[cvsnt] test
___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] CVSNT on OpenSolaris
Tilo Riemer wrote: The right way is adding a check for getopt to the autoconf stuff... There's already a getopt compatibility test in configure.in, which is designed for OSX which has a broken getopt. It shouldn't be hard to modify that test to detect whatever is wrong with the solaris one. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] reg: login error
Shivendra_Vishal wrote: So probably once login is needed when you start your session for getting connected to server and thereafter its not required. Any comments? That's not what login is for. It merely stores your password in the registry, and is only ever needed once unless your password changes. Login with sspi really isn't needed unless there's something special about your setup - in fact it's not recommended. Your error sounds more like errors caused by firewalls/AV breaking the connection. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] SCC plugin for CVSNT?
Bo Berglund wrote: Now for the first time in all these years I have found a need for a plug-in for CVSNT to a development IDE... CVS Suite has one built in that's been used with various IDEs and should work with any SCC compatible system. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] reg: login error
Shivendra_Vishal wrote: Dear All, I'm setting CVSROOT and trying to login to cvs as follows, however CVS is not allowing me to login and gives an error cvs [login aborted]: The :local: protocol does not support the login command. C:\Documents and Settings\Administratorset cvsroot=:sspi:cscsbilifecvs\administra...@cscsbilifecvs:/scm_test SSPI almost never needs login (there are some special cases but most users won't see these), and doesn't need the username if you're not logging in as a different user. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] txvkrSxUqU
xqjkunmyoa wrote: cWhCMu a href=http://eegltisveufv.com/;eegltisveufv/a, [url=http://wbwwmjpbnoko.com/]wbwwmjpbnoko[/url], [link=http://cjkpnjigyztl.com/]cjkpnjigyztl[/link], http://iklirrfmmdwj.com/ Grr... no way to filter this - it's sailing past all my spam filters, explicit rules to block it, everything. I can't even find out how it's getting on the system... they've found a way of bypassing the normal logging of source IP address. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Betr.: CVSNT to CVS
Andreas Krey wrote: Most software is written and then a bit modified, and then stays, which means repositories don't grow over time, but mainly with the code size they contain. Which explains why git often manages to have the repository smaller than the (uncompressed) tar ball of the head. A valid and not uncommon model is to keep the resultant binaries in the tree for each released version. Some even go as far as to keep the installers for all the compiler versions in there as well, so they can guarantee they can replicate a build. Binaries don't compress very well... by the time you had half a dozen in there you'd be talking about some serious repository size. Worst case is those places that put ISO images of their releases in the reopsitory as well! Which basically means that you need to enable compression if a cvs checkout of the head shall be faster than a git clone of the *entire* repo. cvs checkouts are generally always compressed.. but the last time I grabbed redhat it took 2 hours on a fast link. I only wanted a couple of files... I happen to travel by rail, and while there is GPRS etc., it is a bit flaky when you actually move, and it's not exactly fast. You definitely don't want to do lots of 'cvs diff's there. Depends on how you work - I rarely do cvs diffs for example, unless I'm checking a commit, and if I'm doing that it's in the office. Those people would also be pretty indifferent to switching tools; give them a list of old and new commands, and that's it. As Arthur points out, that's not true - those kinds of people are often the most resistant to change. Unfortunately, those (the mergemeisters and generally those who do the difficult stuff) need to be in the boat. If you omit these 10% you got an acceptance problem at least where such people exist. There's stuff that people never do, except in rare exceptions. You'd be surprised - we've had stuff break in cvsnt and it been 18 months before anyone actually noticed. People think 2.5.03.2382 is pretty stable (and it is, generally) but there are several hundred bug fixes since, some of them major.. except they don't use that bit so don't notice. and lots of general (command line) usability stuff. Just how often you have to type repository urls to svn is annoying. (Although the price for the biggest shittyness*popularity product goes to ant.) Never saw the point in ant. I note a trend there. :-) Namely the reliance on the commercial business case (which is nothing to boo at). For evs, definately. It's a big server aimed at commercial customers, so they drive the development. Everything in there so far (AFAIK) is stuff that companies have asked Arthur about. Conflicting commit resolution? Can be handled by the admin when they need it. In practice it doesn't happen - if you have a team in the US and a team in India, for example, they'll be working on different modules. They may call into each other so they need each others source code, but they won't be modifying each other. Where teams are working on the same modules we've got that covered. 