On Thu, Aug 16 2012, Eric Schulte wrote: > Eric Abrahamsen <e...@ericabrahamsen.net> writes: > >> On Thu, Aug 16 2012, Eric Schulte wrote: >> >>> Eric Abrahamsen <e...@ericabrahamsen.net> writes: >>> >>>> On Thu, Aug 16 2012, Eric Schulte wrote: >>>> >>>>>>> Every time I edited a block and clicked "save", it just deleted the >>>>>>> whole block. I got these errors in ~/.elnodelogs/elnode-error: >>>>>>> >>>>>> >>>>>> I'm not sure what could be causing this problem. Did the test suite run >>>>>> successfully for you? >>>>> >>>>> I've now added a POST test to the test suite. So if your problem >>>>> persists you should now see a failing test, and conversely if you are >>>>> now passing the test suite this problem should be eliminated. >>>> >>>> So yes, I did run the tests the first time, and you're right that, >>>> without the "post" test, they went just fine. The documents were also >>>> altered on-disk (ie, the chunks really were deleted). >>> >>> Interesting. >> >> [...] >> >>> Even thought this test is failing, it does show that your elnode server >>> is returning the HTML in response to your POST requests. It looks like >>> it only fails because your Emacs exports *foo* as <em>foo</em> instead >>> of as <i>foo</i>. >>> >>>> >>>> Then I restarted emacs -Q and used your batch.el file. I'm still getting >>>> the same problem, unfortunately: the editable blocks disappear when I >>>> hit "save". I realized I don't actually know whether this is supposed to >>>> edit the simple.org or simple.html files: >>> >> >> [...] >> >>> I've just pushed up some changes to the git repository which add new >>> require statements. Hopefully the errors above were caused by some >>> required functions not being loaded at run time. If the newest from git >>> doesn't work fix these problems, please try running >>> >>> emacs -Q -l batch.el >>> >>> with this updated version of batch.el, and let me know what is printed >>> in the *Messages* buffer in the line which starts as "params:". >> >> The same test failed in the same way, > > Fair enough, this must be a legitimate difference in our HTML export, > I'll make the test appropriately more permissive. > >> so I loaded up the new batch.el file, and here's the "params" line: >> >> params:(("path" . "/simple.org") ("end" . "577") ("beg" . "156") ("org" >> . "")) >> >> Hope that's enlightening! >> > > Yes, very enlightening. It means that the server-side and Emacs Lisp > side are working as expected. However, for some reason, when you hit > [SAVE] the information submitted by your browser includes an empty text > field (this is the "org" pair above). > > This must be a JavaScript issue related to some difference between > browsers. Can I ask, what browser you are using? I've done all of my > testing with browsers in the Firefox family (Firefox and conkeror).
I'm using the version of Chrome provided by Arch linux, it tells me Version 21.0.1180.79 (151411). I just tried the test server in Firefox, and it worked perfectly. Much as I like Chrome, I've occasionally seen it make websites fail in ways you wouldn't think a browser should be capable of. -- GNU Emacs 24.1.50.1 (i686-pc-linux-gnu, GTK+ Version 2.24.11) of 2012-08-14 on pellet 7.8.11