post: 0.4.0 release #37
Conversation
|
Could you update the |
|
cc @jbenet |
|
We can update the date field when we publish this |
|
Why don't we close #30, since it is on a remote we can't push to, and just use this as the main new PR? |
|
To add (from @dignifiedquire):
|
| to [dist.ipfs.io](https://dist.ipfs.io/#go-ipfs) and grab the latest version | ||
| from there. Or alternatively, from the same page you can grab the `ipfs-update` | ||
| binary, and use it to perform the upgrade for you. If you installed from source, | ||
| you can simply update your git repo to the `v0.4.0` tag and run `make install`. |
There was a problem hiding this comment.
We might have to mention gx here -- I had to manually run make gx_upgrade and make gxgo_upgrade here the other day.
There was a problem hiding this comment.
Yeah, i really do think that make install should probably install gx automatically
|
I'll contribute a paragraph about v03x/v04x -- will have to change location again in a bit :) |
|
|
||
| Because of this change, you will need to run a migration (from repo version 2 to | ||
| version 3). If you update with the `ipfs-update` tool, this will be done for you | ||
| automatically. |
There was a problem hiding this comment.
Add a sentence here for the non-ipfs-update case, something like "otherwise, ipfs will bail out with repo version mismatch".
something like this: The Public Gateway and BootstrappersWe provide two essential public services to the IPFS community: the public gateway at https://ipfs.io and the default bootstrap nodes. We're making sure that despite the breaking changes, both continue to work with both 0.4.x and 0.3.x for a while. You can read more about the details of this in an earlier blog post: Migrating ipfs.io from go-ipfs 0.3.x to 0.4.0. We expect to keep this grace period open until the end of April 2016. |
|
@jbenet could you please review? |
|
Overall, looks great. I think there’s a few things we should add — for instance, the entire changelog. I've made a PR with my edits to your branch -- see #41. And we still need pictures. |
| permanently into a single multiplexer that won't work for every language or | ||
| situation, this change allows each IPFS implementation to implement the multiplexers | ||
| of their choosing, and then negotiate which to use during the initial connection | ||
| handshake. |
There was a problem hiding this comment.
What about linking to the multistream spec https://github.com/jbenet/multistream? So that the reader can understand what kind of black magic happens
|
|
||
| ### Breaking changes | ||
|
|
||
| This is a **breaking change** release, with a major refactor and upgrade to the |
There was a problem hiding this comment.
I'm afraid that **breaking change** alone might mislead people in thinking that we broke data linking. Calling it a network breaking change and then explaining in the end of the paragraph that data remains intact will avoid many doubts
|
preview: https://ipfs.io/ipfs/Qmab1no1nx6svsjVG7kvUBKkXS6diRRvocMkEYMNxivHNK/14-ipfs-0-4-0-released/
|
ghost
This squashes a couple of commits by @whyrusleeping and @kyledrake
|
LGTM! |
Made some changes to #30