Logs for #nikola for 2020-05-02

15:48:23 <ncdulo> I had some difficulty last night diagnosing why certain page <title> blocks would have <span> ampersands in them. Tried many different things such as the striphtml() filter and various mako filters. What I determined, is the Typogrify HTML filter was causing that. I was not able to find a solution with existing filters, so I went ahead and modified the filter into a new `typogrify_sans_amp` filter.
15:49:25 <ncdulo> I can submit the code somewhere, if anyone is interested. Read through the plugin docs but did not see how to add it as a plugin. So I just directly import the new filter into my config.
15:49:59 <ncdulo> Mostly just asking if this is something I should put into a GH issue, or if I should post the modified filter anywhere for other people that may encounter the same issue
15:52:54 <ChrisWarrick> ncdulo: perhaps a github issue or pR to fix the thing
15:53:00 <ChrisWarrick> ncdulo: <span> in <title> is a bug
15:54:06 <ncdulo> Yeah, it was giving me a lot of trouble trying to figure it out. Only noticed it because of the HTML Tidy filter going weird on it
15:54:56 <ncdulo> I'll write up an issue and post my code. Not sure if I placed my filter into the right place in the sources
15:55:21 <ncdulo> Though I could definitely just clone the Nikola repo, make the change into the main filters file and submit that as a PR
16:29:34 <KwBot> [nikola] ncdulo opened issue #3404: Typogrify filter places <span> into certain page <title> blocks https://github.com/getnikola/nikola/issues/3404
16:32:58 <ncdulo> I've got a PR for the issue on a fork, working through some of the extra things like Changes.txt before I go ahead and submit that
16:56:45 <KwBot> [nikola] ncdulo opened pull request #3405 (ncdulo:ncdulo-typogrify_sans_amp): Ncdulo typogrify sans amp https://github.com/getnikola/nikola/pull/3405
19:32:41 <KwBot> [nikola] cosmoscalibur opened issue #3406: AttributeError: 'str' object has no attribute 'task' https://github.com/getnikola/nikola/issues/3406
21:11:20 <erdgeist> Hello.
21:16:35 <felixfontein> hi
21:19:45 <erdgeist> I am having trouble with the output-filename detection: It's a multilanguage-site with two languages configured in TRANSLATIONS. The pattern is default: TRANSLATIONS_PATTERN = '{path}.{lang}.{ext}', still my input file updates/2007/foo.de.md will be rendered to render_pages:output/de/updates/2007/foode.html ... not foo.html as I would expect
21:20:37 <erdgeist> Anyone knows where the extra copy of file's language id is added to the output filename and how to prevent it?
21:21:19 <ChrisWarrick> erdgeist: What is your slug for this post?  Which is the default language?
21:22:04 <erdgeist> Default language is 'de'. No extra slug specified
21:22:25 <felixfontein> I guess you should name the file `foo.md` and not `foo.de.md` then
21:22:27 <ChrisWarrick> the default language version is expected to be named foo.md
21:23:02 <erdgeist> I see. Thanks.
22:15:08 <erdgeist> Hah! The other problem I encountered could be solved by just googling our conversation from five years ago ;)   https://irclogs.getnikola.com/logs/2015-02-21/
23:13:59 <erdgeist> ChrisWarrick: I keep running into this "HASH: Out of overflow pages.  Increase page size" bug. I wonder if I can make doit use another database backend, maybe sqlite? According to https://pydoit.org/cmd_run.html this can be done by the --backend option ... can I configure this with nikola?
23:14:46 <ChrisWarrick> erdgeist: you can run nikola with the specific backend
23:14:59 <ChrisWarrick> erdgeist: nikola build --backend=sqlite3
23:15:16 <erdgeist> will give it a shot, thanks
23:45:41 <erdgeist> works much better. Great