One step further to use emacs as a window manager and get rid of other tools :)
Created attachment 249537 [details] build.log Two problems that should be solved before we can include this package: 1. The upstream distfile at http://git.savannah.gnu.org/cgit/identica-mode.git/snapshot/identica-mode-1.0.tar.gz gives me a different checksum on each download. Looks like it is dynamically generated. 2. Processing of the texinfo documentation fails. See attached build.log.
Committed to Emacs overlay: http://overlays.gentoo.org/proj/emacs/changeset/1523 Please test.
Ok, AFAICT, it looks fine after I have added (require 'cl) in ~/.emacs, using emacs-23.2 I don't know if it could be added in the gentoo-site file directly, or by editing the identica-mode.el before compilation, here is the only related information I have found: # grep -n "'cl" identica-mode.el 69:(eval-when-compile (require 'cl)) Second, Maybe an elog/einfo could tell the user to define in ~/.emacs or ~/.emacs.d/, even if i don't like the idea of having passwords in clear-text on the filesystem : (setq identica-username "user") (setq identica-password "mypwd")
(In reply to comment #3) > Ok, AFAICT, it looks fine after I have added (require 'cl) in ~/.emacs, using > emacs-23.2 That is a bug. I can confirmed it and emailed to Gabriel. git blame says that in commit 8a9e6874366 Gabriel changed -(require 'cl) +(eval-when-compile (require 'cl)) So maybe for emacs-23.2 one just needs (require 'cl) back? > I don't know if it could be added in the gentoo-site file directly, or by > editing the identica-mode.el before compilation, here is the only related > information I have found: > # grep -n "'cl" identica-mode.el > 69:(eval-when-compile (require 'cl)) Let's hope for an upstream fix. > Second, Maybe an elog/einfo could tell the user to define in ~/.emacs or > ~/.emacs.d/, even if i don't like the idea of having passwords in clear-text on > the filesystem : > (setq identica-username "user") > (setq identica-password "mypwd") There is also the possibility to have such things in a ~/.netrc file. It is still unencrypted, but at least not in .emacs. I think some time ago I also saw people describing encrypted ~/.netrc files in EmacsWiki. But in general: +1 for some elog
Here is the full list of requirements (of the current git HEAD): (eval-when-compile (require 'cl)) (require 'xml) (require 'parse-time) (require 'longlines) (require 'url) (require 'url-http) (require 'json) Note that json.el is not part of the emacs-22 distribution. We either need to adjust the dependency of the ebuild or distribute json.el ourselves.
(In reply to comment #4) > That is a bug. I can confirmed it and emailed to Gabriel. git blame says > that in commit 8a9e6874366 Gabriel changed > > -(require 'cl) > +(eval-when-compile (require 'cl)) > > So maybe for emacs-23.2 one just needs (require 'cl) back? Yes, this should fix it. > Let's hope for an upstream fix. That would be even better. CL is quite heavy. > > (setq identica-username "user") > > (setq identica-password "mypwd") > > There is also the possibility to have such things in a ~/.netrc file. It is > still unencrypted, but at least not in .emacs. I think some time ago I also > saw people describing encrypted ~/.netrc files in EmacsWiki. But in general: > +1 for some elog Added to the ebuild in the overlay. I've just mentioned .emacs though. (In reply to comment #5) > Note that json.el is not part of the emacs-22 distribution. We either need > to adjust the dependency of the ebuild or distribute json.el ourselves. Thanks for pointing this out. The ebuild now has NEED_EMACS=23. (Generally, we don't add new packages for things that are already part of a released Emacs version. There are some exceptions like org-mode though.)
Hi, I think we are in good shape now. The (require 'cl) statement is fixed upstream by undoing the commit I mentioned. Your patch for the docmentation was also applied. In summary: If you snapshot again today then we should be good to go with the 'vanilla' version. Should I? Will you?
(In reply to comment #7) > Hi, I think we are in good shape now. The (require 'cl) statement is fixed > upstream by undoing the commit I mentioned. Your patch for the docmentation was > also applied. In summary: If you snapshot again today then we should be good > to go with the 'vanilla' version. Should I? Will you? Any chance that we get a release (or at least a proper version tag) in a timely manner? I really dislike snapshots and I'd much prefer something blessed by upstream.
identica-mode-1.1 committed to portage tree. Thanks Cyprien and Thomas.