'normal' as in 'centralized VCS', of course. :-) The problem is the administrative overhead in that, the distributed VCS are a lot more democratic (or anarchic) there. Not so much.. create a branch, setup access control so only that user can write to it. 5 minutes max. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] CVS diffs and poor links (was RE: Betr.: CVSNT to CVS)
Peter Crowther wrote: But... it depends on how you work, as you say. Tools exist to support me in doing my job, and what have I changed? is an operation I wish to perform more and more frequently as I get older and more forgetful. At the same time, I travel more, and that fast, always-on connection is neither fast nor always-on. For this reason, I appreciate having a local copy against which I can diff, even when I'm on a train on a particularly rural part of the West Coast main line*. That's a specific case that's quite fixable - it was something I was planning to work on if I get the chance.. you just need to be able have a copy of the original and have cvs diff smart enough to handle that case. Most of the code to do it is already in cvsnt, in fact. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Betr.: CVSNT to CVS
Andreas Krey wrote: On Thu, 09 Apr 2009 10:00:59 +, Tony Hoyle wrote: ... A valid and not uncommon model is to keep the resultant binaries in the tree for each released version. Someone misunderstood '*source* code control system'. And, since by definition those thingies (released binaries or installers) never change there isn't really a point to put them under version control. It's a revision control system, which is integral part of the wider topic of configuration management. Putting released binaries under revision control is a perfectly valid part of that. Even cvsnt has a module for precompiled external binaries - it makes no sense to put them anywhere else... the repository is the single source of everything required to build a new release, as it should be. If you need to be able to reconstruct the *exact* build environment for an old revision, to fix a bug for a customer, then having the dependent libraries in there and possibly even the compilers is a normal thing to do. Not all environments need this but for those that do need it you have to be able to support it. Let alone in the same repository. Also projects that do this are usually not those you can easily contribute to, if at all. If you're employed to work on a project how easy it is doesn't enter into it. noticing. It's not just the work but also that you don't immediately appear in the global branch name space. Which means it isn't logged/audited, or backed up on the central server. That is not a good thing. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Betr.: CVSNT to CVS
Andreas Krey wrote: It isn't until you decide the stuff is worth of putting into the central. Just like with regular commits. You could get lazy or clandestine, though. That's not how auditing works.. you need to audit everything. In some industries that's a legal requirement. That means right down to the diffs. You don't get to decide what stuff is worth storing. Again, not everyone needs this, but for those that do it's an absolute showstopper if it can be bypassed. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Betr.: CVSNT to CVS
Peter Crowther wrote: From: Tony Hoyle Crucially EVS is *not* a new version control system. It will be whatever you want it to be. It can be CVSNT, or SVN, or TFS or (potentially) Git. Or all of them at once, if you want. Tony, how are you dealing with the semantics and assumptions behind each of those systems? Is there a superset of each system's functionality that allows you to provide 100% compatibility, or are there incompatible assumptions between two or more of them that mean you'll have to choose at some point or elect to have incomplete support for one or more? It pretty much depends on what you're doing with it. 90% of work tends to be some variation of checkout/update/commit, and that'll always work or we wouldn't claim support in the first place. The aim is to get as much of the remaining as possible.. I haven't come across anything fundamentally incompatible in any system so far - the constraints have been all been time/budget based. If you've got a mixture of clients then you might have to agree on how to work together (eg. CVSNT treats the top level objects as branches and tags. The SVN module could (and probably should) split these into subfolders called branches and tags, as is the common SVN convention, but doesn't right now). OTOH I'd assume there would be conventions in place already. Beyond that kind of stuff.. it *should* work the way you want it to - and we'll try to make sure that it does for all our customers. That doesn't mean 100% compatibility necessarily - we implement the protocol, and sometimes what comes back is a little different from a native server - but if there turns out to be something you can't do that you should be able to, that would be an issue we'd need to look at. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Betr.: CVSNT to CVS
