Skim v1.0.15 β "Room to Breathe" π
The one where your email finally gets to use the whole window instead of huddling in a 840px column on the left like it's 2009. π₯οΈ
Every fix below is the work of @gildas-refacto in #29 β who not only found three pixel-level bugs we'd been staring past for months, but wrote commit messages more thorough than most postmortems. Thank you. πββοΈπ
π Fixed
The reading pane is now the reading pane. We were capping the message at 840px wide, which meant a big monitor bought you one skinny column of text, a lot of empty pane, and a scrollbar floating somewhere in the middle of the screen. Cap's gone. The mail gets the room. πͺ
Wide emails are no longer quietly guillotined. Some senders lay their newsletters out wider than that old column allowed, and the overflow justβ¦ didn't exist. No scrollbar, no hint, no right-hand edge. Now a too-wide message scrolls sideways inside its own frame like a civilized document. βοΈβ
The message stopped growing 8 pixels every time you resized the window. Turns out the frame was measuring its own height, adding a margin, then measuring that, then adding the margin again β a small ouroboros of blank canvas piling up under your email, one resize at a time. 788 β 812 β 836 β 860, same mail. It now measures the actual content, lands flush, and stays put. π
Bonus side effect: with that 8px gone, the mail's rounded bottom corners are rounded again. Small thing. Bugged us. π
Verified across 120+ real messages, two narrowβwide cycles each: zero inner scrollbars, zero height drift.
π§Ή Housekeeping
Scoop manifest caught up to 1.0.14, just in time to be out of date again. π¦
π Contributors
@gildas-refacto β the entirety of this release (#29), and the AI reasoning-liveness work back in #28. Two releases running. We're just holding the tag machine at this point. π
Get it here: https://github.com/nikserg/skim/releases/tag/v1.0.15


Replies
so when you say wide emails scroll sideways now, does that apply to html newsletters with fixed table widths too?
curious why the height was being remeasured post-margin instead of caching the initial layout, seems like it'd have caught this in testing.