On Sat, Aug 22 2020, Nam Nguyen <n...@berkeley.edu> wrote: > Jeremie Courreges-Anglas writes: > >> emacs-27.1 was released earlier than I expected, thanks Ken for the >> heads-up. >> >> This is good news. As mentioned in >> https://marc.info/?l=openbsd-ports&m=159380829911290&w=2 emacs now uses >> a "portable dumper" that doesn't involve games with ELF layout and sbrk >> at build time. No more ld -Z or -nopie flags hampering ASLR, no more >> workarounds in libc, etc. >> >> A few changes regarding deps: >> - libgmp is now a direct dep used for bignums, it was already there >> because of gnutls >> - libjansson is now linked in for LSP (Language Server Protocol) support >> (requested by Timo Myyra). Yes, this introduce an indirect build dep >> on cmake which kinda sucks for non-clang archs. If that's a problem, >> well, it was already the case because emacs->gnutls->cmocka->cmake->argh. >> - harfbuzz is used for test layout >> - bye bye ImageMagick >> >> If you notice changes with highlighting, particularly with vc-diff, well >> you're not alone. I find that some foreground and background color >> changes (maybe not intended) make the resulting frames very hard to >> read. But maybe just in the no_x11/tty case? > > vc-diff looks very hard to read now in no_x11. Some deleted lines are > highlighted and some are not, clashing with the highlighted lines of > additions. Even graphical emacs made vc-diff harder to read a bit with - > and + being red and green to blend in with the highlighted line. Strings > and variables now have colors, too.
Thanks for confirming, this prompted me to look again for the causes. The code changes ./lisp/vc/diff.el point out a few variables. With: (setq diff-refine nil) (setq diff-font-lock-syntax nil) I get diff hilighting behavior similar to emacs-26.3. >> Anyway, test reports welcome. Even if the new dumping mechanism is >> called "portable" I'm not sure it has been tested on the wide range of >> architectures we provide. amd64 and sparc64 seem to behave, what about >> yours? > > I tested the amd64 athena flavor. I tested org-mode, vc-diff, gnus and > rcirc. Org-mode changed a keybinding for <s TAB which usually expands to > a code snippet. The keybinding was changed to C-c C-, > > https://orgmode.org/manual/Structure-Templates.html > > Thank you for the update. Thanks for the report! Now that I'm sure that those changes were intended and not the result of fallout from recent hilighting work done in emacs, I guess I'll just push the update. -- jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF DDCC 0DFA 74AE 1524 E7EE
