Summary:
This switches to using stdin/stdout for RPC calls to the sandbox, rather than specially allocated side channels. Plain text error information remains on stderr.
The motivation for the change is to simplify use of sandboxes, some of which support extra file descriptors and some of which don't.
The new style of communication is made the default, but I'm not committed to this, just that it be easy to switch to if needed. It is possible I'll need to switch the communication method again in the near future.
One reason not to make this default would be windows support, which is likely broken since stdin/stdout are by default in text mode.
Test Plan: existing tests pass
Reviewers: dsagal, alexmojaki
Reviewed By: dsagal, alexmojaki
Differential Revision: https://phab.getgrist.com/D2897
Summary:
Includes overhauled choice configuration UI for choice and choice list
columns based on the TokenField library. Features include rich copy
and paste support, keyboard shortcuts for token manipulation, and
drag-and-drop support for arrangement.
Configured choice colors are visible throughout the application, such
as in the autocomplete window for both choice and choice list cells, and
in table cells directly.
Choice cells in particular are now styled closer to choice list cells,
and render their contents as colored tokens. Choice cells now also
use the improved autocomplete component that choice lists use, with
some room for future improvement (e.g. allowing new choice items to be
added inline like in choice list's autocomplete).
Also includes a minor fix for choice list cells where right align
was not working.
Test Plan: Browser tests updated.
Reviewers: jarek, dsagal
Reviewed By: jarek, dsagal
Subscribers: jarek
Differential Revision: https://phab.getgrist.com/D2890
Summary:
- Normally Reference columns can only be used for data entry once the target
table has loaded. When it shows RowID, we shouldn't need to wait.
- Also, fix pasting values between cells of a RowID-showing column.
Test Plan: Added a test for entering data before data has loaded.
Reviewers: paulfitz
Reviewed By: paulfitz
Differential Revision: https://phab.getgrist.com/D2902
Summary:
Branding feedback from AppSumo found a capitalization problem. They also nudged us again to include a link back for the user to manage
their AppSumo account.
Test Plan: manual
Reviewers: dsagal, anaisconce
Reviewed By: dsagal
Subscribers: dsagal
Differential Revision: https://phab.getgrist.com/D2901
Summary:
Does the UI only no backend.
Follow up work:
- Implement a way to remember when a user dimsmis the popups, so
that we don't show her again.
- After users clicks Finish adds a final popup saying "You can repeat this tour from the Help Center" , and in help center home page, have a link "Repeat Grist welcome tour", which opens, say, https://docs.getgrist.com/doc/lightweight-crm#repeat-welcome-tour, where the hash part tells us to repeat the tour.
Test Plan: Tested in project/OnBoardingPopups
Reviewers: paulfitz
Reviewed By: paulfitz
Differential Revision: https://phab.getgrist.com/D2892
Summary: Removed test/aws/, most of app/server/lib/, 3 dirs in app/lambda/, corresponding tests, and more!
Test Plan: a lot of this is quite the opposite...
Reviewers: dsagal, paulfitz
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D2894
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:
mkpydocs.py has been moved to the grist-help repo where it belongs.
Deleting all localstack and test/aws stuff is coming soon - not sure if it should be in this same diff.
Test Plan: this
Reviewers: dsagal
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D2891
Summary:
Run JS with a value for SANDBOX_BUFFERS_DIR, then run test_replay in python with the same value to replay just the python code.
See test_replay.py for more info.
Test Plan:
Record some data, e.g. `SANDBOX_BUFFERS_DIR=manual npm start` or `SANDBOX_BUFFERS_DIR=server ./test/testrun.sh server`.
Then run `SANDBOX_BUFFERS_DIR=server python -m unittest test_replay` from within `core/sandbox/grist` to replay the input from the JS.
Sample of the output will look like this:
```
Checking /tmp/sandbox_buffers/server/2021-06-16T15:13:59.958Z
True
Checking /tmp/sandbox_buffers/server/2021-06-16T15:16:37.170Z
True
Checking /tmp/sandbox_buffers/server/2021-06-16T15:14:22.378Z
True
```
Reviewers: paulfitz, dsagal
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D2866
Summary:
- Implement UI with "Apply to new records" and "Apply on record changes"
checkboxes, and options for selecting which changes to recalculate on.
- For consistency, always represent empty RefList as None
- Fix up generated SchemaTypes to remember that values are encoded.
Included test cases for the main planned use cases:
- Auto-filled UUID column
- Data cleaning
- NOW() formula for record's last-updated timestamp.
- Updates that depend on other columns.
Test Plan: Added a browser test.
Reviewers: jarek
Reviewed By: jarek
Subscribers: paulfitz
Differential Revision: https://phab.getgrist.com/D2885
Summary:
User was not able to delete cards. This patch introduces a context menu for cards, analogous to the one available for rows on a GridView.
Changes:
- Row numbers on a GridView have the same icon as on columns to make context menu more discoverable.
- Context menu for rows and columns, when activated, didn't switch section in rare conditions (i.e. when the section had 2 or more columns selected, one of which had the same rowId as a column in the section that the user switched from).
- Card list layout and a single card layout has the same context menu as in a GridView, available by pressing the context menu button.
Test Plan: Browser tests
Reviewers: dsagal, paulfitz
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D2870
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:
Current appsumo sign-up flow doesn't reach the billing pages.
This diff nudges user on through that extra step.
It also tweaks plan summaries to say what special appsumo
features are in effect (member count prepaid for).
Test Plan: manual
Reviewers: dsagal
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D2882
Summary:
This adds a new landing page for cognito sign-up, intended for
use by new appsumo users.
Their email address is pre-filled and locked down, and sign-up
is by entering a password.
The page is very crude compared to hosted cognito - especially
in error reporting! - but having the address filled in more
than makes up for that.
The flow does not quite connect with the new billing signup.
I think we can do that through the regular "welcome" process,
which will list the user's team site. When the user visits
that site, we could detect that we are on a site with no
domain set yet and for which the user is a billing manager,
and trigger a visit to the appropriate billing page.
Test Plan: manual - hard to test through cognito email step
Reviewers: dsagal
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D2880
Summary:
This links AppSumo sign-ups with Stripe subscriptions
and our billing pages. Different AppSumo tiers are supported by
different coupons on the standard plan. Configuration of this
is in stripe, and then cached in the database.
The front end is tweaked just enough to make completing a sign-up
possible. It is not yet friendly.
Not covered includes:
* Streamlining landing page.
* Making billing pages git clearer summaries of AppSumo states.
* Making flow through Cognito as graceful as possible - default
probably doesn't meet AppSumo requirements.
* Disabling site on cancellation/refund.
* Downgrades when more seats in use than lower tier allows.
Test Plan: api-level tests added. No front-end tests yet.
Reviewers: dsagal
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D2878
Test Plan: Enhanced the test case for memos to check these cases too (fails without this fix).
Reviewers: paulfitz
Reviewed By: paulfitz
Differential Revision: https://phab.getgrist.com/D2876
Summary:
Makes filter counts take other column filters into account.
- Changes the summaries rows to reflect hidden rows:
- hidden rows are added to the `Other Values` summary
- show the unique number of other values as `Other Values (12)`
- Also, adds a sort button to the column filter menu
Test Plan: Adds browser test.
Reviewers: paulfitz, jarek
Reviewed By: jarek
Differential Revision: https://phab.getgrist.com/D2861
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:
This adds appsumo /token and /notification endpoints, with some
tests. The stub implementation is sufficient for AppSumo
activation to succeed (when exposed via port forwarding for testing).
It needs fleshing out:
* Implement upgrade/downgrade/refund and stripe subscription.
* Implement custom landing page and flow.
Test Plan: added tests
Reviewers: dsagal, georgegevoian
Reviewed By: dsagal
Subscribers: alexmojaki
Differential Revision: https://phab.getgrist.com/D2864
Summary: Convert recursion to while loop with stack of things to invalidate.
Test Plan: Loan tracker example works now
Reviewers: dsagal
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D2867
Summary:
User can freeze any number of columns, which will not move when a user scrolls grid horizontally.
Main use cases:
- Frozen columns don't move when a user scrolls horizontally
- The number of frozen columns is automatically persisted
- Readonly viewers see frozen columns and can modify them - but the change is not persisted
- On a small screen - frozen columns still moves to the left when scrolled, to reveal at least one column
- There is a single menu option - Toggle freeze - which offers the best action considering selected columns
- When a user clicks a single column - action to freeze/unfreeze is always there
- When a user clicks multiple columns - action is offered only where it makes sens (columns are near the frozen border)
Test Plan: Browser tests
Reviewers: dsagal, paulfitz
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D2852
Summary:
Grist should not prevent read-only viewers from opening cell editors since they usually provide much more information than is visible in a cell.
Every editor was enhanced with a read-only mode that provides the same information available for an editor but doesn't allow to change the underlying data.
Test Plan: Browser tests
Reviewers: dsagal
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D2842
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:
Replaces https://phab.getgrist.com/D2854
Refactoring of NSandbox:
- Simplify arguments to NSandbox.spawn. Only half the arguments were used depending on the flavour, adding a layer of confusion.
- Ensure the same environment variables are passed to both flavours of sandbox
- Simplify passing down environment variables.
Implement deterministic mode with libfaketime and a seeded random instance.
- Include static prebuilt libfaketime.so.1, may need another solution in future for other platforms.
Recording pycalls:
- Add script recordDocumentPyCalls.js to open a single document outside of tests.
- Refactor out recordPyCalls.ts to support various uses.
- Add afterEach hook to save all pycalls from server tests under $PYCALLS_DIR
- Make docTools usable without mocha.
- Add useLocalDoc and loadLocalDoc for loading non-fixture documents
Test Plan:
Made a document with formulas NOW() and UUID()
Compare two document openings in normal mode:
diff <(test/recordDocumentPyCalls.js samples/d4W6NrzCMNVSVD6nWgNrGC.grist /dev/stdout) \
<(test/recordDocumentPyCalls.js samples/d4W6NrzCMNVSVD6nWgNrGC.grist /dev/stdout)
Output:
< 1623407499.58132,
---
> 1623407499.60376,
1195c1195
< "B": "bd2487f6-63c9-4f02-bbbc-5c0d674a2dc6"
---
> "B": "22e1a4fd-297f-4b86-91a2-bc42cc6da4b2"
`export DETERMINISTIC_MODE=1` and repeat. diff is empty!
Reviewers: paulfitz
Reviewed By: paulfitz
Differential Revision: https://phab.getgrist.com/D2857
Summary:
The loading spinner would always display
'Building Table widget' when creating or changing
a widget. This fixes the title to reflect the selected
widget type.
Test Plan:
Updated existing browser tests to verify the loading spinner
title includes the correct widget type.
Reviewers: dsagal, paulfitz
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D2863
Summary:
Previously, soft-deleted docs in icon view were still accessible from
the Trash and couldn't be permanently deleted.
Test Plan:
Improved the nbrowser test for deleting docs to verify that it can
be done in both view modes.
Reviewers: dsagal, paulfitz
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D2862
Summary: Summaries test works only in a headless mode. In browser mode it failes as section is to narrow.
Test Plan: n/a
Reviewers: dsagal
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D2859
Summary: The `select-all` originally designed the `All` button but it unfortunately evolves into designating both `All` and `None`. Which is confusing. Replaces with the more general `bulk-action`.
Test Plan: Should not break anything.
Reviewers: paulfitz
Reviewed By: paulfitz
Differential Revision: https://phab.getgrist.com/D2860
Test Plan: Wrote unit and browser tests that verify new behavior.
Reviewers: paulfitz, dsagal
Reviewed By: dsagal
Subscribers: alexmojaki
Differential Revision: https://phab.getgrist.com/D2855
Summary:
- Takes advantage of native indeterminate state of html checkboxes
- When an indeterminate checkbox is clicked it turns it into being not checked.
Test Plan: - Added test to projects/UI2018
Reviewers: paulfitz
Reviewed By: paulfitz
Differential Revision: https://phab.getgrist.com/D2846
Summary:
This fixes a bug where deleting a page with the page id
missing from the URL would cause JS errors to be thrown.
Test Plan:
Verified manually in dev environment. Browser test added
that should hopefully replicate the repro steps and catch
any regressions.
Reviewers: dsagal, paulfitz
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D2856
Summary:
This is a somewhat experimental change, that will implement permitted parts of an undo if not all parts are permitted. This is in preparation for trigger columns, where it may become common for a change in a record resulting in a change to an automatic change to another that the user cannot edit directly. How to undo such an action is somewhat unclear. One option is to undo the permitted parts, and then the triggers can rerun.
The general case is a bit of a can of worms, and feels adjacent to merging/rebasing etc.
Oh: it would probably be important in general to communicate to the user that an undo was partial, but this diff doesn't do that. It would need some new plumbing.
Test Plan: added test
Reviewers: dsagal
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D2839
Summary: Cursor position observable was created using GrainJS, but the fields it was using were created using knockout observables. In a result the cursor position wasn't recomputed when a view was changed or an active section was deleted.
Test Plan: Browser tests
Reviewers: dsagal
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D2850
Summary: Records all sandbox pycall arguments and results, saves them to JSON when the environment variable UPDATE_REGRESSION_DATA is set, otherwise checks that they match the saved JSON.
Test Plan: This is tests
Reviewers: dsagal, paulfitz
Reviewed By: paulfitz
Differential Revision: https://phab.getgrist.com/D2851
Test Plan:
Tested manually by enabling/disabling wrapping and changing alignment on reference columns. Existing test updated to check that
reference columns have cell formatting options available.
Reviewers: dsagal
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D2853
Summary:
Bug summary: if in right bar user starts changing name of column, but then clicks on a different column name in table, THAT column will have its name changed.
This bug occurs because the save method is invoked by a blur event on a input field, which is triggered after all computed observables are calculated. Save method gets an observable to update, which by the time a blur event triggers, is changed to a new column.
The solution was to forcefully trigger the blur event as soon as possible - here by subscribing to the cursor position observable.
Test Plan: Browser tests
Reviewers: dsagal
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D2845
Test Plan: Block read access to column A based on the condition rec.B == 1. Then setting B = 1 in a row makes the cell under A grey.
Reviewers: dsagal
Reviewed By: dsagal
Subscribers: paulfitz, dsagal
Differential Revision: https://phab.getgrist.com/D2828
Summary:
A recently added stress test ("deletes documents reasonably quickly" in removedAt.ts) is sporadically failing under postgres. It looks like typeorm's .save() method is in some way unreliable when setting a table with multi-column primary keys, via a ManyToMany relation. This diff replaces the .save() with explicit inserts/deletes.
I modified _repairWorkspaceGuests recently, so thought that change might have been the problem. However under the stress test, failures occur as often in _repairOrgGuests (not changed recently) as in _repairWorkspaceGuests (changed recently).
For reference, see schema diagram at https://grist.quip.com/wWpRAMe058Nl/Home-DB (the table being updated is `group_users`).
Possibly related issue: https://github.com/typeorm/typeorm/issues/4122
Test Plan:
After this change, stress test runs well on postgres locally (no failure 70 iterations); before it would fail on postgres within 3 iterations typically.
Separately: I gave a test that failed a little more time to return, and confirmed it was no slower on average, so I think it was unrelated.
Reviewers: jarek
Reviewed By: jarek
Differential Revision: https://phab.getgrist.com/D2848
Summary:
The `_repairWorkspaceGuests` method is slow for workspaces with large numbers of documents. It makes a query that produces a lot of rows. The query itself is tolerable, but TypeORM processing uses enough CPU to be a likely culprit in some production instability. This diff splits the query into two pieces that are logically independent, but which when combined were resulting in the number of rows being the product of the two pieces. Once split, there is also a where clause that can be applied to one of the pieces.
The purpose of the method is to add every user that a document within a workspace is shared with to a "guest" group of the workspace itself. The design of "guest" groups is not ideal, but this diff leaves the design unchanged and is intended only to speed up operation.
Made some small tweaks to the timing of a flakey test, and temporarily recreated the `samples` directory removed in a previous diff (this is currently breaking tests badly on a fresh worker without a `samples` directory lying around)
Test Plan: added test; existing tests pass
Reviewers: jarek
Reviewed By: jarek
Differential Revision: https://phab.getgrist.com/D2844
Summary:
- this is a core search code refactoring
- this diff should fix the js error that was happening when searching across pages.
Test Plan: Tested manually on dev's environment. Tests shows no regression and successfully fixes the js error.
Reviewers: paulfitz
Reviewed By: paulfitz
Differential Revision: https://phab.getgrist.com/D2837
Summary:
This cleans up a few things about SELF_HYPERLINK urls:
* Use `urlId` rather than `docId`.
* Correctly merge personal org subdomain.
* In dev environment, use clearer port number.
Test Plan: updated test
Reviewers: alexmojaki, dsagal
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D2841
Summary:
Names of private tables and columns were leaking via Code View.
This plugs that leak.
Test Plan: adds test
Reviewers: dsagal
Reviewed By: dsagal
Differential Revision: https://phab.getgrist.com/D2840
