Summary:
This allows limiting the memory available to documents in the sandbox when gvisor is used. If memory limit is exceeded, we offer to open doc in recovery mode. Recovery mode is tweaked to open docs with tables in "ondemand" mode, which will generally take less memory and allow for deleting rows.
The limit is on the size of the virtual address space available to the sandbox (`RLIMIT_AS`), which in practice appears to function as one would want, and is the only practical option. There is a documented `RLIMIT_RSS` limit to `specifies the limit (in bytes) of the process's resident set (the number of virtual pages resident in RAM)` but this is no longer enforced by the kernel (neither the host nor gvisor).
When the sandbox runs out of memory, there are many ways it can fail. This diff catches all the ones I saw, but there could be more.
Test Plan: added tests
Reviewers: alexmojaki
Reviewed By: alexmojaki
Subscribers: alexmojaki
Differential Revision: https://phab.getgrist.com/D3398
Summary: Adds a special user action `UpdateCurrentTime` which invalidates an internal engine dependency node that doesn't belong to any table but is 'used' by the `NOW()` function. Applies the action automatically every hour.
Test Plan: Added a Python test for the user action. Tested the interval periodically applying the action manually: {F43312}
Reviewers: paulfitz
Reviewed By: paulfitz
Differential Revision: https://phab.getgrist.com/D3389
Summary:
This is a fix for a bug discussed in https://grist.slack.com/archives/C069RUP71/p1645138610722889
I still haven't completely wrapped my head around it or figured out how to make a simple reproducible example, but the problem seems to be that a lookup can happen before the column(s) being looked up (the summary helper column in this case) have been computed fully (I think it got interrupted halfway by an OrderError). `do_lookup` would check via `engine._use_node` that the row IDs it found had all been computed already, but there might still be other rows that hadn't been computed yet and would also have values matching the lookup key, so it missed those.
This diff instead calls `_use_node` with no `row_ids` argument, which should ensure that all rows have already been computed.
At first I was worried about how this would affect performance, which led me down an optimisation rabbit hole, hence a bit of unrelated cleanup here and also https://phab.getgrist.com/D3310 . But it doesn't seem to be a problem, and IIUC it should actually make things better, although this code is pretty confusing.
Test Plan: Tested manually that the doc no longer behaves weirdly
Reviewers: dsagal
Reviewed By: dsagal
Subscribers: dsagal, paulfitz
Differential Revision: https://phab.getgrist.com/D3308
Summary:
This is an attempt to optimise Engine._use_node. It doesn't seem to actually improve overall performance significantly, but it shouldn't make it worse, and I think it's an improvement to the code.
It turns out that there's no need to track a stack of compute frames any more. The only time we get close to nested evaluation, we set allow_evaluation=False to prevent it actually happening. So there's only one 'frame' during actual evaluation, which means we can get rid of the concept of frames entirely. This allows simplifying the code and letting the computer do less work in general.
Test Plan: this
Reviewers: dsagal
Reviewed By: dsagal
Subscribers: dsagal
Differential Revision: https://phab.getgrist.com/D3310
Summary: I ran the python tests through a profiler and found that just generating the autocomplete context was a major cost, and this would happen with every schema change. By only generating it when needed for autocomplete, the time for tests reduced from ~38s to ~33s.
Test Plan: this
Reviewers: dsagal
Reviewed By: dsagal
Subscribers: dsagal
Differential Revision: https://phab.getgrist.com/D3312
Summary:
Adds a method Table._num_rows using an empty lookup map column.
Adds a method Engine.count_rows which adds them all up.
Returns the count after applying user actions to be logged by ActiveDoc.
Test Plan: Added a unit test in Python. Tested log message manually.
Reviewers: paulfitz
Reviewed By: paulfitz
Differential Revision: https://phab.getgrist.com/D3275
Summary:
Addresses several issues:
- Error 'Cannot modify summary group-by column' when changing Text ->
ChoiceList in the presence of summary tables.
- Error 'ModifyColumn in unexpected position' when changing ChoiceList -> Text
in the presence of summary tables.