Peter Crowther wrote: From: Tony Hoyle Doesn't matter to us - since all clients are equal if someone wants to do something not supported on one clent they just use a different one. *chuckle* I hope your product manager has flagged that one for review. May I suggest reading up on the curse of knowledge and re-evaluating it? I stand by it. This isn't limiting their choice, just expanding it. A cvsnt client can't copy a subdirectory, and svn client can. The cvsnt user has two choices (a) do it the old way, copy/add the files, or (b) walk over to the nearest svn user, ask them to copy it, and do an update on his own workstation. People get stuck in their ways (this is part of the reason that evs is so nice, as multiple dev groups don't need to have an argument over whether to use cvsnt or svn or whatever, as happens currently). So a hardcore cvsnt user would choose (a) every time. If this is a large tree they're copying it'll take them a while, but they can do it. Someone with a bit of flexibility would choose (b). They might even choose to switch their sandbox to something else permanently - which they can do without affecting their teammates or the project as a whole (except the time taken to switch, which of course might by itself preclude such an action). Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Betr.: CVSNT to CVS
Andreas Krey wrote: There isn't enough coffee in the universe to do that. To make that work you'd need a way to translate whatever your storage is into a git repository, and to be able to accept it back (in another repository). That and the way svn merge tracking works (or rather, does not), can simply not be made to work for the general case. Not at all. It merely needs to be pretend to be git. It's probably not even that hard, provided you can export something that looks like a git repository, and accept that back. Looking at the techincal docs I can't see anyting particularly challenging there (in fact as the client does all the work, it's actually a lot easier than some systems). There doesn't appear to be any docmentation of the native protocol, but that's not unusual these days. It may be possible to have one EVS-with-former-svn-repo, and one EVS-with-former-git-repo, but I thing a generic bridge is impossible. If it weren't, the VCS were actually semantically the same, and they very definitely aren't. evs is *already* a generic bridge. It works. There are rough edges, mostly due to time constraints, but the technical barrier has been sorted. And yes the are all the same, at heart. That's the depressing thing about it.. they all go off writing new ones instead of improving what's there already and end up with something that is basically the same thing with a new skin on it. Sure, you can talk about client/server versus distributed (which is really server/server), or DAV vs something else, or having repository vs file versions.. but those aren't fundamental differences when you break it down to what you're doing. The reason for that should be obvious - you're storing end user actions, and those actions don't change. A rename will always be a rename, a file modification will always be a file modification. Are there compromises in evs? Yes.. tagging with the cvsnt client doesn't *quite* have the same semantics for example. SVN doesn't support unicode, or renames (their rename command is just a copy/delete). TFS is, well, TFS... Doesn't matter to us - since all clients are equal if someone wants to do something not supported on one clent they just use a different one. Alone the way how we work(ed) with cvsnt *cannot* be represented in git, because there merges always take the whole repository. No partial directory or file merges -- the latter also makes migration to svn hard. And you *need* to represent it since that is how the git 'client' talks. That's an implementation detail.. I heard much the same when we talked about supporting SVN and CVSNT simutaneously.. they were supposed to be 'too different'. There's no fundamental difference between a repository merge and a directory/file merge in terms of what actually happens. The only difference is the unit of change you're dealing with. evs has more fine grained object tracking than any system currently out there (that I've seen) and at the same time can represent changes at the repository level if required. It was designed from the ground up to do that. Migration to SVN is hard because it's SVN - The SVN server part of evs was 3 months of trying to work out why the hell they did it like that (and I'm still not sure I answered it). Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Betr.: CVSNT to CVS
