πΎ Archived View for gemi.dev βΊ gemini-mailing-list βΊ 000535.gmi captured on 2024-08-31 at 17:27:48. Gemini links have been rewritten to link to archived content
β¬ οΈ Previous capture (2023-12-28)
-=-=-=-=-=-=-
I?m curious about how authors structure their workflows in geminispace. My own is somewhat of a work in progress ? I write my posts in vim, manually add an entry to my archive gemini feed and then run a bash script that updates my capsule index page and my atom feed. It?s a bit clunky, and I?m pondering how to make it less cumbersome. What?s your workflow like? Cheers ew0k Sent from my TI-83
On 12/12/2020 11:52, Bj?rn W?rmedal wrote: > I?m curious about how authors structure their workflows in geminispace. My own is somewhat of a work in progress ? > > I write my posts in vim, manually add an entry to my archive gemini feed and then run a bash script that updates my capsule index page and my atom feed. It?s a bit clunky, and I?m pondering how to make it less cumbersome. > > What?s your workflow like? > > Cheers > ew0k > > > Sent from my TI-83 > My site isn't a gemlog per se, but I write in gemtext using acme and then I have built a small site generator tool I use to generate feeds, indexes and an HTML version: https://tildegit.org/nihilazo/sitegenerator. It's not great and probably would be difficult to deploy for any capsule other than mine, but that's my workflow. Nico
On 12/12/20 3:22 PM, Bj?rn W?rmedal wrote: > I?m curious about how authors structure their workflows in geminispace. The thing I've worked out so far for myself is based on the use of a content management system called Logarion (written in Ocaml by Orbifx). I first run it to create a new document based on a title I give. Once the file is made I edit it locally with nano, basically in Gemini format minus Logarion's header. (The documents are meant to be in Markdown, but Gemini is like a subset of that anyway.) When the document is written and marked for publication in its header, I run a script which calls Logarion to export them to GMI along with an index listing ordered by date (date.gmi) that Logarion itself provides. The script then copies the generated content into a directory on my server, and as such it is published. I've been thinking about devising a minimalist alternative to this workflow because it's a little bit overkill for what I'm doing. My thought is that the least a Gemini site or log requires is:
I have a very slow workflow, basically I do everything by hand: edit the post, add to the index and to the feed. I like it to take some time and effort for some reason. --
I use a basic static site generator I coded. This is still WIP though https://github.com/raspbeguy/geminer I first wrote it to convert from my PicoCMS markdown sources. This way it generates also index by tags, by author, by whatever you want.
Le samedi 12 d?cembre 2020, 12:52:29 CET Bj?rn W?rmedal a ?crit : > I?m curious about how authors structure their workflows in geminispace. My own is somewhat of a work in progress ? > > I write my posts in vim, manually add an entry to my archive gemini feed and then run a bash script that updates my capsule index page and my atom feed. It?s a bit clunky, and I?m pondering how to make it less cumbersome. > > What?s your workflow like? I just use vim through ssh to write directly on the server. The gemlog index is dynamic and just lists the files in the folder. The downside is that the name of the file is used as post name, I may update the generated index to fetch the first line of posts and use that as link name, not sure. I decided against using the file creation date automatically as post date, because I want to be able to write dateless pages, which will be updated in the future, while a gemlog post with a date is supposed to be a one shot. C?me
I probably have the hackiest setup here, based on shell scripts, duct tape, and some JS on the deployment parts. https://git.ebc.li/admicos/blog I create a new `.gmi` file under `pages/<category>` and add some boilerplate header and footer lines (for my shell-based templating thing) Then I edit the `post_list` file to add the new page. If I put the `.gmi` extension on the filename here, the page becomes Gemini-exclusive. Finally, I push my changes into the repo above, which pings a HTTP endpoint I am listening to via https://nodered.org, clones the repo and runs the shell script inside the `.makeshift.yml` file, that runs the site generator scripts and pushes the results into the proper place. I would make the Node-RED script thing public too, but the only proper way to do so locks you into a single project (And also there probably are some personal information in them). There is an outdated version at https://git.ebc.li/makeshift/makeshift, but it went through some changes at my end, mainly adding error handling and notifications. There are a few screenshots on my Mastodon if you want to visualize the jank going on here: https://toot.ebc.li/@admi/105132131906807252 -- Have a nice /(day|night|week(end)?)/ ~ Ecmel B. Canl?er
I write my micrologs by hand. One document hosts every little updates under subheadings. My wiki outputs its feeds to www and gmi[0][1]. I have no intention to port the whole wiki to gemini, but I began a log last week. Since it's mostly text, I just convert them manually to gmi[2]. I use rsync to upload everything, but I'm trying to find something smart that could also do it from 9front. [0]: gemini://royniang.com/feed.gmi [1]: gemini://royniang.com/journal.gmi [2]: gemini://royniang.com/log/