- Double-evaluation of trigger formulas in some cases.
Fixes include:
- Fixed verification that summary group-by columns match the underlying ones,
and added comments to explain.
- Avoid updating non-metadata lookups after each doc-action (early lookups
generated extra actions to populate summary tables, causing the 'ModifyColumn
in unexpected position' bug)
- When updating formulas, do update lookups first.
- Made a client-side tweak to avoid a JS error in case of some undos.
Solution to reduce lookups is based on https://phab.getgrist.com/D3069?vs=on&id=12445,
and tests for double-evaluation of trigger formulas are taken from there.
Add a new test case to protect against bugs caused by incorrect order of
evaluating #lookup columns.
Enhanced ChoiceList browser test to check a conversion scenario in the presence
of summary tables, previously triggering bugs.
Test Plan: Various tests added or enhanced.
Reviewers: alexmojaki
Reviewed By: alexmojaki
Subscribers: jarek
Differential Revision: https://phab.getgrist.com/D3184
Summary:
Record numbers of rows, columns, cells, and bytes of marshalled data for most calls to table_data_from_db
Export new function get_table_stats in the sandbox, which gives the raw numbers and totals.
Get and log these stats in ActiveDoc right after loading tables, before Calculate, so they are logged even in case of errors.
Tweak logging about number of tables, especially number of on-demand tables, to not only show in debug logging.
Test Plan: Updated doc regression tests, that shows what the data looks like nicely.
Reviewers: dsagal, paulfitz
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D3081
Summary:
Traceback is available on the Creator Panel in the formula editor. It is evaluated the same way as for normal formulas.
In case when the traceback is not available, only the error name is displayed with information that traceback is not available.
Cell with an error, when edited, shows the previous valid value that was used before the error happened (or None for new rows).
Value is stored inside the RaisedException object that is stored in a cell.
Test Plan: Created tests
Reviewers: alexmojaki
Reviewed By: alexmojaki
Subscribers: alexmojaki, dsagal
Differential Revision: https://phab.getgrist.com/D3033
Summary:
The problem: For a data-cleaning column (one that depends on itself) `doBulkUpdateRecord` calls `prevent_recalc(should_prevent=False)``
which is supposed to cause it to get calculated eventually.
But before that it calls `_do_doc_action` -> `apply_doc_action` -> `_bring_lookups_up_to_date` which recalculates
a lookup column which eventually calls `_recompute_step(allow_evaluation=False)` on the data-cleaning column
which shouldn't really do anything significant but it both modifies the set `self.recompute_map[node]`
and then removes it from the map after it's empty.
The solution: when `allow_evaluation=False` and `self._prevent_recompute_map[node]` is nonempty,
ensure `self.recompute_map[node]` is not modified,
and check the map directly (instead of `dirty_rows` which can now be separate) to see if the set is empty before cleanup.
Test Plan: Added a lookup column to test_self_trigger, ensured that this caused the test to fail without the other two changes in engine.py.
Reviewers: paulfitz
Reviewed By: paulfitz
Subscribers: dsagal
Differential Revision: https://phab.getgrist.com/D3006
Summary:
Prefix keys of `LinkingState.filterColValues` with `_contains:` when the source column is a ChoiceList or ReferenceList.
This is parsed out to make a boolean `isContainsFilter` which is kept in each value of `QueryRefs.filterTuples` (previously `filterPairs`).
Then when converting back in `convertQueryFromRefs` we construct `Query.contains: {[colId: string]: boolean}`.
Finally `getFilterFunc` uses `Query.contains` to decide what kind of filtering to do.
This is not pretty, but the existing code is already very complex and it was hard to find something that wouldn't require touching loads of code just to make things compile.
Test Plan: Added a new nbrowser test and fixture, tests that selecting a source table by summary tables grouped by a choicelist column, non-list column, and both all filter the correct data.
Reviewers: dsagal
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D2940
Summary: Deletes code which was previously only used by SharedSharing.ts, which was deleted in D2894
Test Plan: no
Reviewers: paulfitz
Reviewed By: paulfitz
Differential Revision: https://phab.getgrist.com/D2960
Summary: Remove repl.py, REPLTab.js, some wiring code, CSS, and a test in testscript.json.
Test Plan: NA
Reviewers: dsagal
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D2923
Summary:
Dealing with some things that bothered and sometimes confused me:
Make Table.Record[Set] provide the table argument automatically
Remove the classes from UserTable because they're not used anywhere and the Table/UserTable distinction is already confusing. They're not documented for users and they don't show up in autocomplete.
Remove RecordSet.Record because it was confusing me where that attribute was being set, but also this means .Record will work properly for users with columns named 'Record'.
Test Plan: existing tests
Reviewers: dsagal
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D2913
Summary:
The 'user' variable has a similar API to the one from access rules: it
contains properties about a user, such as their full name and email
address, as well as optional, user-defined attributes that are populated
via user attribute tables.
Test Plan: Python unit tests.
Reviewers: alexmojaki, paulfitz, dsagal
Reviewed By: alexmojaki, dsagal
Subscribers: paulfitz, dsagal, alexmojaki
Differential Revision: https://phab.getgrist.com/D2898
Summary:
API signature for autocomplete updated to add column ID, which is
necessary for exposing correct types for 'value'.
Test Plan: Unit tests.
Reviewers: alexmojaki
Reviewed By: alexmojaki
Subscribers: jarek, alexmojaki
Differential Revision: https://phab.getgrist.com/D2896
Summary:
Add python3 suite to testrun.sh
Build virtualenv called sandbox_venv3 in build script
Move xmlrunner.py into grist folder from thirdparty, since sandbox_venv3 doesn't use thirdparty and I don't know where that file comes from - unittest-xml-reporting is different.
Test Plan: this
Reviewers: dsagal
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D2886
Summary:
Trigger formulas can be calculated for new records, or for new records and
updates to certain fields, or all fields. They do not recalculate on open,
and they MAY be set directly by the user, including for data-cleaning.
- Column metadata now includes recalcWhen and recalcDeps fields.
- Trigger formulas are NOT recalculated on open or on schema changes.
- When recalcWhen is "never", formula isn't calculated even for new records.
- When recalcWhen is "allupdates", formula is calculated for new records and
any manual (non-formula) updates to the record.
- When recalcWhen is "", formula is calculated for new records, and changes to
recalcDeps fields (which may be formula fields or column itself).
- A column whose recalcDeps includes itself is a "data-cleaning" column; a
value set by the user will still trigger the formula.
- All trigger-formulas receive a "value" argument (to support the case above).
Small changes
- Update RefLists (used for recalcDeps) when target rows are deleted.
- Add RecordList.__contains__ (for `rec in refList` or `id in refList` checks)
- Clarify that Calculate action has replaced load_done() in practice,
and use it in tests too, to better match reality.
Left for later:
- UI for setting recalcWhen / recalcDeps.
- Implementation of actions such as "Recalculate for all cells".
- Allowing trigger-formulas access to the current user's info.
Test Plan: Added a comprehensive python-side test for various trigger combinations
Reviewers: paulfitz, alexmojaki
Reviewed By: paulfitz
Differential Revision: https://phab.getgrist.com/D2872
Summary: Changes that move towards python 3 compatibility that are easy to review without much thought
Test Plan: The tests
Reviewers: dsagal
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D2873
Summary: sys.setdefaultencoding('utf8')
Test Plan: Will test against user documents to check for changes
Reviewers: dsagal, paulfitz
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D2858
Summary:
This diff discounts indirect changes for access control purposes. A UserAction that updates a cell A, which in turn causes changes in other dependent cells, will be considered a change to cell A for access control purposes.
The `engine.apply_user_actions` method now returns a `direct` array, with a boolean for each `stored` action, set to `true` if the action is attributed to the user or `false` if it is attributed to the engine. `GranularAccess` ignores actions attributed to the engine when checking for edit rights.
Subtleties:
* Removal of references to a removed row are considered direct changes.
* Doesn't play well with undos as yet. An action that indirectly modifies a cell the user doesn't have rights to may succeed, but it will not be reversible.
Test Plan: added tests, updated tests
Reviewers: dsagal
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D2806
Summary:
An incorrect DocAction (as possible from an Undo of a non-last action)
could cause RemoveRecord on an already missing record. This used to
create an invalid undo, and wreak havoc when a series of DocActions
later fails and needs to be reverted.
To fix, consider RemoveRecord of a missing record to be a no-op.
Test Plan: Includes a new test case that triggers the problem.
Reviewers: paulfitz
Reviewed By: paulfitz
Differential Revision: https://phab.getgrist.com/D2717
Summary:
Currently, an undo of a non-last action can leave the doc in an inconsistent
state. For example, it may remove a table, but fail to remove all columns of
it from metadata. We normally check that schema corresponds to metadata, but
stray columns were not visible to this check, and instead caused later table
additions to fail.
This diff fixes the check to fail the action that causes stray columns, and
to restore the doc to a consistent state.
Note that this only handles schema-metadata inconsistencies, but an undo of a
non-last action can easily create other surprises.
Test Plan: Added a test case that triggered inconsistency before, and now triggers a failed undo.
Reviewers: paulfitz
Reviewed By: paulfitz
Differential Revision: https://phab.getgrist.com/D2715
Summary:
- When adding records, negative rowIds may now be specified. They'll be replaced by proper IDs.
- If these negative IDs are used in Reference columns in subsequent actions in
the same bundle of UserActions, they get replaced with the proper rowIds.
- Use this to sync ACLResources and ACLRules from UI in a single batch of UserActions.
- Resolve the TODOs in GranularAccess test, to no longer need to guess resource rowIds.
Test Plan: Added a python unittest for mapping IDs; updated browser tests.
Reviewers: paulfitz
Reviewed By: paulfitz
Differential Revision: https://phab.getgrist.com/D2691
Summary:
The new plans for granular access control are different and handled by
node.js. Some of the same tables will be reused, of which we never made
real use before except for expecting certain specific initial records.
This diff removes the old logic, replacing it with a stub that satisfies
the interface expected by other code.
It also removes several unused UserActions: AddUser/RemoveUser/
AddInstance/RemoveInstance.
Test Plan: Existing tests should pass.
Reviewers: paulfitz
Reviewed By: paulfitz
Differential Revision: https://phab.getgrist.com/D2662
Summary:
- Introduce a new SQLiteDB migration, which adds DB columns for formula columns
- Newly added columns have the special ['P'] (pending) value in them
(in order to show the usual "Loading..." on the first load that triggers the migration)
- Calculated values are added to .stored/.undo fields of user actions.
- Various changes made in the sandbox to include .stored/.undo in the right order.
- OnDemand tables ignore stored formula columns, replacing them with special SQL as before
- In particular, converting to OnDemand table leaves stale values in those
columns, we should maybe clean those out.
Some tweaks on the side:
- Allow overriding chai assertion truncateThreshold with CHAI_TRUNCATE_THRESHOLD
- Rebuild python automatically in watch mode
Test Plan: Fixed various tests, updated some fixtures. Many python tests that check actions needed adjustments because actions moved from .stored to .undo. Some checks added to catch situations previously only caught in browser tests.
Reviewers: paulfitz
Reviewed By: paulfitz
Differential Revision: https://phab.getgrist.com/D2645
Summary:
- Make suggestions less case-sensitive (not entirely case-insensitive, but
allow top-level suggestions to match in all-lowercase)
- Add function signatures to suggestions for Grist functions.
- Excel-like functions that are present but not implemented are no longer
offered as suggestions.
Test Plan:
Added a test case on python side, and a browser test case for how suggestions
are rendered and inserted.
Reviewers: paulfitz
Reviewed By: paulfitz
Differential Revision: https://phab.getgrist.com/D2608
Summary:
this moves sandbox/grist to core, and adds a requirements.txt
file for reconstructing the content of sandbox/thirdparty.
Test Plan:
existing tests pass.
Tested core functionality manually. Tested docker build manually.
Reviewers: dsagal
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D2563
