LittleFS: Update to v1.7.2

[geky]

Release notes

v1.7.2: https://github.com/ARMmbed/littlefs/releases/tag/v1.7.2
v1.7: https://github.com/ARMmbed/littlefs/releases/tag/v1.7.0

Note, the release notes for patches has improved since this release here.

Changes

7e110b4 Added automatic version prefixing to releases
7f7b733 Added scripts/prefix.py for automatically prefixing version numbers
22b0456 Add missing word (and reflow text)
ec4d8b6 Changed release script to generate drafts
c7894a6 Added a handful of links to related projects
1950758 Added 2GiB file size limit and EFBIG reporting
97d8d5e Fixed issue where a rename causes a split and pushes dir out of sync
0bb1f7a Modified release script to create notes only on minor releases
28d2d96 Fix -Wsign-compare error
cb62bf2 Fixed release script issue with fetching recent tags
646b1b5 Added -Wjump-misses-init and fixed uninitialized warnings
e5a6938 Fixed possible infinite loop in deorphan step
6ad544f If stats file doesn't exist lfs_emubd_create will fail.
3419284 Fixed issue with corruption due to different cache sizes

Pull request type

[ ] Fix
[ ] Refactor
[ ] Target update
[x] Functionality change
[ ] Docs update
[ ] Test update
[ ] Breaking change

Reviewers

@ARMmbed/mbed-os-storage

[geky]

@adbridge, I'm working on a script that will create these PRs automatically. Does this format work for your release scripts? Feedback is welcome.

[ciarmcom]

@geky, thank you for your changes.
@ARMmbed/mbed-os-storage @ARMmbed/mbed-os-maintainers please review.

[adbridge]

@geky in the current patching scheme then those merge commits would cause a lot of problems with git cherry-pick or git am. However once we move over to the new releasing/branching scheme in January when we release from Master then this would work fine.

[geky]

@adbridge, do the merges cause a problem for minor releases? I thought the cherry-picking only happens on patch releases.

Unfortunately those merges are necessary for subtrees to work from what I understand.

[adbridge]

@geky no, only a problem for patch releases currently. Not a problem for Minor/Feature releases and as I said above, won't be a problem at all from January onwards :)

[geky]

Sounds like for now I'll keep the auto-generated PRs on minor releases. Patches can be cherry-picked as needed until January : )

[bulislaw]

@jamesbeyond do we run storage test that use littlefs on PR? If not can we run nightly on this?

[jamesbeyond]

@jamesbeyond do we run storage test that use littlefs on PR? If not can we run nightly on this?

These are not greentea test. I think it is run as part of Travis job bellow travis-ci/littlefs

[jamesbeyond]

Oh, actually, there are equivalent version of Greentea test as well,
@geky could you update those test together as well?
https://github.com/ARMmbed/mbed-os/blob/master/features/storage/filesystem/littlefs/TESTS

[SeppoTakalo]

Why "needs: Work" label?

Git subtree merges have been OK with Nanostack team for years, should not be problem here.

[jamesbeyond]

please update Greentea test accordingly @geky
https://github.com/ARMmbed/mbed-os/blob/master/features/storage/filesystem/littlefs/TESTS

[pilotak]

@geky why update to v1.7.2 when there is v2.1.2 available?

[adbridge]

Why "needs: Work" label?

Git subtree merges have been OK with Nanostack team for years, should not be problem here.

Firstly the needs work is due to the comments made by @jamesbeyond . Secondly subtree merges are not handled in the automated release script and we do not have the time to do all these manually.

[0xc0170]

Changes appear to be fixes, what is here functionality change?

Please add Release notes section , see https://os.mbed.com/docs/mbed-os/v5.14/contributing/workflow.html#pull-request-template-functional-changes

[geky]

Changes appear to be fixes, what is here functionality change?

I marked it as a functionality change to suggest it should be on a minor release, otherwise it would break the release script due to subtree merges.

Please add Release notes section

There is a release note section, but I moved it to the top since it seemed like the most important part of the PR body, does the order cause problems?

[geky]

Oh, actually, there are equivalent version of Greentea test as well,
@geky could you update those test together as well?

@jamesbeyond, sorry I haven't been able to keep up with GitHub notifications.

The Greentea tests were based on the local littlefs tests, but they're not in sync. The local tests can manipulate the filesystem from a higher-level than is available in Greentea:
https://github.com/ARMmbed/mbed-os/blob/master/features/storage/filesystem/littlefs/littlefs/tests/test_orphan.sh#L20

I did actually have a script for converting the local tests -> Greentea tests, but it's broken for various reasons:
https://github.com/ARMmbed/mbed-os/tree/master/features/storage/filesystem/littlefs/TESTS/util

I get that it would be valuable to duplicate the tests into Greentea, but I can't right now. There's other tasks in littlefs that have my priority.

They are at least ran in qemu-arm in the littlefs repo:
https://github.com/ARMmbed/littlefs/blob/master/.travis.yml#L73-L86

[geky]

@geky why update to v1.7.2 when there is v2.1.2 available?

Because v2 is a breaking change and Mbed OS users have existing devices with v1. There's a v1 -> v2 migration but it comes with a risk and requires testing effort.

There is https://github.com/armmbed/mbed-littlefs (sorry it's still pending an update), and I'm hoping to merge it into Mbed OS, but we'll still have v1 around.

At least the script to manage updates should work for both versions.

[adbridge]

@geky we need this PR to be merged within the next 3 works so would you have time to update this by then? Or is there somebody else that could help ?

Alternatively could you add this :
https://github.com/ARMmbed/littlefs/blob/master/.travis.yml#L73-L86
to our travis so that it runs against all PRs ? Including this one ?

[jamesbeyond]

Hi @geky , totally understand. It's going to require engineer effort to get them in sync.

I can see there are no corresponding alloc and truncate Greentea test.
Maybe could you help with most straightforward one :
https://github.com/ARMmbed/mbed-os/blob/master/features/storage/filesystem/littlefs/TESTS/filesystem/seek/main.cpp#L538

int -> unsigned, that would help as well 😉

[0xc0170]

@jamesbeyond As this PR bringing fixes, tests update can happen separately or this is blocked on them?

[geky]

@adbridge, oh, the local tests are already running in Mbed OS's travis:
https://travis-ci.org/ARMmbed/mbed-os/jobs/593092561

Are you wanting the tests to run under qemu-arm in Thumb mode as well? There was some reason we didn't want qemu in CI, but if that's not an issue I can add it in.

@jamesbeyond, the problem is that the filesystem API's are different enough between the littlefs C api <-> Mbed OS C++, that translating between the two is non-trivial.

[0xc0170]

@jamesbeyond Did @geky answer testing request? Do we need to do anything else for this PR?

[jamesbeyond]

I can see the Littlefs Test is running in the Travis.
As @geky confirm that the LittleFS c API differ from mbed-os API, I am OK for leave the greentea test as it

[adbridge]

CI started

[mbed-ci]

Test run: SUCCESS

Summary: 11 of 11 test jobs passed
Build number : 1
Build artifacts

[geky]

If someone hypothetically wanted to cherry-pick these changes onto some sort of, I don't know, patch release. These commands should work:

git cherry-pick -m 1 55f961b931a32c941cdba3fd4771c57bd453563d --allow-empty
git cherry-pick -m 1 de8df75ec0b50c772629670534bf264aae96af48

[40Grit]

@AGlass0fMilk a wild @geky appeared ^^^

[geky]

@40Grit, hello, was someone looking for me? Sorry I've not been on top of GitHub notifications for a while.

[0Grit]

@AGlass0fMilk has something in your email inbox I believe.