Re: Strategy of access
I'm not sure how strict you want to implement your policy, I wish :) I could give each user selected rights to each subdirectory in many ways - full access, read only or no access. To do this each directory (entry point to project) must be link up to many groups (with different access rights) not only to one as Unix does. So I consider patching RedHat Linux to use ACL's on repository partition which enable this option. You might want instead to set up a repository administrator group and only give membership in it to trusted users: Let see at this example: user johny has been added to 'project_a' and 'project_b' and has full access (he is in 'cvsadmin' group) but there is no way to add him to 'project_c' with read-only access. If johny is removed from cvsadmin group he will unfortunatelly lose 'write' access to all projects not only 'project_c'. Still searching, Andrew. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: RTFM? [ was Re: cvs [commit aborted]: cannot commit files as 'root']
[ On Monday, July 9, 2001 at 20:22:56 (-0700), David Taylor wrote: ] Subject: RTFM? [ was Re: cvs [commit aborted]: cannot commit files as 'root'] When someone who has long chanted RTFM! as the panacea for countless ills can't find what he wants in the manual (and is thus led to proclaim its absence a failing of the manual), though it exists in the manual...we can all agree that there is a problem here. Oh, I hope I've never really claimed the fine manual was the only panacea for all ills! ;-) I recant if I did! Of course my inability to find something in the manual isn't necessarily due to a failing in the manual either. My self analysis was an attempt to discover why *I* had a problem finding a particular reference. Whether my experiences would match those of anyone else is a very different question, the answer to which is left as an excercise for the reader! ;-) -- Greg A. Woods +1 416 218-0098 VE3TCP [EMAIL PROTECTED] [EMAIL PROTECTED] Planix, Inc. [EMAIL PROTECTED]; Secrets of the Weird [EMAIL PROTECTED] ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: problem with remotehost
krishna rama wrote: We are installed CVS server on Redhat linux server7.0.Remote clients are having windows'98/NT .we want step by step procedure. Have you looked at the manual section on this? http://www.cvshome.org/docs/manual/cvs_2.html#SEC30 we are followed :pserver: method but we geting an error connection refused.we are given 2401 stream tcp /bin/echo echo hello line in the xinetd.conf file. this also given connection refused what happed internally.please help us. Your [x]inetd.conf or /etc/services file is misconfigured. I'm assuming the 'echo' line is a test? It would help if you gave us your original line instead of the echo. As noted in the above link, you have to configure 3 files: /etc/[x]inetd.conf - what was your original line if any? (not the echo) /etc/services - There should be a 'cvspserver' line... $CVSROOT/CVSROOT/passwd - See link. -Matt ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: problem with remotehost
=?iso-8859-1?q?krishna=20rama?= writes: We are installed CVS server on Redhat linux server7.0.Remote clients are having windows'98/NT .we want step by step procedure.we are followed :pserver: method but we geting an error connection refused.we are given 2401 stream tcp /bin/echo echo hello line in the xinetd.conf file. this also given connection refused what happed internally.please help us. Get an experienced RedHat system administrator to set up the server for you. -Larry Jones I don't want to learn this! It's completely irrelevant to my life! -- Calvin ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Branch checkout problem
Karthikeyan writes: When i try to use checkout cvs checkout -r xx1 module1 it says /cvsroot/val-scripts permission denied: No, it doesn't. *PLEASE*, when you report an error, be very careful to report the *exact* text of the message. On the server, you need to set the permissions on $CVSROOT/CVSROOT/val-tags so that everyone can read *and write* it. (In the current development version, this is just a warning rather than a fatal error.) -Larry Jones Do you think God lets you plea bargain? -- Calvin ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: problem with remotehost
From: Matthew Riechers [EMAIL PROTECTED] Heiko Rupp wrote: Just a warning / hint here: I had massive problems with the cvs server on RH 7.0 with errors that souldn't be (cvs update says ignoring ., no such file). Does the local client (linux) work ok? What was the exact command you ran? Client side on RH7 seemed ok. But: the regression tests failed even for non-remote CVS, so I can't guarantee anything. Command was cvs update -PdA With a Server on Suse or Debian or Solaris, all is ok. Heiko ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: How well does CVS handle other types of data?
John Dixon writes: Does anyone have experience using CVS with CAD/CAM binary data files like .MOD (Solution3000) or .DWG (Mechanical Desktop)? Since such file don't allow for concurrent modifications, I can't imagine why you'd want to use the *Concurrent* Versions System to store them. -Larry Jones Some people just don't have inquisitive minds. -- Calvin ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Zip the repository?
Voigt, Ulrich writes: Does anybody know how to use a zipped repository? Unzip it or rewrite CVS. -Larry Jones Just when I thought this junk was beginning to make sense. -- Calvin ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: RTFM? [ was Re: cvs [commit aborted]: cannot commit files as 'root']
[EMAIL PROTECTED] (Greg A. Woods) wrote: [ On Monday, July 9, 2001 at 20:22:56 (-0700), David Taylor wrote: ] Subject: RTFM? [ was Re: cvs [commit aborted]: cannot commit files as 'root'] When someone who has long chanted RTFM! as the panacea for countless ills can't find what he wants in the manual (and is thus led to proclaim its absence a failing of the manual), though it exists in the manual...we can all agree that there is a problem here. Oh, I hope I've never really claimed the fine manual was the only panacea for all ills! ;-) I recant if I did! Of course my inability to find something in the manual isn't necessarily due to a failing in the manual either. My self analysis was an attempt to discover why *I* had a problem finding a particular reference. Whether my experiences would match those of anyone else is a very different question, the answer to which is left as an excercise for the reader! ;-) Well, you're not alone having difficulties with the CVS manual. Either a good number of CVS users like me are dumb or indeed there's a problem with the manual. -- __Pascal_Bourguignon__ (o_ Software patents are endangering () ASCII ribbon against html email //\ the computer industry all around /\ and Microsoft attachments. V_/ the world http://lpf.ai.mit.edu/ 1962:DO20I=1.100 2001:my($f)=`fortune`; http://petition.eurolinux.org/ -BEGIN GEEK CODE BLOCK- Version: 3.1 GCS/IT d? s++:++(+++)++ a C+++ UB+++L$S+X$ P- L+++ E++ W++ N++ o-- K- w-- O- M++$ V PS+E++ Y++ PGP++ t+ 5? X+ R !tv b++(+) DI+++ D++ G++ e+++ h+(++) r? y---? UF --END GEEK CODE BLOCK-- ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: CVS For VMS
Martin Eismann [EMAIL PROTECTED] on 07/09/2001 10:01:10 PM To: [EMAIL PROTECTED] cc: (bcc: Rex Jolliff/YM/RWDOE) Subject: Re: CVS For VMS Federal Record Status Not Determined Hello dear list-members! I am very interested getting an OpenVMS-Alpha-binary (maybe client and server? would be great!!). Upper mentioned wildcard-support would be excellent... Has anybody compiled/linked a OpenVMS-Client from actual sources? I have linked CVS v1.11 on an Alpha (DEC C V5.7-006 on OpenVMS Alpha V7.2-1). Only the client will build. I remember sending this version to the CVS maintainers for distribution. If I get some time I'll build 1.11.1p1 and send in to them also. Rex. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: How well does CVS handle other types of data?
John Dixon writes: Does anyone have experience using CVS with CAD/CAM binary data files like .MOD (Solution3000) or .DWG (Mechanical Desktop)? Since such file don't allow for concurrent modifications, I can't imagine why you'd want to use the *Concurrent* Versions System to store them. -Larry Jones Some people just don't have inquisitive minds. -- Calvin I can't speak for John, but we use CVS because most of our files are plain text source files. However, we have some files, both documentation and UML graphical models, which aren't able to be merged via CVS. We want to use one version control system, so we're putting up with CVS's limitations in this area. To answer your question simply; you can use CVS for binary files. Just use -kb when you import or add the binary files. The downside of this is that even though the file has been marked as binary and thus no changes can be merged, CVS will allow multiple users to get writeable copies of the file using cvs edit. When the second user tries to checkin, they'll find that they can't and thus they'll have to redo their work after getting an updated copy of the file. There are ways to send email notification when the second user gets a writeable copy, (see cvs watchers), which can help as long as all users have good notification of new email. The solution usually touted is to make sure your developers know who's working on which files which seems like a cop-out to me. There's also a patch for cvs on SourceForge that would make cvs edit fail if the file is locked, but I haven't seen any indication that it will ever show up in the mainline CVS. That's a problem for us as we have multiple CVS executables on Solaris, Linux, Windows cmd line, Windows DLL, WinCVS, and CygWin. We may decide to go with our own customized version of CVS at some point to solve this and to add hooks to avoid the DOS/Unix line ending conversion. Does anyone know how to customize the version of CVS built in to WinCVS? Thanks! - Bruce Hill [EMAIL PROTECTED] ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: How well does CVS handle other types of data?
