1314378 - Port TestStartupCache to gtest

Status: RESOLVED FIXED

(Core :: XPCOM, defect)

Description

[Benjamin Smedberg]

TestStartupCache is currently a standalone test, which will not be possible any more when the XPCOM glue is gone. It should be ported to a gtest.

As part of this, I'm also going to remove some startupcache telemetry which I'm sure nobody is using and is of dubious value at all.

Comment 1

[Benjamin Smedberg]

Attachment: Bug 1314378 part A - remove startupcache telemetry of dubious value.

Review commit: https://reviewboard.mozilla.org/r/90418/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/90418/

Comment 2

[Benjamin Smedberg]

Attachment: Bug 1314378 part B - port TestStartupCache to gtest,

Review commit: https://reviewboard.mozilla.org/r/90420/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/90420/

Comment 3

[Nathan Froyd [:froydnj]]

Comment on attachment 8807159 [details]
Bug 1314378 part A - remove startupcache telemetry of dubious value.

https://reviewboard.mozilla.org/r/90418/#review90186

Comment 4

[Nathan Froyd [:froydnj]]

Comment on attachment 8807160 [details]
Bug 1314378 part B - port TestStartupCache to gtest,

https://reviewboard.mozilla.org/r/90420/#review90212

::: startupcache/test/TestStartupCache.cpp:66
(Diff revision 1)
> +{
> +  nsresult rv = NS_GetSpecialDirectory(NS_OS_TEMP_DIR, getter_AddRefs(mSCFile));
> +  mSCFile->AppendNative(NS_LITERAL_CSTRING("test-startupcache.tmp"));
> +  nsAutoCString path;
> +  mSCFile->GetNativePath(path);
> +  char* env = PR_smprintf("MOZ_STARTUP_CACHE=%s", path.get()); // leak

Does this need some sort of annotation for ASan/LSan?

This is kind of gross, but given that we already have the environment variable support...

Comment 5

[Benjamin Smedberg]

Do we run ASAN/LSAN on gtests? I guess MOZ_LSAN_INTENTIONALLY_LEAK_OBJECT wouldn't hurt.

Comment 6

[Andrew McCreight [:mccr8]]

(In reply to Benjamin Smedberg [:bsmedberg] from comment #5)
> Do we run ASAN/LSAN on gtests? I guess MOZ_LSAN_INTENTIONALLY_LEAK_OBJECT
> wouldn't hurt.

We run ASan, but I think LSan is disabled there. I'd be surprised if we didn't leak a bunch of things in these tests, and I can't recall anybody doing the work to fix it up.

Comment 7

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/419e1a311337
part A - remove startupcache telemetry of dubious value. r=froydnj
https://hg.mozilla.org/integration/mozilla-inbound/rev/4585a7b6f3ce
part B - port TestStartupCache to gtest, r=froydnj

Comment 8

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/52cfc93236f0
Build bustage fixup from bug 1314378 and bug 1314955 - remove things from the whitelist which are no longer built, r=lumpy on a CLOSED TREE

Comment 9

[Iris Hsiao [:ihsiao]]

Sorry had to back out these also because there are bustages after backed out bug 1314955

Comment 10

[Pulsebot]

Backout by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/4ab3e15efcfe
Backed out changeset 52cfc93236f0 to fix the bustage after backed out bug 1314955
https://hg.mozilla.org/integration/mozilla-inbound/rev/03614ff20ee0
Backed out changeset 4585a7b6f3ce 
https://hg.mozilla.org/integration/mozilla-inbound/rev/87b4c375fe75
Backed out changeset 419e1a311337

Comment 11

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/6578aec20296
part A - remove startupcache telemetry of dubious value. r=froydnj
https://hg.mozilla.org/integration/mozilla-inbound/rev/ee8798338529
part B - port TestStartupCache to gtest, r=froydnj

Comment 12

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://hg.mozilla.org/mozilla-central/rev/6578aec20296
https://hg.mozilla.org/mozilla-central/rev/ee8798338529