1239516 - ASSERTION: nsTDependentString must wrap only null-terminated strings. You are probably looking for nsTDependentSubstring.: 'mData[mLength] == 0'

Status: RESOLVED FIXED

(Core :: XPCOM, defect)

Description

[Mats Palmgren (inactive)]

I get this assertion when running reftests in a local beta debug build on Linux64.
Looking at the blame I guess it's a regression from bug 1227023.
The call comes from here:
http://hg.mozilla.org/mozilla-central/annotate/ad1f85f172b7/xpcom/base/nsSystemInfo.cpp#l706

Comment 1

[Mats Palmgren (inactive)]

Attachment: fix?

This makes the assertion go away, fwiw.

Comment 2

[Karl Tomlinson (:karlt)]

Comment on attachment 8707680 [details] [diff] [review]
fix?

Thank you.  That is what I was thinking the length argument was for.

Comment 3

[Mike Hommey [:glandium]]

And that's actually what I was meaning to do, but failed to.

Comment 4

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/37d094c1ed1e

Comment 5

[Karl Tomlinson (:karlt)]

Looking through the uses of IsTerminated(), I think this will only cause real problems if secondaryLibrary is used together with nsPromiseFlatCString.

secondaryLibrary is only read from javascript and concatenated among other strings.
I don't know what code does js reads, but I think use of nsPromiseFlatCString would be unlikely.

I ran valgrind --tool=memcheck on amd64 firefox-44.0b8, checked for updates, and no uninitialized reads were reported.

Comment 6

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/37d094c1ed1e

Comment 7

[Karl Tomlinson (:karlt)]

Comment on attachment 8707680 [details] [diff] [review]
fix?

Approval Request Comment
[Feature/regressing bug #]: bug 1227023
[User impact if declined]:
I don't expect any problems from this bug on its own, but the patch is required for the patch in bug 1245745 to apply cleanly to 45.
[Describe test coverage new/current, TreeHerder]:
There are existing tests running this code, but they don't expect any particular string AFAIK.
[Risks and why]: low
[String/UUID change made/needed]: none

Comment 8

[Sylvestre Ledru [:Sylvestre]]

Comment on attachment 8707680 [details] [diff] [review]
fix?

Necessary for bug 1245745

Comment 9

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/releases/mozilla-beta/rev/d9bbf3f7bbf8

Comment 10

[Martin Mokrejs]

For your reference I also hit this with vanilla firefox-44.0:

ASSERTION: nsTDependentString must wrap only null-terminated strings. You are probably looking for nsTDependentSubstring.: 'mData[mLength] == 0', file ../../dist/include/nsTString.h, line 455