Andreas Krey wrote: Waitaminute. You need to convert the full history that EVS holds into a git repo, otherwise the git (client) users will complain because a) the history isn't there and b) if you hand out partialized histories the commits won't have their usual identities. If the git users aren't able to pull from one another, why should they bother at all? No problem with enabling that... Personally I find the concept of transmitting whole repositories between users silly, but if customers wanted it there's no reason why not. Also, as a git user I would expect to be able to pull from one EVS repo and then push one branch into another EVS repo; and would Take a scenario. Your company has an evs server. They let you use any client you like, but there is of course only one centrally, fully backed up and audited central server. If they want another evs server then evs servers can sync between themselves just fine. There's no purpose in trying to use git to do that - in fact it'd be missing important metadata, so would be a very suboptimal solution. It also highlights the fundamental problem with git - someone can take your entire repository history and copy it out of the company, and you can't do a damned thing about it. Corporate customers are big on auditing and security - neither of which git can provide, by design. It also lacks things like locking, again because of its distributed nature. Evs can help as at least the central server has these functions, but it's not entirely a solution. It's a good solution for opensource development, but then evs is just not aimed at that. Just branch out (in cvsnt) one subdirectory, start working in that branch, and merge back a few (not all) files. The history you now have in cvsnt cannot be properly represented in git (and possibly neither in svn). You can represent the state of the files in all commits, but what is missing is the correct merge tracking information. What you describe there is basic version control. I don't see the problem. It's simple enough in SVN (although SVN's mergepoint stuff is pretty bizarre). evs records the merges for all such commits. If the client can't provide them, then.. it couldn't whether the server is evs or anything else, so nothing is lost, it's just a consequence of the choice of client. ... evs is *already* a generic bridge. It works. Between cvsnt and svn? Yes. How do you map the trunk/tags/branches metaphor? How do you import from svn when the used style doesn't match your way? There's a generic mapping. Is said elsewhere that it should probably create virtual directories called tags and branches, but doesn't right now. Import from svn is easy enough. You might have to move some things around to support the cvsnt clients, but it's not a major hardship. There are rough edges, mostly due to time constraints, It is my argument that the time constraints can't be lifted if full interoperatility between cvsnt, svn and git clients is to be achieved, especially on already-existing repositories that did some strange things. The time constraints are mostly because we have limited budget and one active developer. The hard work is done, and if we had someone else I'd probably put them on ironing out the remaining rough edges on SVN first. I see no reason why that particular module shouldn't be near perfect. The same with TFS (subject to implementation of the bug tracking modules for it etc., which is a longer project). That covers the major ones. git doesn't store renames either; they conceptually just the store the current state of the tree after each commit. You can determine things using heuristics, or just treat it as a bunch of adds and removes. The latter approach means you've lost some object history, which is unfortunate but unavoidable if the client doesn't support it. Also, svn:externals aren't anything elsewhere, as are git submodules or cvs modules. There are some possible mappings between each other, but none of them are the only possible way of doing things, and doing the mappings at least requires some project-specific configuration. Those are implementation specific details. They don't actually change what's stored. At some point evs may well sprout its own generic module system that maps across all clients. You can basically do it already with symbolic links (although no clients currently create them except for cvsnt alias tags). Pfft. svn and cvs are almost the same, except for the stupid tags-are-in-file-namespace idea. They both give you the advantage that you can basically hide the actual history from the clients. To interoperate with git you need to *give* them the full history. Don't see the problem with that, provided your IT manager doesn't have a problem with it.. That's the problem: git was designed not to do this. That's not a problem. evs was. Did you design the storage model in a way that
Re: [cvsnt] Betr.: CVSNT to CVS
Andreas Krey wrote: ...and get the flames when the next merge fails because of the mergeinfo that wasn't created because he carried over the files by hand. Depends if the directories are ever intended to be mergable. If they're not, then both options are equivalent. On the other hand 'svn copy' basically only exists to create tags and branches anyway. Who ever needs two directories that are mergeable to each other within a project tree? It happens. eg. customer specific modules, one per directory. You'd copy those to keep them mergable and make the customer specific changes - if the base module changes you merge from one to the other. svn suports this. cvsnt could check that out but couldn't create the copies, or do the merges. IIRC tfs could do it as well. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Betr.: CVSNT to CVS
Andreas Krey wrote: It comes close to the best thing since version control in general, IMHO. That way I can basically do everything offline, even complicated merges, and only need a connection for pushing back the results. I first considered it overkill just as well but then the whole git repository is typically smaller than all the 'pristine copy' files that svn puts in the sandbox. Only for a new repository. Try pulling down something like redhat, for example. If you're working on the same data all the time then there's saving - for an initial download of data you're avoiding a lot of the small ones later (although of course you still need to update from the server to get others changes, and commit to the server to send your changes, so it's debatable how much you're actually saving). If you're only on a project for a short period it's a huge waste especially if it's happening remotely. With git you would just as well have an 'official' repo, but the funny thing is that every developer has a full backup. :-) What's more important to me is that I can continue working even when the central server is offline (like 'powered down for holidays'). Something we considered adding to cvsnt is offline commits, so you could commit locally then sync later. It's not really a common case though - it's rare you don't have *some* internet connectivity. That's the point I was tring to make: EVS can't come to a point where it mimics a git server so that you can treat it like one in all regards. Under normal use you wouldn't be able to tell the difference. You're talking about corner cases - the normal use case is everyone checks out from the central server, works with stuff and checks back in. They're not about to start sending entire repositories to other servers. git might allow them to stay of the main development tree for a time - if that's allowed by company policy - but it doesn't really change how they work. People use a lot less of version control systems than sometimes forums like this would have you believe. They're a tool. They don't spend all day debating the best way to work.. heck, a significant proportion never even merge in the way you or I would.. the fire up WinMerge, view the differences and merge manually, then commit the changes without any merge information. That's the way they like it and they're happy like that. Us going on about cool new features doesn't affect them in the slightest (I believe Arthur has worked on ways to track that kind of merge, but don't know how far he got). If you get 90% of the functionality right then you cover 99% of the users. The others are the noisy ones that complain on forums :) That is just as true for cvs and svn; a full copy (except for some metadata) is just one 'git cvsimport' or 'git svn clone' away. .. and that would be logged in the central server, and be noticed, along with IP and time/date. Pulling down an entire cvs repository is a long process (even more so for svn). We do it for migrate and it can take many hours. For git the whole process is streamlined - even expected, so you wouldn't think anything of someone doing it... so it would be much harder to nail down one particular user as being the problem. Still, I'm amazed at the speed with which git came into existence and is running circles around svn (except in the gui department). I presume git is based on arch, so had an existing system to base itself around. svn always had its problems (I'm no fan of svn.. even less so after I implemented the protocol) but is still extremely popular, and will remain so for some years I expect. No system will gain the ubiquity that cvs had a few years ago, simply because there's so much choice now. A few years ago you had SCCS, RCS, CVS and a whole bunch of expensive commercial ones, most of which didn't even support concurrent development. It wasn't surprising that CVS became so popular. git versions trees, not files. Hence you can't have file-wise merge arrows, and you can't