Error reporting strings are Latin1, not UTF-8.

[eefriedman]

Because the best way to convert from a 16-bit encoding to an 8-bit encoding is obviously truncation.

[michaelwu]

Have you tried using report->ucmessage?

[eefriedman]

Just tried it... it contains the exact same Latin1 string inflated to u16.

[michaelwu]

Hmm I expected it to be the other way around. For example, in http://mxr.mozilla.org/mozilla-central/source/js/src/jscntxt.cpp#650 , ucmessage is the original string before truncation to latin1.

[eefriedman]

Added commit to use ucmessage. You can test for yourself if you want.

[eefriedman]

Oops; not sure how I forgot it was in the standard library. Updated.

[eefriedman]

@michaelwu Ping.

[michaelwu]

@bors-servo r+

[eefriedman]

Nothing is happening here; do I need to file a bug against saltfs or something?

[jdm]

@bors-servo: r=michaelwu I had to re-synchronize homu's queue for rust-mozjs.

[bors-servo]

:pushpin: Commit 9c4a2ba has been approved by michaelwu

[bors-servo]

:hourglass: Testing commit 9c4a2ba with merge bba1b42...

[bors-servo]

:broken_heart: Test failed - travis

[eefriedman]

The build timed out during the Spidermonkey build. The configuration probably needs to be changed somehow.

[jdm]

@bors-servo: retry

[bors-servo]

:hourglass: Testing commit 9c4a2ba with merge 0aae51a...

[bors-servo]

:broken_heart: Test failed - travis

[jdm]

Bah, we're being bitten by libc here. #211 needs to merge first.

[jdm]

@bors-servo: retry

[bors-servo]

:hourglass: Testing commit 9c4a2ba with merge e2458c4...

[bors-servo]

:sunny: Test successful - travis