Discussion:
Broken w3 subrepo?
Jerry James
2015-10-03 17:21:09 UTC
Permalink
I did an hg pull on the packages to get yesterday's gnus prerelease,
and got nothing, and hg log showed a last top-level update from a
couple of weeks ago:

changeset: 1137:3720091d7106
tag: tip
user: Norbert Koch <***@xemacs.org>
date: Sat Sep 19 00:35:08 2015 +0200
summary: pre-release gnus

I thought that maybe the recent trouble with the top-level had
something to do with it, so I did a fresh hg clone of the package
repository. Mercurial reported success, but several subrepositories
are missing:

xemacs-packages/
xemacs-base
xemacs-devel
xetla
xlib
xslide
xslt-process
x-symbol
xwem
zenirc

In short, all subrepos that lie alphabetically after w3 did not get
checked out. I thought maybe running hg update would fetch the rest:

resolving manifests
getting .hgignore
getting .hgsub
getting .hgtags
getting ChangeLog
getting INSTALL
getting Local.rules.inc
getting Local.rules.mk
getting Local.rules.template
getting MAINTAINERS
getting Makefile
getting NEWS.packages
getting Version.rules
getting XEmacs.rules
getting find-macro-err.awk
getting gen-macro-list.awk
getting hack-package-index.el
getting iterate-1.rules
getting iterate.rules
getting meta-iterate.rules
getting meta-targets.rules
getting mule-packages/Makefile
getting package-clean.el
getting package-compile.el
getting package-net-packages.el
getting xemacs-packages/Makefile
resolving manifests
images/open-cap-up.xbm: untracked file differs
abort: untracked files in working directory differ from files in
requested revision (in subrepo xemacs-packages/w3)

This is from a fresh hg clone; I have not touched anything inside the
w3 subrepo, or any other subrepo for that matter. Also, after running
hg update, I get this:

$ hg status
M .hgsubstate

And hg diff shows the entire contents of .hgsubstate as being added.
Indeed, running "hg revert .hgsubstate" erases .hgsubstate!

Is anybody else seeing this?
--
Jerry James
http://www.jamezone.org/
Henry S. Thompson
2015-10-03 20:35:00 UTC
Permalink
Post by Jerry James
I did an hg pull on the packages to get yesterday's gnus prerelease,
and got nothing, and hg log showed a last top-level update from a
... [lots of lossage ensues]
When, exactly?

I happen to have done a clean pull -u at 1147 BST == 1047 UTC
on Thursday (1 October), and had no problems.

The head at that point was

changeset: 1137:3720091d7106
tag: tip
user: Norbert Koch <***@xemacs.org>
date: Sat Sep 19 00:35:08 2015 +0200
files: .hgsubstate
description:
pre-release gnus

I can't see any of the symptoms you describe.

Where are you pulling from (just to be sure)?

ht
--
Henry S. Thompson, School of Informatics, University of Edinburgh
10 Crichton Street, Edinburgh EH8 9AB, SCOTLAND -- (44) 131 650-4440
Fax: (44) 131 650-4587, e-mail: ***@inf.ed.ac.uk
URL: http://www.ltg.ed.ac.uk/~ht/
[mail from me _always_ has a .sig like this -- mail without it is forged spam]
Mike Kupfer
2015-10-03 23:05:44 UTC
Permalink
I just did a fresh clone from
https://bitbucket.org/xemacs/xemacs-packages and didn't have any
problems.
Post by Jerry James
I did an hg pull on the packages to get yesterday's gnus prerelease,
and got nothing, and hg log showed a last top-level update from a
changeset: 1137:3720091d7106
tag: tip
date: Sat Sep 19 00:35:08 2015 +0200
summary: pre-release gnus
Yes, that's the same tip changeset that I got.
Post by Jerry James
I thought that maybe the recent trouble with the top-level had
something to do with it, so I did a fresh hg clone of the package
repository. Mercurial reported success, but several subrepositories
[...]
Post by Jerry James
In short, all subrepos that lie alphabetically after w3 did not get
[...]
Post by Jerry James
images/open-cap-up.xbm: untracked file differs
abort: untracked files in working directory differ from files in
requested revision (in subrepo xemacs-packages/w3)
I don't know what happened there. What do "hg verify", "hg id", and "hg
tip" report in the w3 subrepo?
Post by Jerry James
This is from a fresh hg clone; I have not touched anything inside the
w3 subrepo, or any other subrepo for that matter. Also, after running
$ hg status
M .hgsubstate
I usually see that when I've committed something in a subrepo but there
hasn't been a top-level commit to incorporate the subrepo change. I
don't know if that's the only way you get those results, though. I
suspect it's related to the problems you saw in the w3 subrepo.

mike
Norbert Koch
2015-10-04 19:28:44 UTC
Permalink
* Jerry James <***@xemacs.org>:

Hi Jerry, all,
Post by Jerry James
I did an hg pull on the packages to get yesterday's gnus prerelease,
and got nothing, and hg log showed a last top-level update from a
changeset: 1137:3720091d7106
tag: tip
date: Sat Sep 19 00:35:08 2015 +0200
summary: pre-release gnus
I see something similar. I have a clone of the repository on a
different machine which sits on the same commit.

I just did a new clone with the same result, i.e. I get everything but
only up to 1137. hg status is clean, so I don't get the strange stuff
you see.

In my package maintainer's repo I am more advanced. To be more precise:

***@vserv:test% hg log -l 7
changeset: 1143:b2d449e85940
tag: tip
user: Norbert Koch <***@xemacs.org>
date: Sun Oct 04 20:55:46 2015 +0200
summary: pre-release packages (tramp)

changeset: 1142:ca49d97d3b1e
user: Norbert Koch <***@xemacs.org>
date: Sun Oct 04 20:49:22 2015 +0200
summary: XEmacs Package Release

changeset: 1141:d4b17ec9982d
user: Norbert Koch <***@xemacs.org>
date: Sun Oct 04 20:48:54 2015 +0200
summary: update packages (tramp)

changeset: 1140:355cf5e105dc
user: Norbert Koch <***@xemacs.org>
date: Fri Oct 02 10:05:22 2015 +0200
summary: pre-release packages (gnus)

changeset: 1139:4b3e5b7f593c
user: Norbert Koch <***@xemacs.org>
date: Fri Oct 02 09:16:23 2015 +0200
summary: XEmacs Package Release

changeset: 1138:105d97d2872f
user: Norbert Koch <***@xemacs.org>
date: Fri Oct 02 09:12:17 2015 +0200
summary: package update (gnus)

changeset: 1137:3720091d7106
user: Norbert Koch <***@xemacs.org>
date: Sat Sep 19 00:35:08 2015 +0200
summary: pre-release gnus


To my knowledge I've pushed everything upstream, so I'm honestly
wondering where the changesets have gone. Could someone with more
insight and more kungfu in mercurial (and bitbucket) please have a look?

Thanks,
norbert.
--
Mike Kupfer
2015-10-04 21:05:26 UTC
Permalink
Post by Norbert Koch
I just did a new clone with the same result, i.e. I get everything but
only up to 1137. hg status is clean, so I don't get the strange stuff
you see.
In my package maintainer's repo I am more advanced.
[...]
Post by Norbert Koch
To my knowledge I've pushed everything upstream, so I'm honestly
wondering where the changesets have gone. Could someone with more
insight and more kungfu in mercurial (and bitbucket) please have a look?
Well, the web page at https://bitbucket.org/xemacs/xemacs-packages shows
the most recent activity as being

8 commits
Pushed to xemacs/xemacs-packages
3720091 pre-release gnus
3ea28b7 XEmacs Package Release
6d529b6 update gnus
2d75ad8 pre-release general-docs
c70deb9 XEmacs Package Release
+ 3 more…
viteno · 2015-09-18

So either your push went somewhere else (what does "hg path default"
say?), or Bitbucket lost the changesets.

It should be safe to just push again.

mike
Norbert Koch
2015-10-05 06:03:58 UTC
Permalink
* Mike Kupfer <***@xemacs.org>:

Hello Mike,
Post by Mike Kupfer
Well, the web page at https://bitbucket.org/xemacs/xemacs-packages shows
the most recent activity as being
8 commits
Pushed to xemacs/xemacs-packages
3720091 pre-release gnus
3ea28b7 XEmacs Package Release
6d529b6 update gnus
2d75ad8 pre-release general-docs
c70deb9 XEmacs Package Release
+ 3 more…
viteno · 2015-09-18
So either your push went somewhere else (what does "hg path default"
say?), or Bitbucket lost the changesets.
hg path default looks good
Post by Mike Kupfer
It should be safe to just push again.
You know, been there, done that. But to no avail, all clean.

norbert.
--
Michael Sperber
2015-10-05 16:42:28 UTC
Permalink
Post by Jerry James
images/open-cap-up.xbm: untracked file differs
abort: untracked files in working directory differ from files in
requested revision (in subrepo xemacs-packages/w3)
What does "hg status" say in w3?
--
Regards,
Mike
Mike Kupfer
2015-10-05 20:12:10 UTC
Permalink
Post by Norbert Koch
hg path default looks good
Post by Mike Kupfer
It should be safe to just push again.
You know, been there, done that. But to no avail, all clean.
Huh. That's really strange.

I suppose the next step would be to get debug information
("hg --debug <command>").

mike
Norbert Koch
2015-10-12 09:07:01 UTC
Permalink
* Mike Kupfer <***@xemacs.org>:

Hi!
Post by Mike Kupfer
Huh. That's really strange.
I suppose the next step would be to get debug information
("hg --debug <command>").
I've just pushed the latest changes for edit-utils and prog-modes.
Attached is the debug log.

The changesets in the packages are there. I'm missing the one in the
xemacs-package node. I thought there was one earlier.

norbert.
--
Mike Kupfer
2015-10-12 16:56:18 UTC
Permalink
Post by Norbert Koch
I've just pushed the latest changes for edit-utils and prog-modes.
Attached is the debug log.
Thanks. Hmm. I'm afraid my knowledge of Mercurial isn't enough to know
Post by Norbert Koch
broken pipe
remote: Connection to bitbucket.org closed by remote host.
Unless Mike S. has access to server-side logs, I think it's time for a
support request at Bitbucket (https://bitbucket.org/support).

mike
Jerry James
2015-10-12 19:32:22 UTC
Permalink
Post by Mike Kupfer
Thanks. Hmm. I'm afraid my knowledge of Mercurial isn't enough to know
Post by Norbert Koch
broken pipe
remote: Connection to bitbucket.org closed by remote host.
Unless Mike S. has access to server-side logs, I think it's time for a
support request at Bitbucket (https://bitbucket.org/support).
Sorry for my silence on this issue. I tried doing a clean checkout
again myself, and this time it did not get stuck on w3. The checkout
appears to have finished, but I got the connection closed by remote
host message as well.
--
Jerry James
http://www.jamezone.org/
Loading...