represent them. This is a purposeful difference which affects the way you organize projects, as you can't just keep using the cvsnt way. You'd have to derive them from the changes to the tree, but they're there since a tree is made up of files. ie. just as you can say 'this tree is a merge of these two trees' it's equivalent to saying 'the files in this tree are merges of the files in these two trees'. It's broken, actually: http://blogs.open.collab.net/svn/2008/07/subversion-merg.html Wonder how they screwed that up? I'd have never guessed from the protocol - although requiring the entire merge history to be sent to the client rather than just a mergepoint was probably a clue. (aside: I note the technical language... a pet hate. When I put mergepoints in cvsnt in 2002 I never used any word more techincal than the word 'merge'. Even the word mergepoint was originally just a corruption
Re: [cvsnt] Betr.: CVSNT to CVS
btw. As it's late, I'd like to go on record by saying that I think git is a good thing - I can pull down an opensource project, fix bugs, post on the forum and anyone that wants to can get my fixes/improve them without waiting for the original author to get around to releasing a new version - so you end up with a truly open development system. There's even a good case for putting the cvsnt repository in git (which would be somewhat ironic), as it would benefit from this kind of collaboration. I'm just not sure how well it translates to the corporate space.. and currently that's where all my thinking is, having had my head in evs for so long. Now if someone wants to come up with the idea for a corp-git that handles realtime audit, bug tracking, access control and locking without losing the distributed nature (if that isn't a total contradiction in terms) then I'd be all for it. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Betr.: CVSNT to CVS
Bryce Schober wrote: This last point can't be stressed enough, in my experience. For whatever reason, CVSNT never got the widespreadd acceptance in the software development community that other SCMs have. I disagree. The download figures disagree with you too. CVSNT was as mainstream as it gets.. it was never a Firefox or Redhat, but millions of people use it every day. Of course, from a commercial point of view acceptance means nothing unless it generates revenue. It's been said repeatedly that if 1% of people who used CVSNT bought 1 copy of Suite we could employ a number of developers full time. If that model won't work then we're content to ship to 1% of the number of users and guarantee paying my salary every month. That's a sad thing for someone like me, who loves opensource, to be saying. Yes, I have considered removing opensource support and just leaving commercial. I haven't though.. because I still think opensource is important, and so does Arthur. You think mainstream is important.. It really isn't. cvsnt was never conceived as such.. it happened by accident. Maybe evs will, maybe not. evs however is under a very different model to cvsnt.. it's fundamentally commercial although (most of) the source code is freely available. And so, when you do finally get to that point where the newer features make a solid business case in your scenario, you'll be stuck without an easy upgrade path. The point of EVS is you don't *need* an upgrade path. We just update evs to support the new clients as required. The work seamlessly and there's no work to do and no upgrade cost. Version control really doesn't change - at its core all systems are basically alike. Even the distributed systems aren't fundamentally different (and a git interface to evs has been mooted already). What we've ended up is a multitude of new mutually incompatible systems written all to answer a single need of one developer, but all copying each other. I've long since realized that's a stupid way for the industry to continue. EVS is me going Screw it, I'm not playing any more and writing a system that can talk to anything, given enough time and coffee. From the outset we handle migration from different systems, and because we can pretend to be different types of server, migration out again isn't hard either, should it be needed. Crucially EVS is *not* a new version control system. It will be whatever you want it to be. It can be CVSNT, or SVN, or TFS or (potentially) Git. Or all of them at once, if you want. There is no 'EVS Protocol' and there never will be. So you can upgrade to 'Version Control system of the week' and have *exactly* the same problems a couple of years down the line, or you can use something that's going to grow with you. And if it doesn't... there will be lots of ways to migrate - because nobody has to support evs specifically.. it handles the support itself. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] getting the connection working