On Tue, Jul 10, 2001 at 08:56:09AM -0400, Larry Jones wrote: Since such file don't allow for concurrent modifications, I can't imagine why you'd want to use the *Concurrent* Versions System to store them. It's free and has decent networking capabilities? mrc -- Mike Castle [EMAIL PROTECTED] www.netcom.com/~dalgoda/ We are all of us living in the shadow of Manhattan. -- Watchmen fatal (You are in a maze of twisty compiler features, all different); -- gcc ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Branch merging into main line,
All, I have a dilemma. I am fairly new to CVS. One of my developers needs to merge a branch into main line. There are a lot of files that need to be manually merged. My take on this is that he needs to edit the main line file by adding to it the extra stuff contained in the branch file. Then just check in the main file into CVS just like any other check in. Am I right? I would really appreciate any help I can get with this issue. Thanks David Petruescu This e-mail and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this e-mail in error please notify the system administrator - [EMAIL PROTECTED] Any views or opinions expressed in this e-mail are those of the sender and do not necessarily coincide with those of Wireless, Inc. This footnote confirms that this e-mail message has been scanned by a virus protection program for the presence of a computer virus. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
checking out modules into an existing directory
Hi, I am having a problem checkout out modules into an existing directory. If I try to do this I get the following error: cvs checkout: in directory c:\temp\source: cvs checkout: cannot open CVS/Entries for reading: No such file or directory If I check the module out when there is no existing directory and it creates a new one, everything works fine. The problem is I need to be able to check it out into an existing directory. Anyone know a way to fix this? Thanks, Jason A. __ Do You Yahoo!? Get personalized email addresses from Yahoo! Mail http://personal.mail.yahoo.com/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: checking out modules into an existing directory
Jason Allen wrote: Hi, I am having a problem checkout out modules into an existing directory. If I try to do this I get the following error: cvs checkout: in directory c:\temp\source: cvs checkout: cannot open CVS/Entries for reading: No such file or directory If I check the module out when there is no existing directory and it creates a new one, everything works fine. The problem is I need to be able to check it out into an existing directory. Is that directory already controlled by CVS? Anyone know a way to fix this? Thanks, Jason A. What command did you run? Are you trying to do 'cvs co -d source module'? That doesn't mean 'create directory source\module'. It means 'create a sandbox dir called source for module'. It won't work if CVS doesn't know about that directory. 'cvs co' on an existing sandbox effectively does a 'cvs update' HTH, -Matt ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: How well does CVS handle other types of data?
-Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] John Dixon writes: Does anyone have experience using CVS with CAD/CAM binary data files like .MOD (Solution3000) or .DWG (Mechanical Desktop)? Since such file don't allow for concurrent modifications, I can't imagine why you'd want to use the *Concurrent* Versions System to store them. Here at work, we have binary files scattered through the text files. Since CVS is one of the better available systems for source code, and using two systems is a significant cost, and CVS handles binary files in an adequate manner, that's what we do. I have a repository at home, which serves one developer (me) and therefore doesn't involve concurrent development. I use it partly because I'm familiar with it, but also because it works well over the home network (Linux and Macs) and serves my needs very well. I can have branches for possible development work. I can tag release points and cut branches, so I can fix bugs without interfering with new features. I can easily create diffs between versions. If I screw up maintaining the change log, it's easy to recover from. It's overkill, but what that means is that it satisfies all my current needs and I have room to go from there. There's also the fact that I'm doing my home stuff on as low a budget as I can get away with, and CVS fits into that budget very nicely. It seems to me that, to forego the possibility of concurrent development (which is irrelevant to me) I'd have to forego things that I care about. Can I get a Linux/Mac source code system that does branching for free other than CVS? -Larry Jones Some people just don't have inquisitive minds. -- Calvin Sometimes these random .sigs seem appropriate. :-) ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: [PHP-DOC] cvs: phpdoc /en/features safe-mode.xml
Don't know why that would have anything to do with the faq, but you do need lt; instead of in your code examples... Daniel - Original Message - From: Jeroen van Wolffelaar [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Tuesday, July 10, 2001 3:36 PM Subject: [PHP-DOC] cvs: phpdoc /en/features safe-mode.xml jeroen Tue Jul 10 16:36:23 2001 EDT Modified files: /phpdoc/en/features safe-mode.xml Log: It did build this way... strange... - fixed ?php - lt;?php Index: phpdoc/en/features/safe-mode.xml diff -u phpdoc/en/features/safe-mode.xml:1.1 phpdoc/en/features/safe-mode.xml:1.2 --- phpdoc/en/features/safe-mode.xml:1.1 Tue Jul 10 16:33:01 2001 +++ phpdoc/en/features/safe-mode.xml Tue Jul 10 16:36:22 2001 @@ -28,9 +28,9 @@ /computeroutput Running this script.php programlisting role=php -?php +lt;?php readfile('/etc/passwd'); -? +?gt; /programlisting results in this error when safe mode is enabled: computeroutput ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
CVS on AIX 4.2
Hi, I noticed that there is a compiled version of CVS for AIX 4.3 available for downloadand wanted to see if anyone would know if I could expect to run that same versionunder AIX 4.2 or would you expect this to need to be recompiled to work? If not maybe there exists an already compiled version for AIX 4.2? Background: Doing development with CVS on a handful of machinesand can usea W2KNTorAIX RS6000 box to run as a CVS repository server but don't want to trust the NT version based on recommendations we'vehad to avoid it. Please reply to my email as well sinceI am not yet signed up for the mailing list but will soon. Thanks, Guy GardnerChief RD ScientistDynamic Healthcare Technologies51 Sawyer Rd.Waltham, MA 02453781.642.6200x3246http://www.dht.com
Re: CVS on AIX 4.2
Guy Gardner writes: I noticed that there is a compiled version of CVS for AIX 4.3 available for download and wanted to see if anyone would know if I could expect to run that same version under AIX 4.2 or would you expect this to need to be recompiled to work? It almost certainly will need to be recompiled. Vendors work very hard to make their systems upward compatible, but downward compatibility is a very hard problem and almost never done. As long as you have a C compiler, you should be able to download and build the source with no trouble at all. -Larry Jones No one can prove I did that!! -- Calvin ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: [PHP-DOC] cvs: phpdoc /en/features safe-mode.xml
Um.. I'm terribly sorry about this post. I somehow typed The wrong address in the To: field. Daniel Beckham - Original Message - From: Daniel Beckham [EMAIL PROTECTED] To: info-cvs [EMAIL PROTECTED] Sent: Tuesday, July 10, 2001 4:12 PM Subject: Re: [PHP-DOC] cvs: phpdoc /en/features safe-mode.xml Don't know why that would have anything to do with the faq, but you do need lt; instead of in your code examples... Daniel - Original Message - From: Jeroen van Wolffelaar [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Tuesday, July 10, 2001 3:36 PM Subject: [PHP-DOC] cvs: phpdoc /en/features safe-mode.xml jeroen Tue Jul 10 16:36:23 2001 EDT Modified files: /phpdoc/en/features safe-mode.xml Log: It did build this way... strange... - fixed ?php - lt;?php Index: phpdoc/en/features/safe-mode.xml diff -u phpdoc/en/features/safe-mode.xml:1.1 phpdoc/en/features/safe-mode.xml:1.2 --- phpdoc/en/features/safe-mode.xml:1.1 Tue Jul 10 16:33:01 2001 +++ phpdoc/en/features/safe-mode.xml Tue Jul 10 16:36:22 2001 @@ -28,9 +28,9 @@ /computeroutput Running this script.php programlisting role=php -?php +lt;?php readfile('/etc/passwd'); -? +?gt; /programlisting results in this error when safe mode is enabled: computeroutput ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: checking out modules into an existing directory
