Replaced NTA_ROOTDIR by NTA

[david-ragazzi]

$NTA_ROOTDIR is redundant. Its remotion decreases the number of environment variables to the user set.

[david-ragazzi]

Hi @rhyolight . You can merge now..

[rhyolight]

LGTM, but I'd like @scottpurdy to approve this first.

[scottpurdy]

I don't know what ignore = dirty does for the submodule. And I haven't had a chance to try this. But seems like a move in the right direction.

[david-ragazzi]

I don't know what ignore = dirty does for the submodule. And I haven't had a chance to try this. But seems like a move in the right direction.

ignore = dirty option avoids that git include the word -dirty in the end of the submodule SHA when files in a submodule folder are changed.. It's a headache try finding what was changed and revert this changes, as github doesn't give more details about where occurred the changes.. If these changes aren't reverted, it's impossible synchronize nupic repo with a "dirty" submodule..

[breznak]

On Thu, Apr 10, 2014 at 1:02 PM, David Ragazzi notifications@github.comwrote:

I don't know what ignore = dirty does for the submodule. And I haven't
had a chance to try this. But seems like a move in the right direction.

ignore = dirty option avoids that git include the word -dirty in the end
of the submodule SHA when a submodule folder is changed..

I'm not sure we want this, as git rightfully informs you the something has
changed (?)

It's a headache try finding what was changed and revert this changes, as
github doesn't give more details about where occurred the changes..

git stash will drop the -dirty changes
git diff will tell you what's changed--or that's what I do. (both mus be
done in the nta/ subfolder)

If these changes aren't reverted, it's impossible synchronize nupic repo
with a "dirty" submodule..

[david-ragazzi]

Hi @breznak,

As we should pull PRs directly from nupic.core repo cloned to our machine, not from its /nta subfolder in nupic, I think is simpler ignore the changes were in /nta and so avoid type the commands that you suggested..

I saw several developers that recommend this option:

http://stackoverflow.com/questions/3240881/git-can-i-suppress-listing-of-modified-content-dirty-submodule-entries-in-sta

http://www.nils-haldenwang.de/frameworks-and-tools/git/how-to-ignore-changes-in-git-submodules

bronson/vim-update-bundles#32

Anyway, I liked these commands that you suggested, but my lazyness claims for ignore = dirty.. :-)

[david-ragazzi]

I reverted "ignore = dirty" in order to this PR is related only to NTA_ROOTDIR env variable.. and I created a specific PR for "ignore = dirty" (#794)..

[rhyolight]

@david-ragazzi Ok, don't hate me for this, but I'm going to wait to merge this PR until Wednesday. Our last merge of PR #751 caused considerable difficulties in the Grok pipelines for @jcasner and the Grok crew, which is my fault for not properly notifying them of the coming build changes. In the meantime (while they mop up the damage and prepare for an upcoming Grok release), they have asked for me to hold of any any PRs that could potentially affect their builds.

This is not a good situation for NuPIC, and it won't last long. Grok engineers will be working next week to improve their pipelines so NuPIC merges that break their pipelines won't prevent them from building and releasing against older versions of NuPIC.

Other PRs that will not potentially affect the Grok builds will still get merged in the meantime.

[david-ragazzi]

I'm feel sorry for this @rhyolight . :-(

The problem is I can't imagine the effects of the PRs over Grok and thus it's impossible I advise you about what are the risks involved to merge a PR.. I only run locally, Travis become green, and other people review and test.. This is only what we have to say if something is good or not..

Anyway, I hope this situation normalize ASAP...

[rhyolight]

I hope this gets resolved soon. We can't expect OS contributors to anticipate how their changes will affect Grok pipelines, so right now the responsibility for that lands squarely on me (how uncomfortable! ). So I'm really just covering my own ass here by holding off on this merge.

[jcasner]

All, I agree with @rhyolight, this is a temporary scenario because our build logic is flawed internally. We're actively working to resolve that, so this should be the last time we make this request of the NuPIC community. That said, I also have to balance getting commercial software out the door on time 😄

Our builds are all stable at the moment, and I'll check back in here once we have an a final release candidate so that we can merge this one, hopefully sooner than Wednesday.

@unixorn can you please review this one and make sure your team is prepared for the build changes next week?

[rhyolight]

@jcasner To be clear, I don't think this change will affect Grok builds, but I choose to err on the side of caution.

[david-ragazzi]

@rhyolight @jcasner I feel that most PRs are reviewed having in mind what impact these will have on Grok machines.. From I understand, Grok always is compiled using the current state of the Nupic repo.. Hopefully if you tie Grok to a NuPIC (stable) release, the things will be smooth for Grok and OS community, and problems like this will be hard to happen again.. 🙏

[jcasner]

@david-ragazzi that's what's supposed to happen. The build logic is currently flawed, so instead of taking the latest stable NuPIC build, we seem to be grabbing the latest build. I think this is actually a recent regression in our pipelines, but at any rate, we're working to correct now...

[david-ragazzi]

Hi @rhyolight @jcasner @scottpurdy Have you solved Grook issue? Can you merge this?

[rhyolight]

Hey @david-ragazzi, @jcasner said he would let me know as soon as Grok is released. They are not quite there yet. I'll merge as soon as he tells me.

[jcasner]

👍 from Numenta product side now

[rhyolight]

@david-ragazzi After you resolve the merge conflicts, we are good to merge. 👍

[david-ragazzi]

Ok! Ready!

[rhyolight]

@david-ragazzi Isn't there one more thing? Remove this from the README!

export NTA_ROOTDIR=$NTA

😺

[david-ragazzi]

@david-ragazzi Isn't there one more thing? Remove this from the README!
export NTA_ROOTDIR=$NTA

It's weird.. To me it appears as removed... Please, check "file changes" tab again..

[rhyolight]

😳 uh... too early... need ☕

[david-ragazzi]

😳 uh... too early... need ☕

hehe