VFS additions

[phatpaul]

Adding a vfs_upload example GUI. Depends on chmorgan/libesphttpd#48 (already merged)

Also add helper script to generate filesystem images (stolen from loboris/MicroPython_ESP32_psRAM_LoBo).
(I only got FatFS to work so far, and I don't have time right now to troubleshoot SPIFFS and LittleFS. I put a disclaimer in the readme to this effect)

The example still builds and runs fine without any of this VFS stuff turned on. (the filesystem upload functions just fail if used)

[chmorgan]

Love the concept here but the second commit, the one that pulls in a ton of files from other repositories would be much smaller if we made use of submodules instead. Is that possible?

[chmorgan]

possible to add littlefs as a submodule vs. copying all of the files into the repository?

[phatpaul]

Would love to, can you help figure out how?

[chmorgan]

any reason not to use the main GitHub repository for mkspiffs? Igrr's?

[phatpaul]

I modified that project to get it to work as a submodule in ESP-IDF. I don't know if there is a cleaner way.
phatpaul/mkspiffs@04d419f
Thoughts?

[chmorgan]

@phatpaul maybe reach out to igrr to see what he thinks? It may be that we can perform the configuration parts as a part of the build inside of this repository.

[igrr]

The commit linked above seems to be the correct approach, if you want to include mkspiffs as a submodule.

[chmorgan]

@igrr @phatpaul had to add this commit, phatpaul/mkspiffs@04d419f

Would that commit be accepted into your tree if submitted as a PR?

[igrr]

Yes, it would.

[chmorgan]

@phatpaul can you submit that commit as a PR? I tried but GitHub is reporting an error, probably because it requires some permissions on your fork.

[phatpaul]

Rebased and created PR. See igrr/mkspiffs#57

[chmorgan]

can we use mkfatfs via a submodule as well?

[phatpaul]

I would love to. I looked at doing that but couldn't figure out how.
The problem is that the source directory for that project is several layers deep.
https://github.com/jkearins/ESP32_mkfatfs/tree/master/components/mkfatfs/src

And it is my understanding that you can't check-out a sub-directory of a repo as a submodule.

Ideas?

[chmorgan]

@phatpaul thoughts about asking jkearins to split that off into a separate repository so other people can reuse it? You are correct that git doesn't allow what we need because he embedded his component into the subdirectory. There are some hacks to make it work like add a submodule to another directory like repos/ and then symlink from repos/mkfatfs/components/mkfatfs/src to components/mkfatfs but that seems kind of blah.

[chmorgan]

@phatpaul it may take a little while to discuss with igrr and jkearins but lets see how that goes first. If there is some impasse then we can discuss working around it. Does that sound ok?

[chmorgan]

@phatpaul have you had any luck with those guys? To clarify my concerns if we merge back changes here we'll be copying or relying on forks of external repositories. If we can get those repositories adjusted then we can include them directly and won't have to worry about maintaining forks or copied files. Would it help if I reached out to them? I'd like to do what I can to facilitate getting your PR merged in as soon as possible.

[phatpaul]

Yes please feel free help out. I have been busy at work finishing my esp32-based product.

…

On Sat, Dec 8, 2018, 8:39 PM Chris Morgan ***@***.***> wrote: @phatpaul <https://github.com/phatpaul> have you had any luck with those guys? To clarify my concerns if we merge back changes here we'll be copying or relying on forks of external repositories. If we can get those repositories adjusted then we can include them directly and won't have to worry about maintaining forks or copied files. Would it help if I reached out to them? I'd like to do what I can to facilitate getting your PR merged in as soon as possible. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#12 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ADVVdzW8TxIBtZlhdiLTc27smRmR9NPnks5u3GnMgaJpZM4Ycame> .

[chmorgan]

@jkearns, @igrr, can you take a look at the comments above? We are trying to integrate mkspiffs and mkfatfs in this project as submodules but we are having trouble doing so. Are we missing something?

[chmorgan]

@jkearns have you see the request here to split the mkfatfs into a separate repository so it can be included as a component of another project? (As least that’s what I think the request was, @phatpaul can correct me)

[chmorgan]

@phatpaul ahh ok this PR from a while back. The changes to mkspiffs were integrated by igrr right?

@jkearns just pinging again to see if you saw the request about splitting things up.

[phatpaul]

@chmorgan I will split this up to leave out the building local fs stuff.

I would like to merge this in - The UI for VFS is in here and it does not depend on the mkfatfs.

[phatpaul]

Actually it is already split into 2 commits. What's the best way to merge in just the first commit? Do I need to update the pull-request? Or rewind the branch? I would like to leave the 2nd commit in this or another pull-request.

[chmorgan]

Hi @phatpaul. I'd do something like:

git branch vfs_upload_full
git rebase -i master
git push --force

The branch makes a copy of your current branch and during the rebase you'd drop the commit you'd like to lose. The rebased branch with the single commit should show up here replacing the previous commits after the force push.