This reverts commit c3f94fd7f76f377e4728c715abbb8c7846e9fb25.
The specified commit is working fine for CI and development, but contains performance
degredation (solved in a follow-up pr) that we want to avoid for the next release. This
rever is temporary, and the changes will be re-instated after release.
---
#### Does this PR need a docs update or release note?
- [x] ⛔ No
#### Type of change
- [x] 🤖 Supportability/Tests
#### Test Plan
- [x] ⚡ Unit test
- [x] 💚 E2E
Drive pager usage currently showcases strong coupling between two layers: drive collection logic processing and drive api. This PR separates that coupling by moving the full item enumeration process into the
API, and letting the collection logic process the
results. This acs as both a simplification of complex code, and a clearer separation of ownership between the two layers.
A detrimental side effect of this change is that drive item enumeration has moved from page-streaming
(ie: each page is fully processed before moving on to the next) and onto batch processing (ie: all items are stored in memory and processed in a single pass). Acknowledging that this is an unacceptable regression, a follow-up PR will appear shortly with better handling for stream-processing enumeration from the API layer as a standard part of the pattern for all pager
implementations.
---
#### Does this PR need a docs update or release note?
- [x] ⛔ No
#### Type of change
- [x] 🧹 Tech Debt/Cleanup
#### Test Plan
- [x] ⚡ Unit test
- [x] 💚 E2E
updaing the path package to the current naming convention. No logic changes.
---
#### Does this PR need a docs update or release note?
- [x] ⛔ No
#### Type of change
- [x] 🧹 Tech Debt/Cleanup
#### Test Plan
- [x] ⚡ Unit test
- [x] 💚 E2E
some code movement and removal in selectors and scopes before adding scope-to-reason. Removed code
was only being used in testing.
---
#### Does this PR need a docs update or release note?
- [x] ⛔ No
#### Type of change
- [x] 🧹 Tech Debt/Cleanup
#### Issue(s)
* #3993
#### Test Plan
- [x] ⚡ Unit test
- [x] 💚 E2E
Pull in fix for double hashing some values
----
#### Does this PR need a docs update or release note?
- [ ] ✅ Yes, it's included
- [ ] 🕐 Yes, but in a later PR
- [x] ⛔ No
#### Type of change
- [ ] 🌻 Feature
- [x] 🐛 Bugfix
- [ ] 🗺️ Documentation
- [x] 🤖 Supportability/Tests
- [ ] 💻 CI/Deployment
- [x] 🧹 Tech Debt/Cleanup
#### Issue(s)
* #3895
#### Test Plan
- [x] 💪 Manual
- [ ] ⚡ Unit test
- [ ] 💚 E2E
automatically log when we add a recoverable error or a skipped item to fault. This log will include a stack trace of the call from the location of the logged recoverable. Clues does not have a method for pulling a stack trace out of an error yet; that can be added at a future date.
---
#### Does this PR need a docs update or release note?
- [x] ⛔ No
#### Type of change
- [x] 🤖 Supportability/Tests
#### Test Plan
- [x] ⚡ Unit test
- [x] 💚 E2E
#### Does this PR need a docs update or release note?
- [x] ⛔ No
#### Type of change
- [x] 🧹 Tech Debt/Cleanup
#### Test Plan
- [x] ⚡ Unit test
- [x] 💚 E2E
Adds an new details entry field: itemRef.
This holds a stable, semi-unique identifier to the item represented by that entry.
---
#### Does this PR need a docs update or release note?
- [x] ⛔ No
#### Type of change
- [x] 🌻 Feature
#### Issue(s)
* #3027
#### Test Plan
- [x] ⚡ Unit test
- [x] 💚 E2E
Simple concept for a large footprint of change:
filters should always accept a slice of targets
to compare against, instead of a string. Currently,
we're faking a slice by way of using a const
separator anyway, so we might as well support
slices outright.
Lots of changes (simplification, mostly) in
selectors cascade from this shift. Any updates
you see here are purely to comply with the updated
filters interface.
This step will make concealing PII in selector and filter
logging significantly easier than before.
