Re: cvs on mandrake...
On Fri, 2003-06-27 at 12:44, Kent E wrote: I cant find any cvs for mandrake, are they available? If you mean the Mandrake distribution of Linux, then yes, absolutely. It's just the standard Linux/Unix distribution of CVS. It's probably already in the standard Mandrake distribution (check your CDs), but I don't know where - I don't use Mandrake myself. Alternatively, go to http://ccvs.cvshome.org/servlets/ProjectDownloadList You probably want one of these: 2182556 cvs-1.11.6.tar.bz2 2003-05-25 dprice 06a5bbcd93a780a3bbec788deb1cf35b CVS source distribution for CVS stable release version 1.11.6 in BZip2 format. 1129560 cvs-1.11.6-cvshome.org.8x.1.i386.rpm 2003-05-28 dprice 093f5e936a49326c5f0d02374216 RedHat Linux 8.0 RPM for CVS version 1.11.6. I would recommend the RPM, and believe that Mandrake can read it. For a more detailed explanation of installing CVS, check the free beta chapter of Essential CVS, available at http://www.oreilly.com/catalog/cvs/ Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. My book 'Essential CVS': published by O'Reilly in June 2003. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: how to send email notifications on commits
On Fri, 2003-06-20 at 19:32, Tiziana Biondo wrote: Hi, I'd like to know how to configure the loginfo file to send email notifications on commits; I am using WinCVS 1.2 on Windows2000... Thanks in advance Tiziana This is one way of doing it: __ If you want to send email notifications on all commits, try putting something like this in the loginfo file in the CVSROOT directory of the repository: ALL mail -s %s [EMAIL PROTECTED] The leftmost text (ALL) matches the directories in the repository. An 'ALL' rule matches every directory, a 'DEFAULT' rule matches anything that doesn't otherwise have a match, and any other rule is a regular expression which matches against the repository directories. The expression ^example, for instance, would match anything where 'example' started the line. The rest of the line is the mail command. It's just using the shell command 'mail' - you may, for security's sake, want to make it /usr/bin/mail or whatever the address of your mail program is. The %s is a variable that the loginfo file parses - see the 'loginfo' node of info cvs, or the relevant part of the Cederqvist manual. There are three loginfo-specific variables. %v and %V are the new and old version numbers, and %s is the filename. [EMAIL PROTECTED] can be replaced with any appropriate mailing list, or you can use CVS administrative file variables, or the contents of a shell script. Jenn V. -- Essential CVS: http://www.oreilly.com/catalog/cvs/ [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: cvs login
On Sat, 2003-06-21 at 03:09, [EMAIL PROTECTED] wrote: OK a bit more progress here now this message: cvs [login aborted]: unrecognized auth response from machine: cvs pserver: cannot open dir/CVS/CVSROOT/config: Permission denied (for the archives, for the next person with this problem): I talked to him offlist, he doesn't have CVSROOT/passwd set up, and SystemAuth in CVSROOT/config is set to 'no'. I suspect that fixing these will at least get him past another hurdle. Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. My book 'Essential CVS': published by O'Reilly in June 2003. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Is it possible to make CVS send email notification withchanges introduced into committed file version
On Thu, 2003-06-05 at 22:38, Uri Abramov wrote: Hi, I have two questions : 1. Is it possible to add new administrative file under CVSROOT in such way that it would be treated by cvs as it treads it's native administrative files. I mean always keep in repository file,v and file. If it' possible how can I do it ? Yes. There's a file called .. I think it's called checkoutlist. One moment... ... yes. Checkoutlist. Look for the node 'checkoutlist' in info cvs, it's under 'administrative files'. Simply add the filename (and maybe path from CVSROOT) in the 'checkoutlist' file in CVSROOT, and it will be treated as an administrative file. 2. I'd like CVS send email to every developer who works on certain project when someone from them remove, add or change file, and more than that I'd like to send in the email the diff between old and new versions of a file. Does anybody knows how it can be done? If yes How ? You can use the cvs watch, or the 'loginfo' file, to do this during commits. See the 'notify' administrative file and the 'watch' commands in 'info cvs', or the 'loginfo' administrative file. You would include 'cvs diff' as part of the command run from 'notify' or 'loginfo'. Pipe the results of the diff into an email, and send the email. (Make sure you background the command, to prevent deadlocks.) Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. My book 'Essential CVS': published by O'Reilly in June 2003. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: web site management / security issues