I use gssg [1]. It's the first thing I tried for this and have only been using it for a few days now for my shiny new gemlog [2]. So I can't really compare and contrast other solutions, but I've been enjoying it so far. Basically, I make a new post in `./content/posts/YYYY-MM-DD-title.gmi` and write the things that I want. Then I commit and push to sr.ht. The build runs a generate and pushes the `./public/` output up to S3. Then my server runs a sync every hour to pull down any new posts. So it's relatively low-touch from my end, which I like. I think there's still some enahncements I can make, but overall I can't complain. [1] https://sr.ht/~gsthnz/gssg/ [2] gemini://gem.iwritethe.codes/
On 2020-12-12 12:52PM, Bj?rn W?rmedal wrote: > I write my posts in vim, manually add an entry to my archive gemini > feed and then run a bash script that updates my capsule index page and > my atom feed. It?s a bit clunky, and I?m pondering how to make it less > cumbersome. This is basically what I do. I have a `template.gmi`[a] that has the footers and such that I like, so I just copy that, and write my post in (neo)vim, replacing the placeholders in my template. I then just run my gemlog.sh utility[b] to generate the atom feed and index. I don't find it particularly cumbersome, and I actually manage the entirety of the rest of my site 100% manually, without any scripts at all. I think a big strength of gemini is that a full capsule can be managed manually, or with very simple scripts. I've never felt the need for blog generators or md->gemtext anything like that. I've been considering using CGI for my index and atom feed, so I don't even need to run a script after writing a post, but I've never gotten around to it. A while ago I patched bashblog to include gemini support[c], if anyone is interested in adapting that and writing a cgi atom feed then feel free, and send it to the list when it's done :) [a]: https://github.com/nytpu/gemini.nytpu.com/blob/master/gemlog/template.gmi [b]: https://git.sr.ht/~nytpu/gemlog.sh [c]: https://tildegit.org/team/bashblog/src/branch/master/bb.sh#L892 -- Alex // nytpu alex at nytpu.com GPG Key: https://www.nytpu.com/files/pubkey.asc Key fingerprint: 43A5 890C EE85 EA1F 8C88 9492 ECCD C07B 337B 8F5B https://useplaintext.email/ -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 833 bytes Desc: not available URL: <https://lists.orbitalfox.eu/archives/gemini/attachments/20201212/f80c 85d6/attachment-0001.sig>
Le 12/12/2020 ? 12:52, Bj?rn W?rmedal a ?crit?: > I?m curious about how authors structure their workflows in geminispace. My own is somewhat of a work in progress ? > > What?s your workflow like? I use somewhat of a complicated system for the lolz, but the content itself is managed through kiln (https://git.sr.ht/~adnano/kiln) which makes life much easier for writing (I also use nano because I'm that kind of person). Then I rclone / rsync / sitecopy parts of the generated content into three DreamObjects buckets (some to an FTP server) to get served from different self-hosted servers that I've mounted those buckets to serve the content in to the wild. If you're interested to get a bit of an idea of what kind of wild system I've got up, I've got some info here : gemini://oh.mg/project/ https://oh.mg/project/ -K?vin
It was thus said that the Great Bj?rn W?rmedal once stated: > I?m curious about how authors structure their workflows in geminispace. My > own is somewhat of a work in progress ? ... > What?s your workflow like? I email my entries in. A program accepts the email and adds the entry to the archive, which makes it available for the web, gopher and gemini. -spc
On Sat, Dec 12, 2020 at 12:52:29PM +0100, Bj?rn W?rmedal wrote: >What?s your workflow like? I have a pretty typical setup for git + static sitegen. I just finished switching my weblog and gemlog (seirdy.one) from a shell script to a proper Makefile; code at https://sr.ht/~seirdy/seirdy.one I use Hugo to build both sites, but I tailor the Markdown/Gemtext sources to flow well. Handling multiple output formats is one of Hugo's main advantages. Essentially, Hugo reads output formats for each page and generates the results (inc. an RSS feed)in the build directory. "make build deploy" builds and deploys both sites with rsync + zstd. I don't actually run "make build deploy" myself; that command runs in CI. So I guess you could say that I publish to my gemlog by running "git commit && git push" and waiting for CI to run "make build deploy" I went with a Hugo-based setup when I was building an earlier version of my site with Hugo and I found a post in Drew Devault's gemlog about Gemini and Hugo: gemini://drewdevault.com/2020/09/27/Gemini-and-Hugo.gmi I wouldn't recommend this setup unless you share content between a weblog and gemlog; Hugo is great for websites, but overkill for a simple gemlog. My VPS runs either agate or gmnisrv, depending on my mood. I should probably do an obligatory "meta" post sometime to explain the whole process. /Seirdy -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 898 bytes Desc: not available URL: <https://lists.orbitalfox.eu/archives/gemini/attachments/20201212/2563 8c43/attachment.sig>
I create a new post using `make new TITLE="the title"`. It sends a template through envsubst to pre-fill some things then writes it to a well-named file in a drafts folder. When I'm finished writing, I move the document to the proper directory. I then run `make update` to generate updated indexes and atom feed. Then I commit and push the changes. To deploy, I run `make publish` which ssh's into the server with agent forwarding and does a git pull. I don't do http or gopher, so there is none of that to deal with. -- Jon
---