720759 - Windows warning patrol

Status: RESOLVED FIXED

(Core :: JavaScript Engine, defect)

Description

[Jeff Walden [:Waldo]]

!summon themaid

Comment 1

[Jeff Walden [:Waldo]]

Attachment: The guard object classes need to be exported, if they're going to be used in exported APIs

From the log for js/src/jsutil.h:

user:        Jeff Walden <[e-mail]>
date:        Wed Oct 05 15:21:53 2011 -0700
summary:     Make the guard-object macros, used to ensure a class can't accidentally be constructed only temporarily, JS_FRIEND_API-compatible, since they're being used in friend declarations now.  rs=cdleary/luke, no bug

Gotta do the same thing for mfbt's classes...

Comment 2

[Jeff Walden [:Waldo]]

Attachment: Change a mess of function signatures to take bool*, not JSBool*, all to fix one Windows warning comparing bool to JSBool

First it was one method, then two, then then then.  I drew the line at changing js_TestXMLEquality.  And I think changing JSEqualityOp's not possible now either.  :-\

Comment 3

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

Comment on attachment 591170 [details] [diff] [review]
The guard object classes need to be exported, if they're going to be used in exported APIs

Review of attachment 591170 [details] [diff] [review]:
-----------------------------------------------------------------

lgtm

Comment 4

[Jeff Walden [:Waldo]]

Attachment: Delete EqualStrings(cx, linearStr1, linearStr2, &equal) so everyone has to use the made-for-linear-strings overload

Yes, Virginia, MOZ_DELETE works outside of classes too.

Comment 5

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

Comment on attachment 591173 [details] [diff] [review]
Change a mess of function signatures to take bool*, not JSBool*, all to fix one Windows warning comparing bool to JSBool

Review of attachment 591173 [details] [diff] [review]:
-----------------------------------------------------------------

I like.

Comment 6

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

Comment on attachment 591176 [details] [diff] [review]
Delete EqualStrings(cx, linearStr1, linearStr2, &equal) so everyone has to use the made-for-linear-strings overload

Review of attachment 591176 [details] [diff] [review]:
-----------------------------------------------------------------

Sure, if this works.

Comment 7

[Jeff Walden [:Waldo]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/bd75f26eee25
https://hg.mozilla.org/integration/mozilla-inbound/rev/6e849614a79f
https://hg.mozilla.org/integration/mozilla-inbound/rev/01a5ce645a8b

Comment 8

[Ed Morley [:emorley]]

Backed out for win debug failures: (coalescing means this is the first tbpl push that shopws the failure)
https://tbpl.mozilla.org/?tree=Mozilla-Inbound&rev=b7f926cfa8c8
https://tbpl.mozilla.org/php/getParsedLog.php?id=8796111&tree=Mozilla-Inbound

https://hg.mozilla.org/integration/mozilla-inbound/rev/cac3f64d36d0

Comment 9

[Jeff Walden [:Waldo]]

Sigh; I wasn't thinking that JS_FRIEND_API expands to an export or an import depending on the user.  I needed MFBT_API instead.  Relanded with that, after doing a Windows debug build with that to verify the fix.

https://hg.mozilla.org/integration/mozilla-inbound/rev/ec9a6b8c81d8
https://hg.mozilla.org/integration/mozilla-inbound/rev/5d572733c31e
https://hg.mozilla.org/integration/mozilla-inbound/rev/fd742da19c52

Comment 10

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

https://hg.mozilla.org/mozilla-central/rev/bd75f26eee25
https://hg.mozilla.org/mozilla-central/rev/6e849614a79f
https://hg.mozilla.org/mozilla-central/rev/01a5ce645a8b
https://hg.mozilla.org/mozilla-central/rev/cac3f64d36d0
https://hg.mozilla.org/mozilla-central/rev/ec9a6b8c81d8
https://hg.mozilla.org/mozilla-central/rev/5d572733c31e
https://hg.mozilla.org/mozilla-central/rev/fd742da19c52