Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    danyill
    @danyill
    Well with the BRAND SPANKING NEW Antora includes if they work up at component level what I do might be a convenient way to set attributes across an entire website
    I'm used to publishing all my documentation as discrete zip files (I know...) and I have a subtree with my "regular include Asciidoc files" to help with this
    :fireworks: :fireworks:
    Dan Allen
    @mojavelinux

    I do might be a convenient way to set attributes across an entire website

    exactly!

    though you can also do that using the playbook ;)
    but where it becomes very useful is per component
    danyill
    @danyill
    Ah. true.
    Dan Allen
    @mojavelinux
    that per component (at least per component version) will also be happening in antora.yml in 1.2 or 1.3
    i've needed that multiple times. but as you pointed out, the include can get you there too
    danyill
    @danyill
    I haven't caught up with some of the new features yet. I have to decide whether to stop doing my zip files once I move to Antora. One of the things I don't know how to do is to link to large PDFs (which I do a lot of) and have Antora intelligently re-open the link in the same window it's already open in a sane way (probably possible with Javascript). On a local filesystem this just doesn't matter (because it's fast) but online opening a 50MB file too many times is a problem
    This is part of why I'm so whiny about the resource id thing because when I migrate I have to deal with forward/backwards compatibility and I want to be able to script going between the "old way" and the "Antora structure"
    But I'm sure I can get there just by moving slowly and I'm starting to see the pluses on the resource id approach that's been taken. Things do have a home.
    Joseph Cayouette
    @jcayouette
    @danyill Stop using pdf XD hehe problem solved. "grins"
    @danyill I also have pdf requirements, it can be challenging.
    brb lunch
    danyill
    @danyill
    If only it was possible. I work in the electricity industry and we do everything like it's no later than 2006
    And the company mandates Sharepoint and Word. Antora/Asciidoc is my blackmarket attempt to illustrate what a better life could look like (I have yet to find a problem Sharepoint actually solves)
    Dan Allen
    @mojavelinux
    i want to be clear I'm totally fine with producing PDFs. it's not PDF itself I have a problem with, per se
    it's the idea that you can make a PDF creator that's any good without either a) paying or b) using the brower
    and clearly, b) is the way to go for an OSS project ;)
    I tried to create a PDF converter. it can do some things...but then you hit a very very thick wall
    so now I understand the problem space better
    and I'm convinced that if we use the browser engine, we can produce very nice PDFs with relatively low effort
    we don't have invent font rendering, CSS, or layouts
    danyill
    @danyill
    Tots agree with that. We know that PDF is a challenging format because it doesn't handle styling in a sane fashion.
    For professional publishing where that's done by something else it doesn't matter
    Dan Allen
    @mojavelinux
    we just have to solve some corner cases with things like page numbering, which is actually very easy to solve with a second pass (you just layer objects on top of the existing PDF)
    danyill
    @danyill
    But we best not re-implement, say Latex
    Dan Allen
    @mojavelinux
    exactly.
    danyill
    @danyill
    Mediawiki had this issue with rendering and eventually came to your exact same conclusion
    Dan Allen
    @mojavelinux
    I tried very hard not to come to this conclusion
    danyill
    @danyill
    the offline content generation and the collection extension fwiw
    Dan Allen
    @mojavelinux
    and it used to be a pain because all we had was thinging like wkhtmlpdf. but now browsers are offering this as a first-class tool, so this changes everything
    gotta head out in a minute
    danyill
    @danyill
    Thanks for the convo and for 1.1
    Joseph Cayouette
    @jcayouette
    @mojavelinux Yes the brower is def the right target for producing PDF. Totally agree with you on pdf thoughts. Nice points @danyill
    Ian Tindale
    @u0421793
    About fifteen years ago I was using the Apache FOP project (with XSL-FO) to generate pdf. I wonder if a] that’s still going, and b) if it can be fed with something other than formatting objects these days?
    Aha — https://xmlgraphics.apache.org/fop/faq.html — not touched for most of this decade; and no, just XSL-FO. Never mind.
    Joseph Cayouette
    @jcayouette
    @mojavelinux Thanks for 1.1, just updated.
    @u0421793 we use FOP here at SUSE with our DAPS toolchain. https://opensuse.github.io/daps/
    @u0421793 It does an ok just, but the same pdf headaches apply. I am using asciidoctor to convert adoc sources -> DB5 for our product outputs.
    @u0421793 Pretty hacky atm, but it works.
    Ian Tindale
    @u0421793
    :-)
    Joseph Cayouette
    @jcayouette
    @u0421793 The package we use is xmlgraphics-fop
    Joseph Cayouette
    @jcayouette
    @danyill hm interesting use of asciidoc :D https://github.com/danyill/asciidoctorj-pdf-invoices
    Ian Tindale
    @u0421793
    A big improvement for me would be removal of the requirement of a “version”. I have only one version, the correct one. I’ve identified the version mechanism as probably the thing that causes the links in the menu to give 404 messages. You click on something in the same module, and it works, then another module, it works, then click from there to a different module and it won’t work, and from then onwards, nothing works. I think the problem is that it loses the ‘version’ part of the link and won’t ever gain it back once it is lost. Doing without the problematic and unnecessary “version” mechanism would be a good thing, as it would stop these sort of breakages if people click where I’ve had to tell them not to.
    Ian Tindale
    @u0421793
    Also, when I run antora and it doesn’t work, it says “Add the --stacktrace option to see the cause.”. I do that, but invariably it doesn’t show me the cause, but instead, a ream of incomprehensible technical stuff.
    Dan Allen
    @mojavelinux

    , a ream of incomprehensible technical stuff.

    that's the cause ;)

    yes, we have an open issue to auto-set the version based on the current reference name.
    i'll see if I can find it