On Fri, 2003-06-06 at 02:38, Menge, Oliver wrote: My second question is about client/server cvs, which is rather new to me. The internet host where I would like to have the repository is a W2K machine and I have to live with that. Cederquist says: I have not much idea of Windows' security concept, so maybe somebody could advise me, which access method is safe and does not require (too much of) additional software. ext with ssh is my favourite for a default setup. The server doesn't need any extra setup (though I'm not sure how CVS works on Win 2K - I assume you'd be using CVSNT?), but does need an SSH server installed. The clients need an SSH client that's compatible with the SSH server (make sure you use ssh2 on both, or ssh1 on both). They also need the environment variable CVS_RSH set to the ssh program, or if you're using a CVS client such as WinCVS, the appropriate setting in the client set to the ssh program. You may need to set up ssh keys on the client and server, and appropriate logins. I've also found with putty (on Windows clients), I need to run pageant in the background. After that, it should all 'just work'. Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. My book 'Essential CVS': published by O'Reilly in June 2003. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: new file side-branch to side-branch
On Tue, 2003-06-03 at 02:23, Johnson, Susan wrote: I have a set of new files that exist in side branch BranchA and never in the main trunk. I want to create a version in my build branch BranchB. Is it possible to do so without merging from BranchA to the main trunk and creating BranchB off that main trunk version? The main trunk is where all the developers work and BranchA isn't ready to be merged into the main trunk. Yes, quite easily. Instead of merging into a trunk sandbox, merge into a BranchB sandbox. Jenn v. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. My book 'Essential CVS': published by O'Reilly in June 2003. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Query regarding Branches in CVS
On Sat, 2003-04-05 at 01:20, Amit Sood wrote: Hi all, I have one query regarding the working of CVS I am using CVS to maintain my project files and have released one version of it.I have made a branch of the target at this juncture with tag RELBRANCH1 Now simulatneous development is going on parallely on main branch and this RELBRANCH1. Now if I have one file ABC.txt present in both branches,can I lock this file seperately for both the branches.This means if I have locked for main branch it is still unlocked for RELBRANCH1 and other way around. Do you mean 'locked with cvs admin -l' or locked as in marked for editing with 'cvs edit', or locked in some other way? As I understand it, cvs admin -l locks a specific revision. cvs edit marks the file for editing, but I believe it will mark each branch separately. Also if two users make changes to ABC.txt one user make changes to ABC.txt in main branch and other in RELBRANCH1 and these changes at exactly same moment of time will CVS will be able to handle such a scenario successfully. Yes. That's what branches are for. The changes in the branch version of the file will be saved to the branch, the changes in the main version will be saved to the main trunk. Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. My book 'Essential CVS' will be published by O'Reilly in 2003. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: notify
On Sat, 2003-02-01 at 11:04, Monica Li wrote: Hi, I couldn't change the file notify. Notify is under repository/CVSROOT directory. it also has .v. it means I need checkout before make any changes. right? But it's not under any module. How can I checkout this file, change the file, get rid of the # and make it work? cvs checkout CVSROOT/notify edit notify cvs commit Because the notify file is one of the CVS special administrative files, CVS will commit to the ,v form, and also create a plain-text form in CVSROOT. Do not edit the plain-text form in CVSROOT, you have CVS installed so you may as well get revision control on the administrative files. (Exception: don't checkout CVSROOT/passwd) All of this is explained in the 'Administrative Files' node of info cvs. Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. My book 'Essential CVS' will be published by O'Reilly in 2003. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Branch aliasing
On Thu, 2003-01-30 at 22:18, Simon Scott wrote: Does anybody know of any issues with this or better way to achieve what I want? module A: common components. module B: diverse components. When working on common components, use module A. Consider having a 'floating' tag called 'stable'. When working on diverse components, use module B (either -r legacy or -r future). In your build/test script, checkout A -r stable, and B -r whichever, and compile. I'm also sure there's some way you could use ampersand modules in CVSROOT/modules to get what you're after, but I can't figure it out. It'd start with: A: common B: diverse C: A B But I'm not sure how the branching would work. Probably something like cvs checkout -f -r legacy C, where you've developed A without the branch so the -f fetches the head of A. Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. My book 'Essential CVS' will be published by O'Reilly in 2003. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Help Required:
On Thu, 2003-01-30 at 22:54, Amit Sharma wrote: HI All, I am getting a strange warning message while getting the source from CVS. cvs server: cannot read /CVS/research/CVSROOT/users: Too many open files in system. cvs server: cannot open /CVS/research/CVSROOT/notify: Too many open files in system. I have tried to search on the net for this warning message but was not able to find any Please suggest what does it mean If anybody has idea on this. That warning will be coming from the OS, so check your filesystem limitations. I suspect that you're out of inodes - try running df -i from the command line. Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. My book 'Essential CVS' will be published by O'Reilly in 2003. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Bug tracking.