Mike Kay wrote: Greetings and salutations. Please forgive my lack of knowledge - this is the first CVS server I have setup and I definitely have a lot to learn. I am in a Windows environmentI hope this is the correct place to post - as I am unsure where the problem exists. Here's where I am at. I have CVSNT installed on a Server 2003 box. Version is 2.5.04(Zen)Build 3236. I am using WinSSHD for an ssl connection. I have port 22 open on the firewall and have port forwarding to the default CVSNT port. You're running an SSH daemon and forwarding the connection to port 2401? It sounds like a *very* unusual configuration - you normally either have the ssh server calling cvsnt directly or forward the external port directly to cvsnt. I know of no client that would handle the configuration you're trying out of the box... but it may be possible to set something up. Start by keeping it simple. Forget WinSSHD for now. Setup a basic SSPI connection between the client and server using default configuration, then add elements as required. If you need to encrypt tell the server to require encryption, and that'll give you reasonable security. It's not uncommon for people to just leave it at that and tunnel their connections over VPNs rather than have them over the public internet. Configuration of WinSSHD I can't help with... however basically you'd have to make logins with the particular user allow the execution of the cvs server command only (and set the paths/environment to allow that), and make sure it's a pure binary connection (no cr/lf fiddling or anything similar). It's far easier to just use sserver or encrypted sspi though... we don't generally recommend ssh on Windows as it's complex to setup and the security historically hasn't been that great. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] info.dll - trust verification failed
Olexij Tkatchenko wrote: Hello, a supplier is using our repository via cytrix. Recently he started go get the following message: Executable file 'D:\cvsnt\triggers\info.dll' trust verification failed - executable image not signed He is unable to perform CVS operations. What does this message mean? Is there a workaround for this behaviour? Something has modified the files on your system. Scan for viruses etc. and if you're sure you're safe reinstall. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Using extnt.exe for enabling SSPI with the requiredencryption option
Arthur Barrett wrote: Carlos, This is by design (go to the bottom of this page): http://march-hare.com/cvspro/security.htm Another combination that you did not try (enabling Kerberos in SSPI and disabling NTML) also does not work which seems a little inexplicable so I've raised a bug on that: http://customer.march-hare.com/webtools/bugzilla/ttshow_bug.cgi?tt=1id=5533 That won't work. extnt does not handle encryption except where it's inherent to the protocol (ssh and sserver). eclipse should have a flag to enable encryption on the client - the equivalent of using cvs -x from the command line. This isn't something extnt can help with - it's up to the client to initiate the encryption using the protocol. Actually thinking about it even that may not work work with extnt binary as it would have to know the moment the client requests encryption and start using it... however without the help of eclipse it's a non-starter. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Problem in debian
Claudio Guzman wrote: cvs -t -d :sserver:demo:d...@190.21.50.231:2402:$CVSROOT/demo login 16:45:48: - Tracelevel set to 1. PID is 20884 16:45:48: - Session ID is 519449a44e6c6687 16:45:48: - Session time is Tue Feb 24 19:45:48 2009 16:45:48: - main loop with CVSROOT=:sserver:demo:d...@190.21.50.231:2402:/usr/local/cvs/demo Logging in to :sserver:d...@190.21.50.231:2402:/usr/local/cvs/demo CVSLock 2.2 Ready cvs [login aborted]: Connection to server failed ??? Why are you trying to connect cvs to the lock server? Just forget the port number and let it use the default 2401. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] lock service eating resources
Arkadiusz Miskiewicz wrote: Looks like cvslockd does really weird things. It selects with very low timeout (~1s) which renders 1-1.5% cpu usage all the time on my notebook. It wakes up, checks a couple of things and falls asleep again. When idle it'll not use even a measurable amount of CPU - a few cycles a second. I just looked at one of our sample servers and the usage is 0%, as expected. When under load its CPU usage can get quite significant (although the amount is small compared to the main cvsnt processes). Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Import Tags and History
David Hauck wrote: Hi, It appears that import tag events do not get included in the history file (i.e., subsequent cvs history -T requests do not report these tags). Is there another way to capture/persist (outside of leveraging the audit feature with database backend) these events (perhaps via taginfo?)? In particular I'd like to see event records for both the (newly created) vendor tag and vendor release tag. They're created as an artefact of creating the new RCS file.. import is somewhat a law unto itself regarding this, so there are currently no hooks that can be used to monitor it. I suspect they can be added, but I haven't looked at that code in a while so you'd have to ask Arthur how easy it would be. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Re-using Previously Renamed Branch Tags
