NomadNet Browsing¶
The Browser tab opens NomadNet pages (written in Micron) from nodes on the network.
- Finding nodes: the Browser tab has a pane with two lists: Saved (pages
you saved with
s) and Nodes (every NomadNet node heard, newest first). Pages open beside the pane. You can also go to any address (g), or to your home page (H, set in the settings). - On a phone: scrolling a page down in the web UI shrinks the top bars to the back arrow, so the page gets the screen; scroll up a little, or back to the top, to bring them back.
- Searching:
/(or a click on the line under the lists' tabs) searches both lists by name or address, as the Network tab's search does: every word typed must match, and addresses can be pasted as<hash>orhash:/page/index.mu. The lists narrow as you type, matches are highlighted, a node or page found by its address shows it, and the tabs count the matches ("Nodes 4/120").Enterkeeps the search while you move through the list, andEscclears it. In the web UI, the search box is under the tabs, and finds any node heard, not only the 200 newest listed. - Saving a node: with the Nodes list in hand,
ssaves the selected node's home page without opening it (with the page in hand,ssaves the page). In the web UI, the ☆ beside a node saves it, and ★ removes it. Either way it's the same as saving the home page once open, and a node whose home page is saved is starred. - Micron: most of it is supported, including 24-bit colour, alignment, dividers, literal blocks, links, forms (text fields, checkboxes and radio buttons), inline images and page colours. Long lines wrap between words. Pages look as they do in NomadNet: a line's background colour fills its row, and headings have NomadNet's colours. Also, as NomadNet's Micron guide describes them:
- Tables between
`tlines (`tc30centres one and keeps it within 30 columns), with columns aligned by a| --- | :-: | --: |row, which also makes the row above it a header. Columns are as wide as their widest cells where they fit; otherwise the widest give way and their cells wrap. Links and fields in cells work as anywhere else. - Collapsible sections: a heading written
`+>starts open, and`->starts folded, marked ▾ and ▸ (or a page's own marks, from#!fold).Enteron the heading, or a click, folds or opens it. - Anchors: every heading is one (its text in lowercase, with hyphens
between words), and
`:nameadds one anywhere. A link to#namescrolls to it,#alone scrolls to the next heading, and a link to another page withanchor=nameamong its variables opens that page there. - Partials:
`{address`seconds`fields}loads part of a page on its own once the page shows (with the fields and variables it names), and again every so many seconds if it says so, while the page is on screen. A link top:idreloads the partials with thatpid=idamong their variables. What you typed in the page's fields stays when a partial loads. Partials inside partials aren't loaded. A partial loads only from the page's own node, and never a file. The setting Partials from other nodes (off by default) lets a page's partials come from other nodes too: they are never identified to you, never come from your own node, and the node asked learns that you opened the page. - Text fields of several rows:
`<40x5|notes`>(40 wide, 5 rows). In the TUI the field is edited on one line, with ↵ for each new line; the web UI gives it a text box of that size. - Links: page links,
lxmf@links (open a conversation),rrc://links (open a hub room), and/file/downloads, which are saved todownloads/. - Identifying: pages that personalise content need you to identify.
Toggle it per node with
I. - View source:
u(or the button in the page's title bar) shows the page's Micron with line numbers and the markup coloured. Refreshing keeps the source view, andYcopies the source exactly as the node sent it. - Copying: drag across page text to copy it,
Ycopies the whole page, andLcopies the selected link. - Repositories on a page:
a(Keep repos in the web UI) keeps everyrns://repository the page links to in the Repos tab. - Finding in a page:
f(or Ctrl-F in the TUI, or Find in the web UI) finds text in the page shown, or in its source: every match is marked, andEnter(or↓) goes to the next,↑(orShift+Enterin the web UI) the one before, with how many there are;Escis done. - History:
b(or a right-click) goes back. - Tabs: several pages open at once, each with its own history. A saved
page or node picked from the list (a click or
Enter) opens in a new tab, unless the tab shown is empty, which takes it, or a tab is already on that node's home page, which is shown again;Tin the list opens it in the tab shown instead. With the setting Open pages in new tabs (Settings > Browsing) off, it's the other way round. On a page,Topens the selected link in a new tab (with nothing selected, the page shown again),Wcloses the tab,[and]go to the tab before or after, andAlt+1toAlt+9to that tab. The tabs are named across the top of the address bar (the node, and the page unless it's the node's home page), and a click on one shows it; with the setting off, only once there's more than one. Ctrl+click (⌘+click on a Mac, in the web UI) or a middle-click on a link opens it in a new tab, and on a saved page or node does the other way from a click; + after the tabs opens an empty tab (on a phone, New tab behind ⋯), and × on a tab, or a middle-click on it, closes it. Ctrl+click needs a terminal that passes it on (most do; some keep it for their own links). A page loading in a tab out of sight goes on loading, and is there when you come back. Up to 12 tabs can be open; with many, their names are cut short to fit. - A node that won't answer: while a page loads in the TUI, the address
bar says how finding a path goes ("path request 2 of 3"). A node not heard yet
can take up to a minute. One whose path went stale (it moved, or a node
on the way went down) gets a fresh path asked for once its Link goes
unanswered. To check a node by hand, find its path (
P) or probe it (T) in the Network tab; see Network. - Cache: pages and images are cached on disk for
cache_hours(24 by default). A page's#!c=directive can shorten that, and#!c=0pages are never cached. Form submissions and file downloads always go to the network. - The address bar shows when a page came from the cache.
rrefetches the page and its images. The cached copy is only replaced once the new one arrives, so it still works while a node is unreachable.Rclears the whole cache.- Limits, so a node can't fill the disk or memory: a page or image over 8 MB is refused, an entry over 4 MB isn't cached, and past 128 MB the oldest entries go. A page loads at most 24 images and 16 partials, and a partial reloads no more often than every 5 seconds (the setting Shortest partial reload changes that, down to 1 second; a partial still reloads at most once a day).
- Images:
- Page images and image attachments are drawn with ratatui-image in the best protocol the terminal has: Kitty graphics (Kitty, Ghostty and others), Sixel, iTerm2 inline images (iTerm2, WezTerm), or half blocks everywhere else. They scroll and clip like text, and are decoded in the background so the UI never waits for them, two at a time. A picture larger than 16 MB, or wider or taller than 16384 pixels, isn't shown.
- Windows Terminal (detected by
WT_SESSION, which is also set inside WSL) always gets Sixel. It may not report its cell size, which Sixel needs to size pictures: if they come out too big or small, setRETTUI_CELL_SIZEto your font's cell size in pixels, for exampleRETTUI_CELL_SIZE=9x19. - rettui asks the terminal what it supports at startup with a short, bounded query (at most 1.5 s, and only if the terminal doesn't answer). Konsole's Sixel and Kitty support and WezTerm's Kitty support are skipped, as ratatui-image recommends.
RETTUI_GRAPHICSforces a protocol:kitty,sixel,iterm2orhalfblocks.