Title: RE: Stale lock files with anonymous CVS
FYI I've also started seeing some strange signal 11 errors to do with lock files since 1.11.7 (same in .8 and .9). Yet to get to the bottom of it. I'll let you know when I have something more specific.
NOTICE AND DISCLAIMER:
This email
Damn. I'll try doing it in plain text next time.
NOTICE AND DISCLAIMER:
This email (including attachments) is confidential. If you have received
this email in error please notify the sender immediately and delete this
email from your system without copying or disseminating it or placing any
Sorry for the spam, but this is something I feel strongly about.
-Original Message-
From: Andy Baker [mailto:[EMAIL PROTECTED]
Subject: Save the bunnies
== 'bcc' option used
Hi Folks,
A good friend of ours has run a rabbit rescue centre for 12 years taking
in and attempting to rehome
Don't use CVS?
-Original Message-
From: Rabbit Xu [mailto:[EMAIL PROTECTED]]
Sent: 24 May 2002 02:04
To: [EMAIL PROTECTED]
Subject: How to lock a file and don't allow multiple people to modify it
at the same time.
Hello all:
Becase I don't allow multiple people to modify at the same
Can somebody explain why CVS doesn't call the script
specified in $CVSROOT/verifymsg file when the user
enters a null message?
Surely this is self defeating when we're trying to use
verifymsg to ensure that the log message entered by the
user makes some sense?
Have I missed something obvious?
Thanks both. I agree that there shouldn't really be empty directories. It's
just that we're creating a new module structure and copying files from old
modules to new modules and the developers wanted the directory structure in
place before they started. They'll just have to use update.
Ta
Andy
Can someone explain the logic behind why 'co -r' turns on the '-P' (prune)
option?
To get empty dirs checked out locally (using a tag) I'm having to 'update
-d' after my 'co -r'
TIA
Andy
NOTICE AND DISCLAIMER:
This email (including attachments) is confidential. If you have received
this
Yes please Greg...
-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]
Sent: 07 January 2002 21:39
To: [EMAIL PROTECTED]
Cc: Duncan Sommerville
Subject: Re: conflict algorithm
snip
If anyone's interested I have changes to the sanity.sh tests to
accomodate my change to
cvsstat (http://cvs.sslug.dk/cvs2html/) might be worth looking at
-Original Message-
From: Michael Reed [mailto:[EMAIL PROTECTED]]
Sent: 30 November 2001 18:07
To: [EMAIL PROTECTED]
Subject: listing un-added files?
Hello,
Can anyone recommend a way to list files (or directories) that
cvs -n status works fine in my commitinfo script. Are you running the latest
version of CVS?
-Original Message-
From: Bill Buie [mailto:[EMAIL PROTECTED]]
Sent: 29 November 2001 17:41
To: [EMAIL PROTECTED]
Subject: CVS commitinfo trigger question
So far I have not found anyone who can
Do you really want to delete the old branch? Might there not be revisions on
it you want to keep?
In the cause of safety it may be better to rename the old branch, then
create the new one using a variation on the following
$ cvs admin -nRENAMED-BRANCH:BRANCH file
$ cvs
In article [EMAIL PROTECTED], Andy Baker wrote:
I might be being more thick than usual today,
but shouldn't the following two examples
essentially produce the same result?...
cvs up -jrev1 -jrev2 file
cvs diff -rrev1 -rrev2 file patch
patch file patch
The patch program does not handle three
http://www.cvshome.org/docs/manual/cvs_16.html#SEC134
-Original Message-
From: Mark [mailto:[EMAIL PROTECTED]]
Sent: 01 November 2001 12:44
To: CVS Mailing List
Subject: CVSROOT/config
Where can I find out what each letter for history logging (TOFEWGCMAR)
available for the LogHistory
Belay that... Worked out the CVS (diff3) is doing a /much/ better job
-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]
Sent: 31 October 2001 20:08
To: [EMAIL PROTECTED]
Subject: Re: diff-iculties
In article [EMAIL PROTECTED], Andy Baker wrote:
I might be being more
I might be being more thick than usual today,
but shouldn't the following two examples
essentially produce the same result?...
cvs up -jrev1 -jrev2 file
cvs diff -rrev1 -rrev2 file patch
patch file patch
__
Andy Baker
Prepay Development, One2One
Maybe some variant of the cluggy and ugly...
$ cvs up -Cp file /tmp/$$file ; diff args file /tmp/$$file
?Andy
-Original Message-
From: Matthew Versluys [mailto:[EMAIL PROTECTED]]
Sent: 26 September 2001 07:17
When comparing the working version of a file against a version in
It's not pretty, but this'll give you what you want...
$ cvs -n co -p -rmytag2 . 21 1/dev/null | awk '/^Che/ {f=$3}
/^VER/ {printf(%s %s\n,f,$2)}'
Aj
-Original Message-
From: Gary Farmer [mailto:[EMAIL PROTECTED]]
Sent: 12 September 2001 21:46
To: [EMAIL PROTECTED]
Subject:
I don't understand why a conflict was returned on a keyword when you used
-kk ?
Andy
-Original Message-
From: Venessa Williams [mailto:[EMAIL PROTECTED]]
/*
a.java
* $Header: a.java,v 1.11 2001/04/30 21:31:33 user1 Exp
$
=
* $Header: a.java,v 1.9.2.1 2001/06/15 16:45:05 user2
I give up... can't find any references to ,t files anywhere... what /are/
they???
$ cvs status cvs_tag ; ls -l CVS
===
File: cvs_tag Status: Up-to-date
Working revision:1.1.2.5 Fri Jun 29 13:24:49 2001
...but as you can see, I have committed. Indeed I'm even up to a fifth
revision. Strange. Guess something must have hiccuped and left the file
lying around.
Thanks
Andy
-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]
They hold the creation log message for newly
Something along the lines of the quick'n'dirty below should get you started
(beware of files that don't contain the branch though)
cvs -Q rlog -rbranch file 2/dev/null \
| awk '/^RCS file: / {file=$3}
/^revision [0-9][0-9]*\./ {rev=$2}
/^date: .* author:/
Must be a pretty fundamental change to get the result Mark wants. Is there
any documentation for this patch? I'd like to apply it but I'd rather
appreciate what it does before applying it.
Andy
-Original Message-
From: Derek R. Price [mailto:[EMAIL PROTECTED]]
There's still a patch up
Thanks Derek. I'll give that a go
Andy
-Original Message-
From: Derek R. Price [mailto:[EMAIL PROTECTED]]
http://alumni.engin.umich.edu/~oberon/ccvs.newfmtstrings.1-11.diff
NOTICE AND DISCLAIMER:
This email (including attachments) is confidential. If you have received
this
Ah. I see (next time I'll read the posting properly!). Good 'un that.
Nothing obvious springs to mind.
-Original Message-
From: Mark [mailto:[EMAIL PROTECTED]]
Since taginfo is in a sense, a pre-op trigger, the tag won't be
placed in the RCSfile until the taginfo script successfully
My experience with cvs would suggest that it's always a good idea to trap
stdout and stderr and examine for potential problems. I've seen situations
where I get a 0 status returned even though something I consider to be a
problem has occurred. That's not a criticism of cvs btw, just that I see
A key point of CVS is the Concurrency. If you need to lock like that then
CVS isn't the tool for you.
As for permissions on specific directories - this has been discussed quite
recently here. Either group your users and set the permissions on the
directories accordingly, or write some sort of
Nah. ISTR that admin -n will only work if it can latch on to something that
already exists
-Original Message-
From: Donald Sharp [mailto:[EMAIL PROTECTED]]
I just tried using the admin -n command in a test repository. I'm
getting:
NOTICE AND DISCLAIMER:
This email (including
cvs update -d
Andy
-Original Message-
From: McCann, Brian [mailto:[EMAIL PROTECTED]]
I know how to create a new directory in the repository by
cvs add
cvs commit
but how can I have the other developers get this directory without checking
out the whole repository again?
NOTICE AND
As a specific tag can only appear once within a file, reapplying the tag to
a file that already contains it will move the tag to the specified revision
- providing you use the '-F' option (see cvs -H rtag).
Andy
-Original Message-
From: Helliwell, Matthew [mailto:[EMAIL PROTECTED]]
AIUI, on the server side you will actually be in the /tmp/cvsblah
directory when your script is called so a quick examination of 'pwd' should
give the game away.
Andy
-Original Message-
From: Schell Walter [mailto:[EMAIL PROTECTED]]
we use cvs server 1.11 on linux and wincvs as client.
Yes, if you can identify the a.b.0.c revision, but tread with care!
Oh, and give the user a slap '-)
Andy
-Original Message-
From: Donald Sharp [mailto:[EMAIL PROTECTED]]
Sent: 31 May 2001 19:18
To: [EMAIL PROTECTED]
Subject: deleting branches.
I have a user who manged to delete
Fairy snuff, though I feel that cvs should still pass the name(s) of the
file(s) the user has requested be operated on - not that it's of great
importance.
Cheers
Andy
-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]
Andy Baker writes:
If cvs has no work to do
I suggest you'll need to write a commitinfo script that'll examine any
commits and reject them if they're on one of the locked codestreams. See
http://cvshome.org/docs/manual/cvs_18.html#SEC165
Andy
-Original Message-
From: Kevin Yang [mailto:[EMAIL PROTECTED]]
Sent: 25 May 2001 22:45
If cvs has no work to do on a tag operation (eg the tag already exists on
the revision), is there a reason why it calls a (given) script defined in
taginfo /without/ passing the file name? It passes all the other info.
Surely it should either not bother to call the script at all or it should
pass
re-presented.
Thanks for the input
Andy
-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]
Sent: 18 May 2001 19:55
To: [EMAIL PROTECTED]
Cc: [EMAIL PROTECTED]
Subject: Re: Is this a bug?
Andy Baker writes:
After editing a log message during the commit process you
After editing a log message during the commit process you quit from the
editor (eg :q! in vi) the following prompt (correctly) appears.
Log message unchanged or not specified
a)bort, c)ontinue, e)dit, !)reuse this message unchanged for
remaining dirs
Action: (continue)
But can you handle the situation where different sets of files have
different valid PR's? I've been trying to figure a way round that for
ages...
Andy
-Original Message-
From: Chris Cameron [mailto:[EMAIL PROTECTED]]
Subject: RE: Triggers links to defect tracking systems
We use a
I'd like to verify a log message depending on branch tag. Can't see a way to
get to the filename (and hence cvs status) via verifymsg?
Any thoughts, or do I have to hack into the cvs code?
TIA
Andy
NOTICE AND DISCLAIMER:
This email (including attachments) is confidential. If you have
Or ksh...
$ cvs commit -m"$(echo "Line One\nLineTwo")" file
Andy
-Original Message-
From: Gianni Mariani [mailto:[EMAIL PROTECTED]]
Sent: 13 April 2001 15:26
To: James A. N. Stauffer; CVS
Subject: RE: commit -m "Line One\rLineTwo"
It seems this depends more on your shell
Can anybody shed some light on this?
I thought I could configure cvs to write lockfiles outside of the standard
repos structure, but having set the $CVSROOT/CVSROOT/config LockDir variable
to /tmp/lockfiles I'm finding that cvs still attempts to create a
$CVSROOT/project-a/,filea, type lock file
Thanks Larry
Andy
NOTICE AND DISCLAIMER:
This email (including attachments) is confidential. If you have received
this email in error please notify the sender immediately and delete this
email from your system without copying or disseminating it or placing any
reliance upon its contents. We
NOTICE AND DISCLAIMER:
This email (including attachments) is confidential. If you have received
this email in error please notify the sender immediately and delete this
email from your system without copying or disseminating it or placing any
reliance upon its contents. We cannot accept
You could try getting commitinfo to call a quick script to check on who's
committing stuff into the repos...
Andy
-Original Message-
From: Fredrik Liljegren [mailto:[EMAIL PROTECTED]]
Sent: 03 May 2000 11:30
To: '[EMAIL PROTECTED]'
Subject: Read-only local (ssh) access
How do I manage
My apologies to everyone in this group who received an out of office auto
reply from me.
I hadn't figured on Outlook97 doing this. Duh.
Andy
44 matches
Mail list logo