Ihor Radchenko <yanta...@gmail.com> writes: > org-git-version is very useful when people report bugs. > M-x org-submit-bug-report supplies org-git-version output for bug > subject. Thus, we can easily check which git commit their build > corresponds to.
Let's keep `org-git-version'. If we manage to release Org more often (minor and major versions), I doubt keep `org-git-version' will remain that useful in practise, though. Let's revisit the topic then. > I am not very familiar with all the code paths our Makefile and > autoloads take from setting ORG-VERSION to generating the appropriate > Elisp constants. > > However, I do note that mk/targets.mk contains the following: > > ifneq ($(wildcard .git),) > ORGVERSION ?= $(subst release_,,$(shell git describe --match release\* > --abbrev=0 HEAD)) > ifeq ($(ORGVERSION),) > # In elpa.git, there are no tags available. Fall back to using > # the org.el header. > ORGVERSION := $(patsubst %-dev,%,$(shell $(BATCH) --eval "(require > 'lisp-mnt)" \ > --visit lisp/org.el --eval '(princ (lm-header "version"))')) > GITVERSION ?= $(ORGVERSION)-g$(shell git rev-parse --short=6 HEAD) > else > GITVERSION ?= $(shell git describe --match release\* --abbrev=6 HEAD) > endif > GITSTATUS ?= $(shell git status -uno --porcelain) > else > -include mk/version.mk > GITVERSION ?= N/A > ORGVERSION ?= N/A > endif > > Note that we already have a way to parse Org version from lisp/org.el, > similar to what the commit you referenced does. > It is just that this code path is not used by default. I'd favor using it by default. When using Org from the main branch of the git repository, M-x org-version RET should return this: "Org mode version 9.6-dev (release_9.5.5-822-g0a6a56 @ [load-path])" > We can remove the current default and simply use > > ORGVERSION := $(patsubst %-dev,%,$(shell $(BATCH) --eval "(require > 'lisp-mnt)" \ > --visit lisp/org.el --eval '(princ (lm-header "version"))')) > GITVERSION ?= $(ORGVERSION)-g$(shell git rev-parse --short=6 HEAD) > > all the time. > > I do not know if more involved fix is required (because I am not > familiar enough with the relevant code). Can you provide a patch for the above suggestions? I'll test and see if more fixes are needed, even though I'm also not that familiar with the code either. -- Bastien