3, most importantly get it working with the app and demonstrate performance improvements on the things we care about
• release alpha and instructions on how we can test and give you feedback

the pattern we discussed last week for the speech package has a problem in that the client bundle does not receive the custom components in the same way the server can. I think the solution is that the bundle needs to be created at the implementation level rather than the package level. This requires coming up with a pattern that includes a build process in the implementation.

lbryio/lbry-sdk#803

lbryio/lbry-sdk#803

• finish testing the spee.ch vanilla implementation on staging
• merge the spee.ch-as-a-package branch into master
• update the repo name in www.spee.ch package.json
• update readme on all the spee.ch repos to show their connections and use
• add spee.ch-components repo to zenhub

For lbryio/lbry-sdk#1121
Use DeferredList wherever possible to run deferreds in parallel to reduce individual component startup time.

hold another session with Steven where we work on implementing spee.ch and porting over his old content

After this is done, break down lbryio/lbry-sdk#1061

Research the phenomenon of under estimating to make project more palatable for green light and then making quick progress because of the short deadline (in reality, if more effort was put into estimating, it would have shown the deadline is not realistic). Does this have a name? Is it a problem? Are there solutions?

Cannot start this until v20 is released.

• add to lbryschema
• support file that are not lbryfiles
• changes to db
• changes to filemanager and filedownloader (make it an interface)

lbryio/lbry-sdk#762

why use spee.ch?
why run a spee.ch implementation?
what is different from status quo?

assuming lbrycrd changes are merged

• speech is running lbryumx
• victor is done with working on lbryumx
• bonus: test https://github.com/lbryio/lbryschema/issues/41 on speech staging

@jackrobison commented on Tue Feb 06 2018

file json objects (returned get, file_list) by should include the fields:

-claim_name
-claim_id
-txid
-nout
-outpoint
-channel_name
-channel_claim_id

These fields should also be filters for use in file_delete and file_list

finish core functionality II in lbryio/lbry-sdk#1155

10 because 5 for each