Update citeproc-js #178

Closed
gracile-fr opened this Issue Jul 12, 2015 · 13 comments

Projects

None yet

4 participants

@SteveRidout
Member

Thanks for letting me know.

I'm off to visit friends and family for a couple of weeks and when I return I'll try upgrading it, hopefully the transition from 1.0 to 1.1 hasn't made any breaking changes.

When I get back I'm going to put a message up on the editor site, the github repo, and the uservoice help page asking for a web developer to volunteer to maintain the editor. I don't work for Mendeley, I don't work in academia, and I don't write papers so have no use for CSL myself. I have nothing invested in this anymore apart from the pleasure of seeing people still using it, it's very nice but not enough to justify spending much more time on it.

@gracile-fr

Oh yes, no hurry. Actually, I was wondering if you were still maintaining it, so thanks for your fast reply! (cc @rmzelle @fbennett @adam3smith @cpina )

@adam3smith
Member

Thanks -- we have some funds from CSL that we could offer as a bounty, especially if someone were interested in doing a little more than just maintenance. That'd have to be approved first, but in the past there was always strong support for anything involving making CSL more user friendly.

@SteveRidout
Member

Just tried updating but ran into a new citeproc error so have left it alone for now. By manual trial and error tracked the introduction of the error down to changes between citeproc 1.1.5 and 1.1.6. Here are the details in case @fbennett wants to investigate:

Uncaught TypeError: state.sys.variableWrapper is not a function

Here's the top of the stack trace (using citeproc 1.1.6 - https://bitbucket.org/fbennett/citeproc-js/raw/bc738a142a948006590f303777dcc9e4d8b51304/citeproc.js):

CSL.Output.Formats.html.@showid/true @ citeproc.js?bust=$GIT_COMMIT:13055
CSL.Output.Queue.string @ citeproc.js?bust=$GIT_COMMIT:2617
CSL.Output.Queue.string @ citeproc.js?bust=$GIT_COMMIT:2554
CSL.Output.Queue.string @ citeproc.js?bust=$GIT_COMMIT:2554
CSL.Output.Queue.string @ citeproc.js?bust=$GIT_COMMIT:2554
CSL.getCitationCluster @ citeproc.js?bust=$GIT_COMMIT:4221
CSL.Engine.process_CitationCluster @ citeproc.js?bust=$GIT_COMMIT:3951
CSL.Engine.processCitationCluster @ citeproc.js?bust=$GIT_COMMIT:3933
CSL.Engine.appendCitationCluster @ citeproc.js?bust=$GIT_COMMIT:3491

@SteveRidout
Member

I've added a "Maintainer needed" note to the about page and the uservoice page - if anyone contributes I'll review pull requests, at least for a while.

@adam3smith
Member

@fbennett - taking another stab at this. See Steve's stack trace above. When I run current citeproc-js to regenerate the styles for preview in the visual editor, I similarly get
CITEPROC ERROR: Citeproc exception: TypeError: Cannot set property 'offset' of undefined
and
CITEPROC ERROR: Citeproc exception: TypeError: Object [object Object] has no method 'variableWrapper'

Older versions work fine. Any ideas/solutions where we could/should be looking? @fcheslack -- I know you've been working with citeproc-js in node for citeproc-node : do you have any insights (I tried simply copying over csl_json and the citeproc version from there, but that was probably too simplistic).

@fbennett
Member

The visual editor uses an option setting "csl_reverse_lookup_support" that has to be set in the source before the processor is instantiated, rather than at run-time. It looks like it's breaking when the option is enabled. The test suite is only set up to test options that can be enabled after instantiation. This apparently wasn't being checked, and a bug crept in. I'll look into it.

@fbennett
Member

(It looks like it's breaking reliably on authors, at the least.)

@fbennett
Member

Try now with the latest. It may be fixed.

@adam3smith
Member

seems to run fine now -- I'll close this once successfully updated.

@adam3smith
Member

yup, all worked swimmingly, new site deploying now.

@adam3smith adam3smith closed this Nov 23, 2015
@adam3smith
Member

and, of course, thanks!

@fbennett
Member

Yo, that's great. The code for reverse navigation tagging is pretty hard to follow, and I was worried that it might be necessary to wade into it to get it going again. All I did in the end was to ignore tree nodes that had crept in with no reverse-lookup info registered on them, and apparently that was all it needed. Whew.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment