feat: add from fs #1777
feat: add from fs #1777
Conversation
ghost
assigned 
alanshaw
force-pushed
the
feat/add-from-fs
branch
from
a1b56d5
to
fa9158e
Compare
alanshaw
force-pushed
the
feat/add-from-fs
branch
from
fa9158e
to
200e4fb
Compare
| const Path = require('path') | ||
| const isString = require('lodash/isString') | ||
| const pull = require('pull-stream') | ||
| const glob = require('glob') |
There was a problem hiding this comment.
can't we just use https://www.npmjs.com/package/fast-glob ? it's faster and supports streams so we can go directly to pull-stream
There was a problem hiding this comment.
I tried it, and it's not significantly faster than glob, in fact in most tests it's slower (left is glob, right is fast-glob).
There was a problem hiding this comment.
even using streams instead of glob events ? seems weird but if its faster like this im good with it
There was a problem hiding this comment.
I know weird, maybe it is the stream to pull stream conversion? Anyway even weirder is that the benchmarks for fast-glob do not run against glob 😱 They benchmark globby, bash-glob, tiny-glob and glob-stream...but not regular glob 🙄.
There was a problem hiding this comment.
lol that kinda wtf and globby is based on fast-glob with some sindre sugar on top
anyway this looks fine and if your benchs show better numbers lets merge it
alanshaw
force-pushed
the
feat/add-from-fs
branch
from
ca34d52
to
e24fc37
Compare
License: MIT Signed-off-by: Alan Shaw <alan.shaw@protocol.ai>
This PR adds a new method [`addFromFs`](https://github.com/ipfs/interface-ipfs-core/blob/master/SPEC/FILES.md#addfromfs) allowing users to more easily add files from their file system without having to specify every single file to add. In the browser the user will receive a "not available" error. I've pulled out a module `glob-source.js` - call it with some file paths and it returns a pull stream source that can be piped to `ipfs.addPullStream`. This PR comes with the following added benefits: * `ipfs add` on the CLI uses `glob-source.js` - **nice and DRY** * `glob-source.js` uses the events that the `glob` module provides allowing the globbing to be a `pull-pushable`, which means that matched paths can begin to be added before all the globbing is done - **faster** * `ipfs add` now supports adding multiple paths, fixes #1625 - **better** * `ipfs add --progress=false` doesn't calculate the total size of the files to be added anymore! It didn't need to do that as the total was completely discarded when progress was disabled. It means we can add BIGGER directories without running into memory issues - **stronger** License: MIT Signed-off-by: Alan Shaw <alan.shaw@protocol.ai>
License: MIT Signed-off-by: Alan Shaw <alan.shaw@protocol.ai>
License: MIT Signed-off-by: Alan Shaw <alan.shaw@protocol.ai>
License: MIT Signed-off-by: Alan Shaw <alan.shaw@protocol.ai>
License: MIT Signed-off-by: Alan Shaw <alan.shaw@protocol.ai>
License: MIT Signed-off-by: Alan Shaw <alan.shaw@protocol.ai>
alanshaw
force-pushed
the
feat/add-from-fs
branch
from
5d486a0
to
158d37d
Compare
License: MIT Signed-off-by: Alan Shaw <alan.shaw@protocol.ai>
License: MIT Signed-off-by: Alan Shaw <alan.shaw@protocol.ai>
|
@hugomrdias is this gtg now? |
This PR adds a new method
addFromFsallowing users to more easily add files from their file system without having to specify every single file to add.In the browser the user will receive a "not available" error.
I've pulled out a module
glob-source.js- call it with some file paths and it returns a pull stream source that can be piped toipfs.addPullStream.This PR comes with the following added benefits:
ipfs addon the CLI usesglob-source.js- nice and DRYglob-source.jsuses the events that theglobmodule provides allowing the globbing to be apull-pushable, which means that matched paths can begin to be added before all the globbing is done - fasteripfs addnow supports adding multiple paths, fixes Cannot add multiple files from command line #1625 - betteripfs add --progress=falsedoesn't calculate the total size of the files to be added anymore! It didn't need to do that as the total was completely discarded when progress was disabled. It means we can add BIGGER directories without running into memory issues - stronger