801471 - Failed to convert nsresult to HRESULT in nsDataObj.cpp on windows build

Status: RESOLVED FIXED

(Core :: Widget: Win32, defect)

Description

[Oleg Romashin (:romaxa)]

Attachment: nsresult HRESULT conversion fix

Build fix

Comment 1

[Oleg Romashin (:romaxa)]

Attachment: One more nsresult vs HRESULT conversion fail

Comment 2

[Trevor Saunders (:tbsaunde)]

Comment on attachment 671282 [details] [diff] [review]
One more nsresult vs HRESULT conversion fail

r=tbsaunde

Comment 4

[Brian R. Bondy [:bbondy]]

For personal reference, do nsresult and HRESULT match up enough for a cast between them to be better than just returning a general failure?

For example I've done this before: return SUCCEEDED(hr) ? NS_OK : NS_ERROR_FAILURE;

Comment 5

[Masatoshi Kimura [:emk]]

(In reply to Brian R. Bondy [:bbondy] from comment #4)
> For personal reference, do nsresult and HRESULT match up enough for a cast
> between them to be better than just returning a general failure?
> 
> For example I've done this before: return SUCCEEDED(hr) ? NS_OK :
> NS_ERROR_FAILURE;

My patch did that. See the duped bug.

Comment 6

[Brian R. Bondy [:bbondy]]

Oops ya that's the patch I was looking at.

Comment 7

[Aryeh Gregor (:ayg) (no longer with Mozilla)]

(In reply to Brian R. Bondy [:bbondy] from comment #4)
> For personal reference, do nsresult and HRESULT match up enough for a cast
> between them to be better than just returning a general failure?

A few specific values agree, like S_OK == NS_OK and E_FAIL == NS_ERROR_FAILURE, but the large majority don't correspond.  In particular, all module-specific errors are likely to be totally uncorrelated.

However, the only difference between different error codes is so that different reasons for failure can be distinguished.  From that perspective, it might be more useful to propagate back the actual error even though it will be bogus as an nsresult, because it will make debugging easier -- if it hits NS_ENSURE_SUCCESS etc. and is printed to the console, or gets back to JS in an exception message.

Comment 8

[Oleg Romashin (:romaxa)]

remote:   https://hg.mozilla.org/integration/mozilla-inbound/rev/25f31c96a280
remote:   https://hg.mozilla.org/integration/mozilla-inbound/rev/8c6f15a77695

Comment 10

[Ed Morley [:emorley]]

https://hg.mozilla.org/mozilla-central/rev/25f31c96a280
https://hg.mozilla.org/mozilla-central/rev/8c6f15a77695