Add proof of concept for T.Output() #1

suntala · 2023-09-26T09:12:14Z

jba · 2023-10-02T18:27:59Z

src/testing/testing.go

+	return OutputWriter{c}
+}
+
+type OutputWriter struct {


jba · 2023-10-02T18:30:14Z

src/testing/testing.go

+			parent.mu.Lock()
+			defer parent.mu.Unlock()
+			if !parent.done {
+				parent.output = append(parent.output, parent.decorate(str, depth+1)...)


This Write method is at a lower level than logDepth. It isn't responsible for adding line info or a final newline, so it shouldn't call decorate. It does need to handle indentation, though. I'm not sure where that happens.

Co-authored-by: Aleks Fazlieva <[email protected]>

jba · 2025-01-31T17:10:49Z

I took another careful look at this. I do think you're on the right track: outputWriter.Write should buffer lines, direct the output to the right place (this test, parent test, etc.) and prepend four spaces. Here is what I think needs to be done:

Move to gerrit so we can have a better review experience, others can comment, and your CL will be linked from the proposal issue. Gerrit also allows having multiple CLs stacked on top of one another, which I think will be valuable. See https://golang.org/doc/contribute.html.
My gut tells me that you should have a function outputWriter.writeLine([]byte). Ultimately all output is a line at a time, so that will be a useful building block. I think you might want it to rewrite the exising code (see below). If you want the arg to be string instead of []byte, that's fine to start. I'm not sure if the arg should include the newline or not; you decide.
Rewrite logDepth and decorate to use outputWriter. In fact, I think the first CL should do that rewrite and make sure all tests pass. It wouldn't even expose t.Output. It would just be an internal change. I think that would increase confidence that you're on the right track, and you could temporarily defer handling various edge cases.
The second CL would export t.Output and add tests for things that can't happen with t.Log, which I think all have to do with partial lines. For example, what happens if you have an incomplete line and the test ends? You'll need to flush the buffer (adding a final newline).

Co-authored-by: Aleks Fazlieva <[email protected]>

This reverts commit 7b0d211.

Co-authored-by: Aleks Fazlieva <[email protected]>

suntala · 2025-02-03T13:50:06Z

@jba thank you so much for the guidance. Doing this in two CLs, with the first one being a rewrite of the log/logDepth is a good idea. Since your message, we did the refactor and got tests working. Next steps are going to be cleaning up the code and moving to gerrit. We're aiming to get them done in the next couple of days.

jba · 2025-02-03T20:28:11Z

Sounds good. I'll hold off reviewing until it's on gerrit.

Co-authored-by: Aleks Fazlieva <[email protected]>

…of variable-sized makeslice CL 653856 enabled stack allocation of variable-sized makeslice results. This CL adds debug hashing of that change, plus a debug flag to control the byte threshold used. The debug hashing machinery means we also now have a way to disable just the CL 653856 optimization by doing -gcflags='all=-d=variablemakehash=n' or similar, though the stderr output will then typically have many lines of debug hash output. Using this CL plus the bisect command, I was able to retroactively find one of the lines of code responsible for golang#73199: $ bisect -compile=variablemake go test -skip TestListWireGuardDrivers [...] bisect: FOUND failing change set --- change set #1 (enabling changes causes failure) ./security_windows.go:1321:38 (variablemake) ./security_windows.go:1321:38 (variablemake) --- Previously, I had tracked down those lines by diffing '-gcflags=-m=1' output and brief code inspection, but seeing the bisect was very nice. This CL also adds a compiler debug flag to control the threshold for stack allocation of variably sized make results. This can help us identify more code that is relying on certain stack allocations. This might be a temporary flag that we delete prior to Go 1.25 (given we would not want people to rely on it), or maybe it might make sense to keep it for some period of time beyond the release of Go 1.25 to help the ecosystem shake out other bugs. Using these two flags together (and picking a threshold of 64 rather than the default of 32), it looks for example like this x/sys/windows code might be relying on stack allocation of a byte slice: $ bisect -compile=variablemake go test -gcflags=-d=variablemakethreshold=64 -skip TestListWireGuardDrivers [...] bisect: FOUND failing change set --- change set #1 (enabling changes causes failure) ./syscall_windows_test.go:1178:16 (variablemake) Updates golang#73199 Fixes golang#73253 Change-Id: I160179a0e3c148c3ea86be5c9b6cea8a52c3e5b7 Reviewed-on: https://go-review.googlesource.com/c/go/+/663795 Auto-Submit: Keith Randall <[email protected]> Reviewed-by: Dmitri Shuralyov <[email protected]> Reviewed-by: Keith Randall <[email protected]> Reviewed-by: Keith Randall <[email protected]> LUCI-TryBot-Result: Go LUCI <[email protected]>

suntala · 2025-05-15T09:23:43Z

golang#73703 and golang#71575 are merged.

jba reviewed Oct 2, 2023

View reviewed changes

suntala force-pushed the suntala/t-output-poc branch from f8a90ab to 61379e3 Compare October 20, 2023 19:13

suntala closed this Jan 15, 2025

suntala reopened this Jan 25, 2025

suntala and others added 4 commits January 25, 2025 10:39

Add proof of concept for T.Output()

cebc979

Co-authored-by: Aleks Fazlieva <[email protected]>

Allow partial input to T.Output()

d1e21b8

Co-authored-by: Aleks Fazlieva <[email protected]>

Refactor outputWriter.Write()

3de3663

Co-authored-by: Aleks Fazlieva <[email protected]>

Fix writing multiple lines

c949982

Co-authored-by: Aleks Fazlieva <[email protected]>

suntala force-pushed the suntala/t-output-poc branch from 3fc09b7 to c949982 Compare January 30, 2025 17:35

Remove question comment

fee380e

suntala and others added 11 commits February 1, 2025 12:14

Fix chatty case

e915748

Co-authored-by: Aleks Fazlieva <[email protected]>

Fix append to parent

fa4275c

Co-authored-by: Aleks Fazlieva <[email protected]>

Refactor log so it uses T.Output

0afdf55

Co-authored-by: Aleks Fazlieva <[email protected]>

Refactor output write method to split string instead of indexing

30db305

Fix T.Output indentation

a812ef4

Fix T.log when called without final new line

c1be55f

Unexport outputWriter provider and fix T.Log's panic message

70a15a2

Rename function for adding call site information

202b2e1

Add decoration function to output writer

7b0d211

Revert "Add decoration function to output writer"

bc4ee5d

This reverts commit 7b0d211.

Refactor decoration functionality in outputWriter.Write

be41190

Co-authored-by: Aleks Fazlieva <[email protected]>

suntala and others added 6 commits February 4, 2025 06:04

Move tests to demonstrate output writer behavior

1e83283

Refactor handling of newlines in output writer

67c43ba

Use string builder when writing output

bea12e3

Append newline if there is input to be buffered

25ac5a7

Refactor decoration functionality when writing output

ab6a9db

Fix panic message

a42a03f

suntala and others added 3 commits February 4, 2025 20:47

Remove some todo comments

38e7deb

Co-authored-by: Aleks Fazlieva <[email protected]>

Fix return value of outputWriter.Write

0d19c84

Extract appending to output from Write

7648014

suntala force-pushed the suntala/t-output-poc branch from b98c1b4 to 7648014 Compare February 4, 2025 22:03

suntala added 2 commits February 4, 2025 23:45

Update comments

0ad06f7

Remove outputWriter demo tests

5e72226

suntala force-pushed the suntala/t-output-poc branch 4 times, most recently from d865109 to 820826a Compare February 5, 2025 10:21

Update doc comments

00014da

Co-authored-by: Aleks Fazlieva <[email protected]>

suntala force-pushed the suntala/t-output-poc branch from 820826a to 00014da Compare February 5, 2025 10:24

suntala closed this May 15, 2025

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Add proof of concept for T.Output() #1

Add proof of concept for T.Output() #1

Uh oh!

suntala commented Sep 26, 2023

Uh oh!

jba Oct 2, 2023

Uh oh!

jba Oct 2, 2023

Uh oh!

jba commented Jan 31, 2025

Uh oh!

suntala commented Feb 3, 2025 •

edited

Loading

Uh oh!

jba commented Feb 3, 2025

Uh oh!

suntala commented May 15, 2025

Uh oh!

Uh oh!

Add proof of concept for T.Output() #1

Add proof of concept for T.Output() #1

Uh oh!

Conversation

suntala commented Sep 26, 2023

Uh oh!

jba Oct 2, 2023

Choose a reason for hiding this comment

Uh oh!

jba Oct 2, 2023

Choose a reason for hiding this comment

Uh oh!

jba commented Jan 31, 2025

Uh oh!

suntala commented Feb 3, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

jba commented Feb 3, 2025

Uh oh!

suntala commented May 15, 2025

Uh oh!

Uh oh!

suntala commented Feb 3, 2025 •

edited

Loading