5:15 p.m.
The following code (and several others like it) is used to determine the
value of user-init-file,
(concat "~" init-file-user "/.emacs")
However, cygwin32 doesn't seem to handle the ~username convention.
Is there are suggested way to work around this problem? Or should the code
be changed? For instances, the HOME environment variable, if set, could be
used.
(let ((home-directory
(or (getenv "HOME") (concat "~" init-file-user))))
... (concat home-directory "/.emacs") ...
--
Paul Stodghill <stodghil(a)cs.cornell.edu>
http://www.cs.cornell.edu/stodghil/
10:32 p.m.
New subject: 21.0 beta41, cygwin32, ~init-file-user doesn't work.
Paul Stodghill wrote:
The following code (and several others like it) is used to determine
the
value of user-init-file,
(concat "~" init-file-user "/.emacs")
However, cygwin32 doesn't seem to handle the ~username convention.
This is handled internally by XEmacs (well, I presume that it is;
passing ~user/file to open() won't work).
For me (on Linux), init-file-user is "" (i.e. empty), so the above
expression evaluates to "~/.emacs", which should equate to the same
thing as $HOME/.emacs (at least it seems to here).
Does NT have any notion of per-user home directories? If it doesn't,
then ~username isn't really meaningful, I guess.
--
Glynn Clements <glynn(a)sensei.co.uk>
12:33 a.m.
New subject: 21.0 beta41, cygwin32, ~init-file-user doesn't work.
Some time ago, Glynn Clements wrote...
|+
| Does NT have any notion of per-user home directories? If it doesn't,
| then ~username isn't really meaningful, I guess.
|-
Yes and no.
Yes, because there is such a concept. There are HOMEDRIVE and HOMEPATH
variables which specify it. In a well designed network environment,
one can log in on any machine in a domain, and get the home directory
automatically mapped ("mounted" in unix terms). The local system tkaes
an unused drive letter, and maps user's home directory share to it.
User's registry hive is also mapped from the domain controller, so
user's personal settings are available.
No, becuase in a domainless workstation setup home directory defaults
to the root of the system drive.
The first setup is good and trustworthy, we have a real home directory
in normal sense. The second is just bogus.
Kirill
4:27 a.m.
New subject: 21.0 beta41, cygwin32, ~init-file-user doesn't work.
"Kirill M. Katsnelson" <kkm(a)kis.ru> wrote:
Some time ago, Glynn Clements wrote...
|+
| Does NT have any notion of per-user home directories? If it doesn't,
| then ~username isn't really meaningful, I guess.
|-
Yes and no.
Well, would the following patch help? I submitted it back in
March, but it never made it into the source tree (and I didn't get any
feedback why it was bad). Here's the changelog, which describes what it
does:
-------------------------------------------------------------------------------
Wed Mar 11 11:37:17 1998 Darryl Okahata <darrylo(a)sr.hp.com>
* emacs.c: Rearrange NT initialization order so that
environment/registry variables will be properly entered into
Vprocess_enviroment.
* nt.c: If $HOME isn't set, initialize $HOME to be
"$HOMEDRIVE$HOMEPATH", if HOMEDRIVE and HOMEPATH are set (this is
for Windows NT). Otherwise, default HOME to the current
directory, which is what XEmacs currently does (without this
patch).
Should this be changed to force HOME to be C:/, if HOME isn't set?
Setting it to "C:/" is probably consistent with NTEmacs, but it's
different from the way XEmacs currently works (without this
patch).
-------------------------------------------------------------------------------
The attached patch is relative to B30 (yea, this is old). If people
like it, I'll migrate it to B41 and submit it to -patches.
-- Darryl Okahata
Internet: darrylo(a)sr.hp.com
DISCLAIMER: this message is the author's personal opinion and does not
constitute the support, opinion, or policy of Hewlett-Packard, or of the
little green men that have been following him all day.
2:08 p.m.
New subject: 21.0 beta41, cygwin32, ~init-file-user doesn't work.
Some time ago, Darryl Okahata wrote...
|+
| "Kirill M. Katsnelson" <kkm(a)kis.ru> wrote:
|
| > Some time ago, Glynn Clements wrote...
| > |+
| > | Does NT have any notion of per-user home directories? If it doesn't,
| > | then ~username isn't really meaningful, I guess.
| > |-
| >
| > Yes and no.
|
| Well, would the following patch help? I submitted it back in
| March, but it never made it into the source tree (and I didn't get any
| feedback why it was bad). Here's the changelog, which describes what it
| does:
Thanks Darryl, this is the right change. If NT sets the evniromanet,
why do not use it?
Btw, do these variables exist on '95?
| Should this be changed to force HOME to be C:/, if HOME isn't set?
| Setting it to "C:/" is probably consistent with NTEmacs, but it's
| different from the way XEmacs currently works (without this
| patch).
I do not think so. I think that XEmacs must generate warnings like "Cannot
determine home directory" or such. Otherwise, if things go silently,
it is 1) terrible mess having C:/.xemacs/* stuff 2) unobvious place
to put .emacs in. ("Wherever I put .emacs, XEmacs ignores it". Personally,
I would try C:\ *after* anything else, if tried at all...)
| The attached patch is relative to B30 (yea, this is old). If people
| like it, I'll migrate it to B41 and submit it to -patches.
Please do! :)
Kirill
9680
days inactive
9681
days old
4 comments
4 participants
participants (4)
-
Darryl Okahata -
Glynn Clements -
Kirill M. Katsnelson -
Paul Stodghill