This is the new hash on my machine

[freeformz]

I assume some stuff changed in gps. Discovered when I ran dep status and got nothing.

[sdboyer]

yeah gps changed to no longer include stdlib imports in the hashing calculations (derp). i think (read: really really hope) it'll be a lot more stable now.

[sdboyer]

...however, i get a different one when i dep ensure locally from the current tip of master, having built dep from the tip. ugh. i guess i have some investigating to do

[sdboyer]

i'm gonna add a method to print the input components to the hash mechanism. This is just too annoying to debug without something like that.

sdboyer/gps#137

[sdboyer]

New func now exists, will have up a PR soon

[freeformz]

So, even after #89 was merged my hash is still this new one.

[freeformz]

@sdboyer Can you point me back at the code gps uses to hash? I think I'd like to change it a bit (or at least understand it).

[sdboyer]

@freeformz sure - hashing logic is in hash.go

I'm working on some updates to it right now that will trim things down a little further - sdboyer/gps#125

[jessfraz]

I get a different hash now: "memo": "ef5a4c5ae6fa16e842be8c0bc1ce7128d5d3f4790ca444da9ed6eb0b9b4344fd",

[sdboyer]

I'm almost done with the improvements to the input hashing algorithm - sdboyer/gps#143. I'm nearly done with the implementation; mostly I'm just finishing covering the bases with tests. There's a lot of subtle cases that should be clearly and explicitly documented/tested.

There's not much point in worrying about this until that's done, as a) the algorithm will change and b) the output from the new HashingInputsAsString() func won't be easily readable. The PR introduces line breaks between each input, and adds "section headers" to indicate what the type of info is in each section. Makes it far easier to know wtf is going on.

[sdboyer]

I added a (hidden) hash-inputs subcommand in #97, so we can now easily compare our inputs to figure out why we've got these differences happening. When building from that branch, here's what I have when running from the root of the hoard project:

$ hoard hash-inputs
-CONSTRAINTS-
github.com/Masterminds/semver
b-2.x
github.com/Masterminds/vcs
svc-^1.8.0
github.com/pkg/errors
svc->=0.8.0, <1.0.0
github.com/sdboyer/gps
b-master
-IMPORTS/REQS-
github.com/Masterminds/semver
github.com/Masterminds/vcs
github.com/pkg/errors
github.com/sdboyer/gps
-IGNORES-
-OVERRIDES-
-ANALYZER-
hoard
0.0.1

[freeformz]

$ hoard hash-inputs
-CONSTRAINTS-
github.com/Masterminds/semver
b-2.x
github.com/Masterminds/vcs
svc-^1.8.0
github.com/pkg/errors
svc->=0.8.0, <1.0.0
github.com/sdboyer/gps
b-master
-IMPORTS/REQS-
github.com/Masterminds/semver
github.com/Masterminds/vcs
github.com/pkg/errors
github.com/sdboyer/gps
-IGNORES-
-OVERRIDES-
-ANALYZER-
hoard
0.0.1

[freeformz]

The memo I get now is cdafe8641b28cd16fe025df278b0a49b9416859345d8b6ba0ace0272b74925ee

[sdboyer]

The memo I get now is cdafe8641b28cd16fe025df278b0a49b9416859345d8b6ba0ace0272b74925ee

that's from a simple hoard ensure?

[sdboyer]

i've got the same digest, now. it's weird that i didn't commit with the same on sunday, but i may have taken a little shortcut that explains that.

[freeformz]

Yes hoard ensure.

[jessfraz]

I get cdafe8641b28cd16fe025df278b0a49b9416859345d8b6ba0ace0272b74925ee when I use the hoard built from #97

[freeformz]

I think this can be closed now, so doing so.