The directory I'm checking it out into, isn't controlled by CVS (no CVS folders). I am using cvs co -d dest module, I have also tried to just name the module on the server the same as the existing directory and then check it out into one directory higher. I get the same problem. Thanks for your help, Jason A. - Original Message - From: Matthew Riechers [EMAIL PROTECTED] To: Jason Allen [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Sent: Tuesday, July 10, 2001 2:47 PM Subject: Re: checking out modules into an existing directory Jason Allen wrote: Hi, I am having a problem checkout out modules into an existing directory. If I try to do this I get the following error: cvs checkout: in directory c:\temp\source: cvs checkout: cannot open CVS/Entries for reading: No such file or directory If I check the module out when there is no existing directory and it creates a new one, everything works fine. The problem is I need to be able to check it out into an existing directory. Is that directory already controlled by CVS? Anyone know a way to fix this? Thanks, Jason A. What command did you run? Are you trying to do 'cvs co -d source module'? That doesn't mean 'create directory source\module'. It means 'create a sandbox dir called source for module'. It won't work if CVS doesn't know about that directory. 'cvs co' on an existing sandbox effectively does a 'cvs update' HTH, -Matt ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs _ Do You Yahoo!? Get your free @yahoo.com address at http://mail.yahoo.com ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Alpha VMS check checkout problems in 1.11.1
Mike, I see what you mean by the problem when checking out a large number of files with the message of cvs [export aborted]: cannot rename file. I have the same problem after building 1.11.1p1 under OpenVMS 7.2-1. There is the work around of repeating the command until everything is checked out but that takes extra time. In testing this I did a checkout of a path that contains about 2,800 files. There is definitely a pattern when it fails. After it fails I used a DIR [.path]*.*;/grand to gather some stats. Keep in mind that the counts include all the [.CVS] directories and their contents. Directories Files Increase 18 318-- this may not have been the first set 18 555 237 46 880 325 60 1166 285 60 1403 237 60 1639 236 66 1893 254 70 2147 254 88 2442 295 102 2725 283 102 2964 239 128 3232 268 When the number of directories did not increase about 236 or 237 were processed. I am wondering if there might be some limit on the number of files that is causing checkout to stop. I just tested the cvs export and got the following: cvs [export aborted]: cannot rename file ./fc05_hgi_get_dist.for_new_ to ./fc05_hgi_get_dist.for: no such device or address DIPSD-dir [.alpha...]*.*;* /grand Grand total of 3 directories, 238 files. It also stopped. It also fits the pattern! Have you also seen this on the import and checkin commands? Was this the same with prior versions? Thanks, Dale ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
New version of CVS access control patch available
I have ported the access control patch to the newest verson of CVS, fixed some bugs, and added some new features. It's available on my web page at http://members.home.net/minyard. The access control path adds per-directory access control lists to CVS (pserver) and the ability to do full user management (password changing, adding/deleting users) from the CVS command line. It's certainly not perfect, but it does a pretty good job. If someone (maybe me) adds an SSL implmentation based on this, it would be pretty secure. This release adds: * Lots of bug fixes and cleanups. The new CVS directory/repository stuff was cleaned up and made much better, this cleaned up some of my patch. * The ability to specify a prefix for the repository that the server will tack on to the beginning of the root directory. So if your repository is in /home/cvs/root, you can set the prefix to /home/cvs and the clients will only need to specify :pserver:a@b:/root. Plus, you can move the repository on the server's filesystem without affecting the users. * For the access permissions, you can optionally allow them to propigate to subdirectories. I don't personally like this very much, but a lot of other people do, so I added it. If you use this patch, an upgrade is highly recommended. -Corey ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: How well does CVS handle other types of data?
