dmsmith at crosswire.org
Sat Dec 30 09:07:26 MST 2017
JSword does it’s own thing. The filter architecture is entirely different from SWORD’s. It converts Plain, GBF, ThML, TEI, … into OSIS. Then the toggles are passed to XSLT as parameters. The XSLT converts OSIS to HTML and it does the toggling.
STEP’s XSLT was based on BibleDesktop’s. As a fork, it handles things a bit differently than BibleDesktop. For example, it does morph segmentation, but BD does not.
SWORD has a filter for each toggle that modifies the input (e.g. OSIS, ThML) that the render filters get.
> On Dec 30, 2017, at 10:22 AM, David Haslam <dfhdfh at protonmail.com> wrote:
> Thanks for the detailed report, Karl.
> It's just as I feared.
> It also indicates that STEP Bible must do its own thing for such modules.
> It displays alternate segments in a different colour.
> Best regards, David
> Sent from ProtonMail Mobile
> On Sat, Dec 30, 2017 at 15:02, Karl Kleinpaste <karl at kleinpaste.org <mailto:karl at kleinpaste.org>> wrote:
>> On 12/30/2017 08:31 AM, DM Smith wrote:
>>> The module matches what the filter expects.
>>> The filter does something.
>>> Third possibility: front end doesn’t handle the attribute.
>> I just spent an hour stepping through Xiphos' handling of WLC Gen.1.1 with and without Morpheme Segmentation set.
>> First, Xiphos correctly turns the option on and off. That's kind of a given, but I had to check on it anyway. There's a single area where all options are turned on/off in a uniform way, driven by the GTK menu files, and how Xiphos discovers the current setting during each chapter redraw. Mostly, I had to check that I had spelled the option right in all the relevant places, to be sure the right effect was being caused.
>> Second, the actual text output that results from Morpheme Segmentation being on vs. off is identical, as discovered by trapping the display routine that accepts the engine's result when requesting the verse. This image is the result of copy/pasting gdb's textual output from a terminal into gedit and then capturing an image of that.
>> The filter has no effect on the text. If someone wants to look at how osismorphsegmentation.cpp does this, feel free, but objectively what comes out from requesting the verse doesn't care whether the option is on.
>> More simply, "diatheke -b WLC -f xhtml -k gen.1.1" with and without "-o M" proves they're the same, as shown with both diff and md5sum.
> sword-devel mailing list: sword-devel at crosswire.org
> Instructions to unsubscribe/change your settings at above page
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the sword-devel