---
#### Does this PR need a docs update or release note?
- [x] ⛔ No
#### Type of change
- [x] 🤖 Supportability/Tests
#### Issue(s)
* #2024
#### Test Plan
- [x] ⚡ Unit test
Updates and corrects input aliasing according to
the following rules (in priority order):
1. if the library name is usable, use it
2. if not, alias to the package name
3. if the package name is weird, alias sensibly
4. in case of collision, alias more distant imports
5. aliases should be consistent throughout
---
#### Does this PR need a docs update or release note?
- [x] ⛔ No
#### Type of change
- [x] 🧹 Tech Debt/Cleanup
#### Issue(s)
* #1970
#### Test Plan
- [x] ⚡ Unit test
Renames a lot of the names in scopes that contain
the word "filter" with the word "info". The basis here
is to reduce confusion in the libraries. Filter as a term
is overloaded. It currently means: 1/
the set of scopes that comparisons _must_ pass
to succeed; 2/ the pkg and struct that handles
comparison logic; and 3/ the set of scopes that
compare values in details.ItemInfo instead of in
the item's storage or logical path. The third
category is the primary overload here, and is
better off with a different name.
---
#### Does this PR need a docs update or release note?
- [x] ⛔ No
#### Type of change
- [x] 🧹 Tech Debt/Cleanup
#### Test Plan
- [x] ⚡ Unit test
## Description
Change to storing OneDrive files by ID instead of their OneDrive display name. This allows delta token-based incrementals to use an exclude list to remove previously backed up items from a backup during hierarchy merging
Also updates the following:
* selectors to match on file display name instead of kopia file name
* ShortRefs for OneDrive files to update when the file display name is updated
## Does this PR need a docs update or release note?
- [ ] ✅ Yes, it's included
- [ ] 🕐 Yes, but in a later PR
- [x] ⛔ No
## Type of change
- [x] 🌻 Feature
- [ ] 🐛 Bugfix
- [ ] 🗺️ Documentation
- [ ] 🤖 Test
- [ ] 💻 CI/Deployment
- [ ] 🧹 Tech Debt/Cleanup
## Issue(s)
* closes#1535
## Test Plan
- [x] 💪 Manual
- [x] ⚡ Unit test
- [x] 💚 E2E
With onedrive storage file names being changed from the file display name to the file id, we need a more granular form of indentification when using selectors to choose which values count as matchable fields.
This change modifies the selector PathValues to return slices of strings for each category instead of a single string, and the reducer matches on any. This will allow each service to decide what values are considered equivalent (id, shortRef, a value inside the info, etc) for each property.
---
#### Does this PR need a docs update or release note?
- [x] ⛔ No
#### Type of change
- [x] 🌻 Feature
#### Issue(s)
* #2708
#### Test Plan
- [x] ⚡ Unit test
## Description
Renaming the funcs in the fault
package to be more clear about
their purpose and behavior. Largely
just find&replace changes, except
for fault.go and the fault examples.
## Does this PR need a docs update or release note?
- [x] ⛔ No
## Type of change
- [x] 🧹 Tech Debt/Cleanup
## Issue(s)
* #1970
## Test Plan
- [x] ⚡ Unit test
- [x] 💚 E2E
## Description
Adds a new reference to the details ent: location-
ref. The location holds the human-readable
version of the item's location in whatever m365
service sourced the item. Hookup is incomplete,
following PRs will fill out functionality.
Also adds a LocationPather interface to data_
collections to pass this data back and forth
between producers and consumers.
Should be safe to merge into main.
## Does this PR need a docs update or release note?
- [x] 🕐 Yes, but in a later PR
## Type of change
- [x] 🌻 Feature
- [x] 🐛 Bugfix
## Issue(s)
* #2423
## Test Plan
- [x] ⚡ Unit test
## Does this PR need a docs update or release note?
- [x] ⛔ No
## Type of change
- [x] 🧹 Tech Debt/Cleanup
## Issue(s)
* #1970
## Test Plan
- [x] ⚡ Unit test
- [x] 💚 E2E
## Description
Adds make command for code formatting.
## Does this PR need a docs update or release note?
- [ ] ✅ Yes, it's included
- [ ] 🕐 Yes, but in a later PR
- [x] ⛔ No
## Type of change
<!--- Please check the type of change your PR introduces: --->
- [ ] 🌻 Feature
- [ ] 🐛 Bugfix
- [ ] 🗺️ Documentation
- [ ] 🤖 Test
- [ ] 💻 CI/Deployment
- [x] 🧹 Tech Debt/Cleanup
## Description
Adds clues and fault handling to selectors pkg.
Some bleed upward into the CLI occured from
where the cli directly calls selectors.Reduce.
## Does this PR need a docs update or release note?
- [x] ⛔ No
## Type of change
- [x] 🧹 Tech Debt/Cleanup
## Issue(s)
* #1970
## Test Plan
- [x] ⚡ Unit test
- [x] 💚 E2E
## Description
Now that resource owners are identified via
the selector itself, rather than each scope, we
can remove the resource owner data from
scope production and data.
## Does this PR need a docs update or release note?
- [x] ⛔ No
## Type of change
- [x] 🌻 Feature
## Issue(s)
* #1617
## Test Plan
- [x] ⚡ Unit test
- [x] 💚 E2E
## Description
Checks for resource owner matches in the top
of the reduce func using the selector owners,
instead of waiting until the path match check.
## Does this PR need a docs update or release note?
- [x] ⛔ No
## Type of change
- [x] 🌻 Feature
## Issue(s)
* #1617
## Test Plan
- [x] ⚡ Unit test
- [x] 💚 E2E
## Description
replaces the common/maps.go code with the
golang experimental maps package. The
maps package provides standard api
such as Keys, Values, Copy, and Clone.
https://pkg.go.dev/golang.org/x/exp/maps
## Does this PR need a docs update or release note?
- [x] ⛔ No
## Type of change
- [x] 🐹 Trivial/Minor
## Test Plan
- [x] ⚡ Unit test
## Description
Adds webUrl scopes to the sharepoint selector.
Also introduces the idea of a scope categories
that can broadly match across all leaf types
within a service. This category union should be
reserved for root categories, and properties
that can be used interchangably with the root.
This is part 2 of exposing webURLs as an alternative
to siteIDs for sharepoint backup and restore. The
next change providing a webURL => siteID lookup
within the graph package.
## Type of change
- [x] 🌻 Feature
## Issue(s)
* #1616
## Test Plan
- [x] ⚡ Unit test
## Description
Adds a method to all scopes that produces a
discrete clone of the original scope. The result is a shallow clone of the scope, thus retaining all
original values, but is made discrete by replacing the resource owner target with a specific value.
## Type of change
- [x] 🌻 Feature
## Issue(s)
* #1621
## Test Plan
- [x] ⚡ Unit test
## Description
For folder-level scopes (ie, scopes that compare
folder-hierarchy path segments), this change
replaces the standard "equals" and "prefix"
string comparators with the new PathContains
and PathPrefix comparators.
Next change is to interpret user inputs in the
cli to determine whether the comparator should
use contains or prefix behavior.
## Type of change
- [x] 🌻 Feature
## Issue(s)
* #1224
## Test Plan
- [x] ⚡ Unit test
- [x] 💚 E2E
## Description
Add aws xray observability tooling, plus ways to quickly and
easily run the tools/load tests. Next steps are 1/ running the
daemon container alongside the load tests within github, and
2/ creating client credentials in aws's xray service so that
gathered trace info is propagated from the daemon proxy
to the xray service.
## Type of change
- [x] 🤖 Test
- [x] 💻 CI/Deployment
## Issue(s)
* #902
## Test Plan
- [x] 💪 Manual
- [x] 💚 E2E
## Description
Builds a scope.Category() -> path.CategoryType trans- former func into scopes to standardize the relationship between the two properties.