[ On Tuesday, July 10, 2001 at 11:28:17 (-0700), Mike Castle wrote: ] Subject: Re: How well does CVS handle other types of data? On Tue, Jul 10, 2001 at 08:56:09AM -0400, Larry Jones wrote: Since such file don't allow for concurrent modifications, I can't imagine why you'd want to use the *Concurrent* Versions System to store them. It's free and has decent networking capabilities? So, do you use a hammer on a machine screw simply because it's in your hand and it's got a good weight and balance to it?!?!? What nonsense!!! PLEASE TRY TO USE THE RIGHT TOOL FOR THE JOB! -- Greg A. Woods +1 416 218-0098 VE3TCP [EMAIL PROTECTED] [EMAIL PROTECTED] Planix, Inc. [EMAIL PROTECTED]; Secrets of the Weird [EMAIL PROTECTED] ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: RTFM? [ was Re: cvs [commit aborted]: cannot commit files as 'root']
[ On Tuesday, July 10, 2001 at 16:40:04 (+0200), Pascal Bourguignon wrote: ] Subject: Re: RTFM? [ was Re: cvs [commit aborted]: cannot commit files as 'root'] Well, you're not alone having difficulties with the CVS manual. Either a good number of CVS users like me are dumb or indeed there's a problem with the manual. I think part of the problem is with this word read. It's been quite a long time since I sat down and actually read the manual. I tend to use the 's' key in info a lot, and that can result in worse side-tracking and unapplicable references than typing sex into Google. I think that if people (including even me!) actually _read_ the manual then there would be far less complaint. Texinfo manuals are, after all, generally written in such a way as to lend themselves far more to being read than to being used as a reference, and the CVS manual is no different in this respect. I still use (not read, exactly) the old troff manual sometimes simply because it's a better reference despite being effectively deprecated and out-of-date. -- Greg A. Woods +1 416 218-0098 VE3TCP [EMAIL PROTECTED] [EMAIL PROTECTED] Planix, Inc. [EMAIL PROTECTED]; Secrets of the Weird [EMAIL PROTECTED] ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: How well does CVS handle other types of data?
[ On Tuesday, July 10, 2001 at 11:15:23 (-0700), Bruce Hill wrote: ] Subject: Re: How well does CVS handle other types of data? The downside of this is that even though the file has been marked as binary and thus no changes can be merged, CVS will allow multiple users to get writeable copies of the file using cvs edit. When the second user tries to checkin, they'll find that they can't and thus they'll have to redo their work after getting an updated copy of the file. If you even use cvs edit -- not everyone does, or needs to even. You also can't easily merge branches if any changes have been made to binary files on each branch -- Greg A. Woods +1 416 218-0098 VE3TCP [EMAIL PROTECTED] [EMAIL PROTECTED] Planix, Inc. [EMAIL PROTECTED]; Secrets of the Weird [EMAIL PROTECTED] ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Strategy of access
[ On Tuesday, July 10, 2001 at 10:16:16 (+0200), [EMAIL PROTECTED] wrote: ] Subject: Re: Strategy of access Let see at this example: user johny has been added to 'project_a' and 'project_b' and has full access (he is in 'cvsadmin' group) but there is no way to add him to 'project_c' with read-only access. Well, with basic UNIX permissions if you've set up CVS so that some projects are read-only to some users then implicitly all users with access to the CVS repository will have read-only access to all projects. However real issue is that if you give a user write access to the CVSROOT module then you cannot literally prevent them from having read or write access to anything and everything (though total lack of accountabilty requires a much more sophisticated attack). Thus membership in the cvsadmin group must only be given to highly trusted users. Your policy should not require that members of the cvsadmin group ever be restricted to read-only access of any given module -- they should be explicitly trusted with write access to all of $CVSROOT. With very very very careful use of per-file (er, per-directory) ACLs you can provide much finer grain control over read-only access, but of course you'll have to implement the server on a system which supports such ACLs, and of course you'll really have to totally avoid cvspserver (though you must avoid cvspserver anyway if you're this concerned about implementing solid technical access controls!). If johny is removed from cvsadmin group he will unfortunatelly lose 'write' access to all projects not only 'project_c'. No, that's not true at all. I don't understand the problem you seem to see here. If you've set up the CVSROOT module access properly then users do not need to be members of the cvsadmin group just to have write access to one or more projects. The CVSROOT module is just a separate project from that point of view. Yes there is the CVSROOT/history file, and optionally some files that might be written by programs called from specs in CVSROOT/*info files. However you don't need write access to the $CVSROOT/CVSROOT directory just to be able to write to some file within that directory. Depending on your specific circumstances you can either make the history file world-writable or create some cvsusers group which all CVS users must be a member of (and which would group-own the group-writable history file). -- Greg A. Woods +1 416 218-0098 VE3TCP [EMAIL PROTECTED] [EMAIL PROTECTED] Planix, Inc. [EMAIL PROTECTED]; Secrets of the Weird [EMAIL PROTECTED] ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs