[FIX] evaluation: Ensure dependency invalidation for array formulas #5209
+23
−2
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
|
LucasLefevre
force-pushed
the
17.0-fix-array-recompute-deps-rar
branch
from
c31bc5f to
c57d35e
Compare
How to reproduce: -> see the attached test What is happening? During the evaluation of SUMIFS, it will fetch its ranges. First it will first fetch an empty range (E4:E7) since MUNIT hasn't been evaluated yet. Afterwards, it will fetch H4, which in turn will request C4 and therefore force the evaluation of the spreaded function MUNIT. AT this point, there is a disparity of values because the values of the range (E4:E7) have changed. To address those situations, the evaluator invalidates the dependencies of all the spreaded zone. Which means marking A1 to be recomputed. Unfortunately, A1 is still being evaluated and since #4589 (specifically commit 8d22e86) we remove the cell from the queue after we evaluted it. This revision changes the approach taken in 8d22e86 to allow legitimate invalidations take place while avoiding useless reevaluations. Task: 4252800
rrahir
force-pushed
the
17.0-fix-array-recompute-deps-rar
branch
from
84d917d to
ec22cf3
Compare
|
@rrahir @LucasLefevre because this PR has multiple commits, I need to know how to merge it:
|
|
robodoo rebase-ff |
|
Merge method set to rebase and fast-forward. |
robodoo
pushed a commit
that referenced
this pull request
Nov 15, 2024
How to reproduce: -> see the attached test What is happening? During the evaluation of SUMIFS, it will fetch its ranges. First it will first fetch an empty range (E4:E7) since MUNIT hasn't been evaluated yet. Afterwards, it will fetch H4, which in turn will request C4 and therefore force the evaluation of the spreaded function MUNIT. AT this point, there is a disparity of values because the values of the range (E4:E7) have changed. To address those situations, the evaluator invalidates the dependencies of all the spreaded zone. Which means marking A1 to be recomputed. Unfortunately, A1 is still being evaluated and since #4589 (specifically commit 8d22e86) we remove the cell from the queue after we evaluted it. This revision changes the approach taken in 8d22e86 to allow legitimate invalidations take place while avoiding useless reevaluations. Task: 4252800 Part-of: #5209 Signed-off-by: Lucas Lefèvre (lul) <[email protected]>
robodoo
pushed a commit
that referenced
this pull request
Nov 15, 2024
The evaluation of splillable function is complex and required some bugfix over time. Some of the bugfixes have some side effects difficult to assess and a specific one is the introduction of infinite in the evaluation which are hidden behind a ceiling of iterations. While infinite loops can occur in very specific cases, they are not common and it makes sense to detect them to spot undesired behaviours. This revision adds a log in debug mode to avoid spam to capture the attention of the developper. closes #5209 Task: 4252800 Signed-off-by: Lucas Lefèvre (lul) <[email protected]>
|
robodoo r- |
|
there's a test which is console.debugging (most probably the test testing there's a max number of iterations. |
rrahir
force-pushed
the
17.0-fix-array-recompute-deps-rar
branch
3 times, most recently
from
5454a99 to
42c1983
Compare
The evaluation of splillable function is complex and required some bugfix over time. Some of the bugfixes have some side effects difficult to assess and a specific one is the introduction of infinite in the evaluation which are hidden behind a ceiling of iterations. While infinite loops can occur in very specific cases, they are not common and it makes sense to detect them to spot undesired behaviours. This revision adds a log in warn mode to capture the attention of the developper. Task: 4252800
| setCellContent(model, "A1", "=MFILL(2,1,D1+1)"); | ||
| setCellContent(model, "C1", "=MFILL(2,1,B1+1)"); | ||
| expect(getEvaluatedCell(model, "A1").value).toBe(31); | ||
| expect(getEvaluatedCell(model, "B1").value).toBe(31); | ||
| expect(getEvaluatedCell(model, "C1").value).toBe(30); | ||
| expect(getEvaluatedCell(model, "D1").value).toBe(30); | ||
| expect(spy).toHaveBeenCalledWith("Maximum iteration rea ched while evaluating cells"); |
There was a problem hiding this comment.
fixed
Suggested change
| expect(spy).toHaveBeenCalledWith("Maximum iteration rea ched while evaluating cells"); | |
| expect(spy).toHaveBeenCalledWith("Maximum iteration reached while evaluating cells"); |
LucasLefevre
force-pushed
the
17.0-fix-array-recompute-deps-rar
branch
from
42c1983 to
a7ab544
Compare
|
robodoo r+ |
robodoo
pushed a commit
that referenced
this pull request
Nov 15, 2024
How to reproduce: -> see the attached test What is happening? During the evaluation of SUMIFS, it will fetch its ranges. First it will first fetch an empty range (E4:E7) since MUNIT hasn't been evaluated yet. Afterwards, it will fetch H4, which in turn will request C4 and therefore force the evaluation of the spreaded function MUNIT. AT this point, there is a disparity of values because the values of the range (E4:E7) have changed. To address those situations, the evaluator invalidates the dependencies of all the spreaded zone. Which means marking A1 to be recomputed. Unfortunately, A1 is still being evaluated and since #4589 (specifically commit 8d22e86) we remove the cell from the queue after we evaluted it. This revision changes the approach taken in 8d22e86 to allow legitimate invalidations take place while avoiding useless reevaluations. Task: 4252800 Part-of: #5209 Signed-off-by: Lucas Lefèvre (lul) <[email protected]>
robodoo
pushed a commit
that referenced
this pull request
Nov 15, 2024
The evaluation of splillable function is complex and required some bugfix over time. Some of the bugfixes have some side effects difficult to assess and a specific one is the introduction of infinite in the evaluation which are hidden behind a ceiling of iterations. While infinite loops can occur in very specific cases, they are not common and it makes sense to detect them to spot undesired behaviours. This revision adds a log in warn mode to capture the attention of the developper. closes #5209 Task: 4252800 Signed-off-by: Lucas Lefèvre (lul) <[email protected]>
This was referenced Nov 15, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
How to reproduce:
-> see the attached test
What is happening?
During the evaluation of SUMIFS, it will fetch its ranges. First it will first fetch an empty range (E4:E7) since MUNIT hasn't been evaluated yet.
Afterwards, it will fetch H4, which in turn will request C4 and therefore force the evaluation of the spreaded function MUNIT.
AT this point, there is a disparity of values because the values of the range (E4:E7) have changed.
To address those situations, the evaluator invalidates the dependencies of all the spreaded zone. Which means marking A1 to be recomputed.
Unfortunately, A1 is still being evaluated and since #4589 (specifically commit 8d22e86) we remove the cell from the queue after we evaluted it.
This revision changes the approach taken in 8d22e86 to allow legitimate invalidations take place while avoiding useless reevaluations.
Task: 4252800
Description:
description of this task, what is implemented and why it is implemented that way.
Task: TASK_ID
review checklist