## Type of change
- [x] 🌻 Feature
## Issue(s)
* #1133
## Test Plan
- [x] ⚡ Unit test
## Description
adds extensible options to folder-level scopes that allows the caller to specify whether they want a
prefix-comparison matcher or a contains-comparison matcher.
Also corrects the behavior of the prefix filter so that it accurately follows the "target is prefix of input" specification, rather than the reverse.
## Type of change
- [x] 🌻 Feature
## Issue(s)
* #1133
## Test Plan
- [x] ⚡ Unit test
- [x] 💚 E2E
## Description
Adds some go runtime diagnostics tracking to
load testing, including some trace regioning.
Unfortunately, I couldn't find any third party trace library that didn't depend on a sidecar server
to sample against the application on. Therefore,
just starting with something basic.
## Type of change
- [x] 🤖 Test
## Issue(s)
* #902
## Test Plan
- [x] 💪 Manual
- [x] 💚 E2E
## Description
Fixes an issue with the path matching logic in selectors where if the path was a substring of what was
specified as the scope in the filter, the item would still be matched.
e.g.
Given 2 items - `/fold/contact1` and `/folderA/folderB/contact2` and a
selector `er.Include(er.ContactFolders("AnyUser", []string{"folderA/folderB"}))`,
the selector would match both items because `fold` is contained in the selector scope `folderA/folderB`
The fix is to invert the comparison - we check if the selector scope is contained in the item path instead.
In the example above, the selector string `folderA/folderB` will then only match the second items path.
## Type of change
<!--- Please check the type of change your PR introduces: --->
- [ ] 🌻 Feature
- [x] 🐛 Bugfix
- [ ] 🗺️ Documentation
- [ ] 🤖 Test
- [ ] 💻 CI/Deployment
- [ ] 🐹 Trivial/Minor
## Test Plan
<!-- How will this be tested prior to merging.-->
- [ ] 💪 Manual
- [x] ⚡ Unit test
- [ ] 💚 E2E
## Description
Moves the `path` package to the `pkg` package so other code outside of Corso can use it if they need it
## Type of change
<!--- Please check the type of change your PR introduces: --->
- [ ] 🌻 Feature
- [ ] 🐛 Bugfix
- [ ] 🗺️ Documentation
- [ ] 🤖 Test
- [ ] 💻 CI/Deployment
- [x] 🐹 Trivial/Minor
## Issue(s)
* closes#908
## Test Plan
<!-- How will this be tested prior to merging.-->
- [ ] 💪 Manual
- [x] ⚡ Unit test
- [x] 💚 E2E
## Description
Over-restrictive scope correlation caused the
reduce processor to only include filters which
matched the data type of the scope. Ironically,
this allowed a superset of information to match,
by evading the _all-match_ expectations of
filter scopes.
Also replaces the data category consts inside /details
with the path category types, since those are acting as
better canonical owners of data type identification
throughout the app.
## Type of change
- [x] 🐛 Bugfix
## Issue(s)
* #890
## Test Plan
- [x] 💪 Manual
- [x] ⚡ Unit test
## Description
Other components may need to rebuild the directory hierarchy of items. As the paths Corso deals with can be hard to properly parse at times, store that information in the Corso backup details. The hierarchy can be rebuilt by following the `ParentRef` fields of items. The item at the root of the hierarchy has an empty `ParentRef` field.
Also hide these folders from end-users. They are not displayed during backup list nor are they eligible as a target for restore
## Type of change
<!--- Please check the type of change your PR introduces: --->
- [x] 🌻 Feature
- [ ] 🐛 Bugfix
- [ ] 🗺️ Documentation
- [ ] 🤖 Test
- [ ] 💻 CI/Deployment
- [ ] 🐹 Trivial/Minor
## Issue(s)
<!-- Can reference multiple issues. Use one of the following "magic words" - "closes, fixes" to auto-close the Github issue. -->
* closes#862
* closes#861
* closes#818
merge after:
* #869
## Test Plan
<!-- How will this be tested prior to merging.-->
- [ ] 💪 Manual
- [x] ⚡ Unit test
- [x] 💚 E2E
## Description
The details entry shorthash can be treated as
equal to the leaf item ID. This adds support to
the selector reduce step to compare the leaf val
to either the path item ID or the shortHash.