On Tue, 2002-12-10 at 18:17, Mike Ayers wrote: Steven Buroff wrote: I would like to vote for this feature too. This open source. Only patches count as votes. CVS's support for bug tracking is poor to nonexistent and many people have commented on it and requested better support. Tags don't really do it. That's because CVS is not a bug tracking tool. It's an archive system. Only an archive system. If you want to do more than just archiving, you must find tools that do those other things and/or integrate them yourself. Sorry if this sounds harsh, but too many people just don't seem to grasp this. CVS can be integrated with bug tracking systems. Bugzilla is a bug tracker, and on its website it mentions how to integrate CVS and Bugzilla together. Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. My book 'Essential CVS' will be published by O'Reilly in 2003. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Bug Tracking
On Thu, 2002-12-12 at 01:33, Zieg, Mark wrote: CVS's support for bug tracking is poor to nonexistent and many people have commented on it and requested better support. That's because CVS is not a bug tracking tool. It's an archive system. Only an archive system. If you want to do more than just archiving, you must find tools that do those other things and/or integrate them yourself. Accepting the harsh truth of Mike's statement, may I ask the group whether anyone has any particular recommendations or useful script to share which can usefully link Bugzilla-stored issues with CVS revisions? I'm thinking along the lines of bug ID's which may be entered into log comments, etc. I can and will write such a thing if need be, but I'm wondering what wheels already exist that people can personally endorse. It already exists. http://www.bugzilla.org/docs216/html/integration.html Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. My book 'Essential CVS' will be published by O'Reilly in 2003. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Merging in CVS
On Sat, 2002-11-23 at 03:35, MacMunn, Robert wrote: I am new to CVS. I am testing out merging. When I merged 2 files I got extra lines teling me where the merged lines where. Is there any way around this ? This means that the same lines were changed in the two files, and CVS doesn't know which lines the humans want put where. It's called a 'conflict', and it requires human intervention. nightly_kituibuild_sol_flash_LOG_Sep27_12.36.10_151 # # # === #a # # 1.2.2.4 , and were changed in the first file, and a, and were changed in the second file. Both were line numbers foo, foo+1 and foo+2, so CVS couldn't just plug them in as it normally does. The way to prevent this from happening is to ensure that noone edits the same lines of the same revision. :) Yes, that's overly simplistic. If you need more help or a better explanation, ask. Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: Merging in CVS
On Sat, 2002-11-23 at 05:28, MacMunn, Robert wrote: It isn't a slick interface. In Clearcase it is the merge tool itself that gives you the ability to deal with the conflicts easily. Check the various CVS graphical clients - some of them display the conflicts in merges in a merge tool. It sounds to me like that's what you're after. CVS just allows you to do the same thing in your own choice of editor. Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Newbie question re: ssh
On Wed, 2002-11-20 at 00:25, Robert Koster wrote: I have set up a repository, and want to use :ext w/ SSH (without keys). Pros/cons? Con: this isn't possible. ssh operating without keys will forcibly invoke username/password login, which will gag the CVS client. Huh? I haven't tried without keys in a while but I'm pretty sure my client still just asks me for the password and lets me type it Apparently some of the graphic clients don't understand the password prompt. The command line client is fine, though. Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Branches and Dates
On Wed, 2002-11-20 at 10:06, Larry Jones wrote: Eric Siegerman writes: On Tue, Nov 19, 2002 at 10:16:47AM -0500, Larry Jones wrote: You need to specify both the branch and the date. Unfortunately, not all CVS subcommands allow both -D and -r at the same time [...] Why not? Existing inconsistency coupled with profound indecision over exactly how to best provide the capability. Some places in CVS allow you to specify a branch and a date in a single package, like update -j's branch:date syntax. Other places, like update, allow specifying -r and -D at the same time. Still other places, most notably diff, allow -r and -D at the same time, but interpret them as specifying two different revisions rather than jointly specifying a single revision. Trying to unify all that is a daunting task, particulary when you try to ensure that the result will be upward compatible (for all of clients, servers, and users). Thank you for clarifying that. Hrm. Given that some places (such as diff) need to be able to specify -r and -D to mean two distinct revisions, would the branch:date syntax be the most effective answer? And given that it's an expansion of functionality, is there a problem with implementing it in the next version and then allowing other development teams to add it as they go? Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: How to make update -d ignore a top-level directory?
