7:04 p.m.
PATCH 21.5
Over on xemacs-beta, Stephen J. Turnbull <stephen(a)xemacs.org> wrote:
Why not just make it a commit-time constant? That mailbox is great!
:-) I hoped somebody would like that. My favorite of all time,
though, was the email address of Patrice Godefroid when he was at Bell
Labs: god(a)att.com.
No, I don't see any reason to keep it. AFAIK the reason to have
it in
the first place was to make it sort of possible to identify the build
that compiled the elc, but the VCS id is much more accurate about the
code base.
But there are going to be files that vary anyway; I'm pretty sure
there's code protected by compile-time features (like Mats's
gnus-use-idna), even in core.
Sure. The use case of interest is compiling on systems that vary only
in CPU architecture. They will have the same sets of software
installed, with exactly the same versions, etc. This change will make
it possible to verify that we're getting identical results in such
cases, which means the files can be shared across all architectures
(we can produce a "noarch" package, to use RPM terminology).
Anyway, here's the patch.
diff -r be31f7878b0d lisp/ChangeLog
--- a/lisp/ChangeLog Tue Sep 23 16:50:48 2014 -0600
+++ b/lisp/ChangeLog Thu Sep 25 11:58:18 2014 -0600
＠＠ -1,3 +1,8 ＠＠
+2014-09-25 Jerry James <james(a)xemacs.org>
+
+ * bytecomp.el (byte-compile-insert-header): do not insert email
+ addresses or timestamps in comments to get reproducible builds.
+
2014-08-05 Aidan Kehoe <kehoea(a)parhasard.net>
* keymap.el:
diff -r be31f7878b0d lisp/bytecomp.el
--- a/lisp/bytecomp.el Tue Sep 23 16:50:48 2014 -0600
+++ b/lisp/bytecomp.el Thu Sep 25 11:58:18 2014 -0600
＠＠ -2156,11 +2156,7 ＠＠
finally return res)))
(setq comments
(with-string-as-buffer-contents ""
- (insert "\n;;; compiled by "
- (or (and (boundp 'user-mail-address) user-mail-address)
- (concat (user-login-name) "＠" (system-name)))
- " on "
- (current-time-string) "\n;;; from file " filename "\n")
+ (insert "\n;;; compiled from file " filename "\n")
(insert ";;; emacs version " emacs-version ".\n")
(insert ";;; bytecomp version " byte-compile-version "\n;;; "
(cond
--
Jerry James
http://www.jamezone.org/
_______________________________________________
XEmacs-Patches mailing list
XEmacs-Patches(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-patches
1:57 a.m.
Jerry James writes:
PATCH 21.5
Over on xemacs-beta, Stephen J. Turnbull <stephen(a)xemacs.org> wrote:
> Why not just make it a commit-time constant? That mailbox is great!
:-) I hoped somebody would like that. My favorite of all time,
though, was the email address of Patrice Godefroid when he was at Bell
Labs: god(a)att.com.
$ host olympus.net
CNAME att.com
<wink/>
On the DMARC front, I've been hanging out with a guy who styles
himself superuser(a)gmail.com, which is arguably bigger than god＠att
these days. :-) (Except of course Googlers get ＠google addresses.)
Sure. The use case of interest is compiling on systems that vary
only
in CPU architecture. They will have the same sets of software
installed, with exactly the same versions, etc. This change will make
it possible to verify that we're getting identical results in such
cases, which means the files can be shared across all architectures
(we can produce a "noarch" package, to use RPM terminology).
OK that seems like a good use case.
(with-string-as-buffer-contents ""
- (insert "\n;;; compiled by "
- (or (and (boundp 'user-mail-address) user-mail-address)
- (concat (user-login-name) "＠" (system-name)))
I'm not sure I agree with this, though. Why not have the build do
xemacs -eval '(setq user-mail-address "buildbot(a)builder.com")'
I think it's useful to be able to easily distinguish between packaged
compiles and local ones: "That doesn't sound like our build. Check
the first line of the .elc; ours always say 'compiled by
buildbot(a)builder.com'." Time is different, there's no point that I
can see in setting time to a constant.
- " on "
- (current-time-string) "\n;;; from file " filename "\n")
+ (insert "\n;;; compiled from file " filename "\n")
(insert ";;; emacs version " emacs-version ".\n")
(insert ";;; bytecomp version " byte-compile-version "\n;;; "
(cond
--
Jerry James
http://www.jamezone.org/
_______________________________________________
XEmacs-Patches mailing list
XEmacs-Patches(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-patches
_______________________________________________
XEmacs-Patches mailing list
XEmacs-Patches(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-patches
2:55 a.m.
On Thu, Sep 25, 2014 at 6:57 PM, Stephen J. Turnbull <stephen(a)xemacs.org> wrote:
$ host olympus.net
CNAME att.com
<wink/>
On the DMARC front, I've been hanging out with a guy who styles
himself superuser(a)gmail.com, which is arguably bigger than god＠att
these days. :-) (Except of course Googlers get ＠google addresses.)
Heh. Still, that's a good one.
I'm not sure I agree with this, though. Why not have the build
do
xemacs -eval '(setq user-mail-address "buildbot(a)builder.com")'
Good point. That would be fine.
I think it's useful to be able to easily distinguish between
packaged
compiles and local ones: "That doesn't sound like our build. Check
the first line of the .elc; ours always say 'compiled by
buildbot(a)builder.com'." Time is different, there's no point that I
can see in setting time to a constant.
Agreed. How about this patch, then?
diff -r be31f7878b0d lisp/ChangeLog
--- a/lisp/ChangeLog Tue Sep 23 16:50:48 2014 -0600
+++ b/lisp/ChangeLog Thu Sep 25 19:54:02 2014 -0600
＠＠ -1,3 +1,8 ＠＠
+2014-09-25 Jerry James <james(a)xemacs.org>
+
+ * bytecomp.el (byte-compile-insert-header): do not insert
+ timestamps in .elc comments to get reproducible builds.
+
2014-08-05 Aidan Kehoe <kehoea(a)parhasard.net>
* keymap.el:
diff -r be31f7878b0d lisp/bytecomp.el
--- a/lisp/bytecomp.el Tue Sep 23 16:50:48 2014 -0600
+++ b/lisp/bytecomp.el Thu Sep 25 19:54:02 2014 -0600
＠＠ -2159,8 +2159,7 ＠＠
(insert "\n;;; compiled by "
(or (and (boundp 'user-mail-address) user-mail-address)
(concat (user-login-name) "＠" (system-name)))
- " on "
- (current-time-string) "\n;;; from file " filename "\n")
+ "\n;;; from file " filename "\n")
(insert ";;; emacs version " emacs-version ".\n")
(insert ";;; bytecomp version " byte-compile-version "\n;;; "
(cond
--
Jerry James
http://www.jamezone.org/
_______________________________________________
XEmacs-Patches mailing list
XEmacs-Patches(a)xemacs.org
http://lists.xemacs.org/mailman/listinfo/xemacs-patches
3740
days inactive
3741
days old
2 comments
2 participants
participants (2)
-
Jerry James -
Stephen J. Turnbull