Lack of features in Gitksan corpus

[kchall]

In one Gitksan corpus (Barbara_stories_April_9_orthnomorph_nocaps.corpus in the test files Dropbox folder), the feature files seem to be correct and allow PCT to organize the segments into IPA charts, but when a phonological search is conducted, there are no features listed in the "basis for search" section, and features aren't even a choice in the "environment" section. This seems to be a problem with this particular corpus, as the feature file works fine with other corpora, and all other feature files run in to the same problem with this corpus.

[kchall]

Further investigation shows that this is a recurrent problem with trying to associate a feature file with a corpus that has been created from running text. I was able to "fix" the problem in another instance by exporting the created corpus as a tab-delimited .txt file and re-create it from that new .txt file, and then associating the feature system to it. So it seems to be a real issue, and somehow connected to corpora that are created from running text. @mmcauliffe

[bhallen]

@kchall I'd like to fix this bug, but first I need to be able to replicate it. Could you please send me the Gitksan running text file so that I can do this? I don't have access to the general-purpose Dropbox folder anymore.

[mmcauliffe]

Sorry, I probably should have closed this a while ago. I'm almost 100% positive that I fixed on the develop branch when looking at the lack of features for Avery and Michael's project. I'm guessing that the recent update from @kchall was on the io-standardize branch?

[kchall]

Hmm, I was looking at it in v. 1.0.1.

It does seem to be working ok now on the develop branch. Sorry for the confusion!