On Tue, 2002-11-19 at 10:02, Shankar Unni wrote: How can I achieve what I want: to be able to: * not have to set CVSROOT in the environment. * Be able to issue an update -d from the root of my work area, and * not get that junk directory? Is there a way? I'm trying to avoid having to do a radical nms-ectomy (:-/) on my repository.. In CVSROOT in your repository there's a configuration option in one of the files - I think it's called config - that is 'TopLevelDir'. Set it to yes, and the directory you're in when you run cvs checkout gets a CVS subdirectory. I believe that will do what you want. Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Moving to Pserver from .rhosts
On Sun, 2002-11-17 at 07:25, Greg A. Woods wrote: [ On Saturday, November 16, 2002 at 12:18:19 (+0100), Fredrik Wendt wrote: ] Subject: Re: Moving to Pserver from .rhosts True, but to us, there's actually only one developer, person X. Then why even bother with remote access via CVS? If that one developer wants to work on files on some other host then they can easily copy them over to it as necessary, and back again when they're done. I'm doing single-user development with CVS at the moment, and using a remote host. (Yes, Greg, I'm using ssh.) Why CVS? For the version and revision control. Why a remote host? Because this way I automatically have two copies, in two different locations, at all times. We do also have a backup system, which backs the repository up to tape, but I'm comfortable knowing that whatever gets cvs committed is sent to a different machine, with a different hard disk, in a different building. Besides, the remote host is the one that already had a repository which was already in our backup script. :) Occasionally the right tool for the job is the one which takes least effort. But speaking of minimal effort: ssh takes, IMO, less effort to set up than correctly setting up pserver, especially if you have a package manager on the server and are running Unix or Linux. Install ssh on both machines, generate keys, set CVS_RSH to ssh, set your CVS connection method to :ext:. Done. Yes, SSH is a little more annoying to set up on Windows. I recommend recent versions of Putty. Do read the docs - you'll need to use the key generator and to set 'pageant' (putty key agent) running. Plink is the command line ssh client. Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Reverting a vendor-branch import
On Sun, 2002-10-27 at 00:13, [EMAIL PROTECTED] wrote: I have performed a vendor-barnch import. I have started the merging procedure, and half way to this I want to revert (undo) the whole import thing. Before doing the import I have tagged my tree with, say, T1. How can I do it? 1. Smart thing, to tag before doing the import. :) 2. I know this is one way to revert to a tagged point, I'm sure there are others. This is taken from info:cvs#Merging_two_revisions cvs update -j (current revision) -j (old revision) filename (yes, this is icky for bulk reversions. You could tag the 'now' point, then do cvs update -j nowtag -j thentag over the whole sandbox.) You can find other solutions at: http://cvsbook.red-bean.com/cvsbook.html#Examining_And_Reverting_Changes Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: using cvs to contol system files
On Thu, 2002-10-24 at 05:21, [EMAIL PROTECTED] wrote: Has anyone out there used to cvs to version control system files? For instance, files in /etc. I'm working with 2 system admins who want to do this. We were going to make the 'live' files a sandbox that they would share. I was hoping to have them edit files by logging in to their regular accounts, su to root, edit a file, exit su, cvs commit filename. The problem is the permissions involved. They each have a umask of 022 on regular accounts. So a 'cvs add dirName' creates a directory in the repository without group write. No problem, we'll set this one by hand. The CVS/Entries file however is left as owned by the last user to commit and without a group write. This is a problem for the next admin to commit. Is there a way around this. Setting the umask in the admins regular accounts to 002 seems like a poor option. Just getting started on this problem. Any advice is appreciated. The way I do it is use cvs with make. Use CVS to store secure versions of the files, then every time you want to modify them, create a sandbox, make your changes, then run 'make'. EG, for a set of mail files for a bunch of machines: cvs -d (repository) checkout mail cd mail vi (host).aliases cvs commit make The make script copies the changed aliases file to the proper host and directory, ensures permissions and ownerships are correct and runs newaliases. (It would probably be 'safer' if it did a cvs export of the relevant file instead of copying the sandbox version, but this is just a home network so some imperfections are ok. (G) ) You could also have a script that runs whenever a change is committed, that does the export/check permissions/run stuff like 'newaliases' or restart services. Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: deleting modules
On Tue, 2002-10-22 at 23:04, Stefan Immel wrote: $ REPO=/usr/local/cvsroot ; # change this to whatever you use $ MODULE=my_module; # change this to your module name $ cd $REPO $ tar zcvf ~/$MODULE.tgz $MODULE ; # make a backup just in case $ rm -rf $MODULE ; # all gone! Thanks, that worked for me. I thought cvs would keep some more infos in the CVSROOT directory. It can, most often in the CVSROOT/modules file, sometimes in the *info files. But if there is information there, it is likely to have been added manually at some point. It can also have information in CVSROOT/history and CVSROOT/val-tags. History would be historical data used by the 'cvs history' command. val-tags is a list of tags used in the module. Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: How do i tag only on branched module
On Mon, 2002-10-21 at 17:29, Ravish A.S wrote: Hi, I am newbie to cvs. I have created two branches from main trunk. Then I tried to labeled the branches separately. it worked fine. Then later i have added new files to each branches. This time I tried to label branches , but it didn't label the newly added files to branches, it did for exist files. Anybody know this problem or am I doing any mistake ?. How do I label the newly added files to branch? .. please help me out. Which command did you use to tag the branches? If you were in a sandbox/working copy that contained the files you wanted to tag, with the revisions you wanted tagged, then used 'cvs tag', it should have worked the way I think you wanted it to. Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: AW: Multiple cvs users in one checkout area
On Mon, 2002-10-21 at 17:05, Andreas Otte wrote: Von: Mr. Cristian Romanescu [mailto:c.romanescu;finsiel.ro] Why doesn't each user checkout his own version of the project and works on it and then commits changes? Why must you have one checkout area? This is the role of cvs so that each user to retrieve a version to work on it The sources are sources for database objects (stored procedures, triggers, ...) which are applied against a database, no user has it's own database, so it makes no sense to work in different areas, everbody will see everybody else's changes in the database anyway. The private sandbox isn't there. When I've done this sort of work (Java client accessing an oracle database), we each had our own sandbox for the source, which accessed shared Java oracle libraries. So we used one database, out of different sandboxes, and compiled against a shared set of database access libraries. Perhaps you could do something like that? You can make symbolic links in your build script, if necessary, or find some other way to use the shared resource when testing, but the individual sandboxes/working copies, for development. Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Binary files and cvswrappers
On Fri, 2002-10-18 at 01:39, [EMAIL PROTECTED] wrote: Having finished - I now realise that there are one or two extensions that I should have included in the cvswrappers (actually the extensions weren't used in the first version of the product but appeared in the last 10 releases). Secondly - how am I meant to deal with files that don't have an extension but could be either a binary or a shell script? This isn't a complete answer, but to set the files that have been missed to default to binary mode, use cvs admin -kb (filename). If a file defaults to binary mode, all CVS operations will treat it as binary henceforth. This doesn't answer your problem with importing, but should help with the existing files. Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: (no subject)
On Tue, 2002-10-15 at 17:33, John Noronha wrote: How can one remove a branch after merging it with the main branch ? There is no specific way that I am aware of to mark a branch 'closed' or 'not to be used anymore'. If you really want to remove the historic record of the branch ever having been there, you can use cvs admin -o and cvs tag -d -B. But I'd back up the repository first! Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Updating third party software
On Tue, 2002-10-15 at 03:14, Kaz Kylheku wrote: The problem with the vendor branch is that it starts at version 1.1 of every file (1.1.1.1 really, but that's an internal detail). You cannot start a vendor branch somewhere halfway up your trunk, so vendor branching is useless for the case when someone takes some release of your code and becomes a patch; you then have to track their patches on a real branch, the methods that you describe. cvs import -b (vendor branch id) allows you to branch off at a later point. (me, I'd like -b to accept a tagname. Maybe I should try to write a patch) Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: cvs diff on repository
On Sat, 2002-10-12 at 03:47, Larry Jones wrote: [EMAIL PROTECTED] writes: cvs rdiff: [10:00:50] waiting for automan's lock in /home/cvs_automation/CVSROOT It will loop there until I Ctrl-C it. Apparently rdiff doesn't like to be run from within loginfo. Any way around it? Use the -n global option to ignore the locks. Alternatively, background the cvs rdiff command in the script. The commit is waiting for the cvs rdiff to finish (because it's called in a *info script), but the rdiff is waiting on the commit's lock. If you background the rdiff, the commit won't be waiting for it anymore. Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Add proxy support to cvs?
