On Tue, Aug 30, 2011 at 5:35 PM, Jukka Zitting <[email protected]> wrote: > Hi, > > On Tue, Aug 30, 2011 at 9:07 PM, <[email protected]> wrote: >> + assertContains("zażółć gęślą jaźń", content); >> + assertContains("ZAŻÓŠĆ GĘŚLĄ JAŹŃ", content); > > I think it would be best if we used \uNNNN escapes for non-ASCII > characters in .java files. Our Maven build already standardizes to > UTF-8, but there's no guarantee that someone who later edits the file > uses the correct encoding settings.
Hmm, thinking more about this: are we sure we can't make full use of UTF8 in our source files? It'd make the source much more readable for those of us working on non-ascii tests... Also, if someone does edit the file and (say) writes it incorrectly in the wrong encoding, their tests will suddenly fail locally and they'd know something is up? I think expecting Tika devs to grok source encoding issues is reasonable? Or are people actively using editors that can't handle UTF8 or something...? Failing that, if we really must only use ASCII for Tika's sources... shouldn't we fix Maven to enforce this, so that I see an error when compiling if I use non-ASCII? Mike McCandless http://blog.mikemccandless.com