## Type of change
- [x] 🌻 Feature
## Issue(s)
* closes#572
## Test Plan
- [ ] 💪 Manual
- [x] ⚡ Unit test
- [x] 💚 E2E
* Have exchange data collection store path.Path
Still complies with the old FullPath() string interface until we update
that.
* Pass path.Path to NewCollection for exchange
Basically fixes up errors introduced by previous commit.
* Fixup exchange recovery path indices
All exchange paths now use the path struct, meaning the service,
category, and user elements are in the standard positions.
* use path package in selector reduction (#822)
Currently, during a reduction process, scopes
compare their values to the raw split on repoRef.
This causes some brittle indexing to retrieve
values from the rr, carrying assumptions that are
difficult to track across changes. This PR trades
the string split for the paths package to better
integrate identification of the path values.
Adds some mocks and amends some error
behaviors in order to fit paths into the current
testing schema.
Co-authored-by: Keepers <ryanfkeepers@gmail.com>
* Remove pathType const
The path package const only provides information on the category, not
the service for the path.
* Fix var name/package name clashes
## Description
Adds foundational cli integration tests for backup/restore
of events and calendars
## Type of change
- [x] 🤖 Test
## Issue(s)
#501
## Test Plan
- [ ] 💪 Manual
- [ ] ⚡ Unit test
- [x] 💚 E2E
swaps the corso go module from github.com/
alcionai/corso to github.com/alcionai/corso/src
to align with the location of the go.mod and
go.sum files inside the repo.
All other changes in the repository update the
package imports to the new module path.
The path package changed the standard format
of fullPath and repoRef design. This should have
failed tests before being pushed to main, but was
able to slip in falsely while github actions were
configured to pass all tests until failed.
Will help make later PRs easier as the category will already be known.
## Description
<!-- Insert PR description-->
## Type of change
Please check the type of change your PR introduces:
- [ ] 🌻 Feature
- [ ] 🐛 Bugfix
- [ ] 🗺️ Documentation
- [ ] 🤖 Test
- [x] 🐹 Trivial/Minor
## Issue(s)
<!-- Can reference multiple issues. Use one of the following "magic words" - "closes, fixes" to auto-close the Github issue. -->
on path to:
* #456
## Test Plan
<!-- How will this be tested prior to merging.-->
- [ ] 💪 Manual
- [x] ⚡ Unit test
- [ ] 💚 E2E
The filters package allows callers to specify both a target
to match on, and behavior of the comparison. While data-
type scopes always equate to "equals", the control over
different comparison behavior is useful for info-type
filters. This change integrates filters into scopes for
built-in control of those comparisons.
Though enumerated as ints, the selector service
categories are stored and managed as strings. The
only time that we use the int is when passing their
iota const to and from functions. Using a string
type instead of an int allows us to get rid of the
string builders and AtoI funcs (though not the
stringer requirement in the Iface, sadly).
* Fix wsl lint errors in pkg package
* Fix wsl lint errors in most of internal package
Leave some sub-packages out that have higher churn at the moment.
Scope production was still using service-type specific
factories instead of a generic factory set. This has been
centralized so that all service instances share the same
scope production concerns.
Additionally, category comparator funcs now use generic
comparators as well, which allows for the removal of the
isType() and includesType() comparator funcs.
Scope filtering is currently hardcoded to the exchange
use case. In order for future work to rely on boilerplate
rather than re-writing the full filtering logic on each new
type, as much of that code as is possible has been moved
into a generic toolset.
Centralizes as many of the exchange scope funcs as
possible into scopes.go. Ensures exchangeScopes comply
with the scoper interface. Reshuffles some test helper
code in selectors to a centralized file.
Introduces selectors/scopes.go, which is the base file for
managing service-agnostic scope logic. Funcs like "isAny",
"contains", "getValue" and etc are expected to be housed
here, so that service instances only need to provide thin
wrappers, mostly typecasting, around their behavior.