On Wed, 2002-10-09 at 17:17, Fabian Cenedese wrote: Hi At the moment I'm using WinCvs because it's the only way to get over our company proxy. It's not possible with native cvs. But I'd like to work with normal cvs so I can also use it on the Linux box. The guys from WinCvs added proxy support so it's not impossible. Wouldn't that be something for cvs also? There have been quite some questions/requests on this list about this subject. If WinCvs could use cvs, why not in return use something from WinCvs? If you can use ssh to get through your company proxy, you can set the CVS_RSH environment variable to ssh (or ssh2), and use the :ext: method to access the server. If you need to use ssh with options or parameters, you can even set CVS_RSH to a shell script that calls ssh with those parameters. What form of proxy is it? What type(s) of proxy(s) is(are) WinCVS supporting? There're a lot out there, without more information it's impossible to actually say 'oh, you can use blah'. Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Tag locking change
On Thu, 2002-10-10 at 04:26, Adam Bregenzer wrote: It has nothing to do with the client, it's all *server* side. I see no reason for it to bve tied to an update, I don't even know how to execute a server-side script on update and wouldn't want to anyways. While there are several better ways to do it and you probably can't run an rtag command from a script run from a *info file it is an example. You can run any cvs command from a script run from a *info file. Just make sure the command is backgrounded so you don't have locking conflicts. Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Tag locking change
On Thu, 2002-10-10 at 05:40, Adam Bregenzer wrote: On Wed, 2002-10-09 at 14:58, Greg A. Woods wrote: My point was that you can do the same thing more reliably with a cvs update in a working directory that is the DocumentRoot. But then it wouldn't be automagic, it would be whenever cvs update is run. Certianly every minute would be 'acceptable' but this solution is fine. Plus in this case the repository and the live site have different directory structures. Try a *info script like: cd /wherever /usr/bin/cvs -d cvs:/home/cvs export -r $1 whosiwhatsit if [ $? -eq 0 ]; then cd whosiwhatsit /usr/bin/make /usr/bin/make install else echo Export failed. fi $1 is obviously a tag. :) You can use update instead of export, and -D now instead of -r $1 if you want. You may also want to replace the echo with a mail. The make and the make install should contain whatever is necessary to do your build and/or rearrange the structure. Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Read only file
On Wed, 2002-10-09 at 02:03, dd hh wrote: Hi everyone, Thanks a lot for the help :-) . But I am not able to understand why read only directory is required in CVS. Why can't we have directories which has both read and write permission and in that directories files with read only permission and some writable files too? Coz in normal unix this works perfectly fine. Could anyone explain it ? It's an artifact of the way CVS writes the RCS files (the files in the repository), and it's part of CVS attempting to make the writing an atomic operation. Rather than editing the actual RCS file, CVS copies the file to a backup, writes the changes to that, then renames the backup to the original file. This means that it doesn't need write access to the RCS file, only to the directory it's in. This ALSO means that read/write permissions in the repository can be modified on a directory-by-directory level, not a file-by-file level. I hope this clarifies things for you. Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: file locking in cvs
On Tue, 2002-10-08 at 22:18, Iqbal Shaikh wrote: Hi all, Would like to know if we need to edit any admin file in CVSROOT to enable locking of files in cvs. Check info cvs, especially 'Multiple Developers' and the subheadings under Multiple Developers called 'Watches' and 'Choosing a model'. Somewhere in there, you'll also find a reference to the 'rcslock' program which is in the contrib directory of the source. I suspect by your question that you're trying to use cvs admin -l and cvs admin -u to create reserved checkouts. Unfortunately, without using the 'rcslock' program, that doesn't work. Those commands use the rcs locking mechanism, which is honoured by RCS but I believe not by CVS (at least, not by default). CVS is designed for the concurrent development of files, and uses a merging model - it's explained in info cvs#Choosing_a_model. The best way I know to simulate exclusive development or file locking is to use cvs watch and perhaps the rcslock program. Read the info cvs sections I've pointed you at, you'll probably be able to take it from there. Ask if you need more help after that. :) Jenn V. -- Do you ever wonder if there's a whole section of geek culture you miss out on by being a geek? - Dancer. [EMAIL PROTECTED] http://anthill.echidna.id.au/~jenn/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