David Hauck wrote: Hi again, I've just been informed that this mechanism *is not* supported in non-EVS versions of CVSNT. Specifically, the deletion of branch tags, in particular when at least one alias tag is present, is not supported. The method in the post you mention is still supported and is the recommended way to do it. What isn't supported (but isn't specifically prevented) is removing the last reference to a branch - you can end up with no reference to any of the revisions in the branch, which causes problems. The Evs branching mechanisms are completely different (in fact in evs you can't delete a branch in the way cvs does at all, merely mark it as deleted at that point - a branch is a versioned object just like everything else in evs). Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] CVSNT 2.5.04 build 3236 server really slow
Peter Crowther wrote: error. Given that the default for many firewalls is now to drop packets rather than return any ICMP status, I can entirely see why a naïve configuration could cause this problem. From the LAN? I can see why a firewall would drop packets on the WAN (I've never agreed with the reasons, but it's arguable) but one that would do it on the LAN seems to be somewhat broken IMO. The authors have probably been reading too much Gibson :p I wonder if we should devise something in cvsdiag that could analyse the health of the network - timing DNS/RDNS lookups, checking the reachability of the default route, external connectivity etc. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] CVSNT 2.5.04 build 3236 server really slow
Peter Crowther wrote: I'm somewhat more interested as to how a correctly configured DNS for this environment would either take 3 seconds to fail, or return an IP address that can't then be pinged. Surely it should fail immediately without returning an address? So I agree with you that the DNS needs work, just not about the symptom :-). Correctly configured DNS in any configuration will return almost immediately.. it'll either fail or succeed. I'm not actually sure how a machine not connected to the internet could take 3 seconds to resolve a name.. that's a new one on me. Even if it somehow returns an unpingable address as there's no default route the machine will just return immediately with a failure, so no harm done. If for some reason there's a default route given by a misconfigured dhcp server that router will immediately return an ICMP host unreachable or similar, so that doesn't really slow things down. What we appear to be seeing in some cases is a combination of misconfigurations where a network is setup as if it is connected to the internet but isn't.. that's quite creative - I can't think of a way to do it (well not a sane way anyway.. certainly nothing that could be done accidentally). Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Ignore merging a certain text file
Attila Strba wrote: Hi Guys, is there a possibility to setup in the sandbox a rule to avoid merging a certain file? In Keil uVision if the *.uv2 text file is merged it usually gots corrupted. The file can't be in ignore list because if creating a new sandbox the project needs this file. thanx for any help, You could just mark it as binary, then it'll never get merged. If you don't want changes to it to be committed as new revisions, then mark it as static. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Bug id format?
Gerhard Fiedler wrote: Hello, is the bug id a string or a number? I didn't find anything in the documentation, so I ran some quick tests with a string that contained letters, dashes and numbers (123-abc), and it seemed to work. Was this by chance or by design? It's just an ID... it depends on the bug system you're using what you treat it as. There are some limitations to the characters you can put in it but I can't remember them off the top of my head. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Vanilla install on SLES
Tom Halliley wrote: When I do: cvsnt -ttt login I notice that the wrap_add lines, e.g., 12:31:05: - wrap_add(*.ncb -kb, 0, 0, 0, 0) list the default wrappers, not what's in my cvswrappers file. Is this a hint? That's just normal startup. You should find authentication logs in your system logs somewhere that'll say why the login is being rejected. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Vanilla install on SLES
Tom Halliley wrote: I added this PAM configuration: tools...@source1:~ cat /etc/pam.d/cvsnt #%PAM-1.0 auth includecommon-auth account includecommon-account password includecommon-password session includecommon-session ... but still got the same authorization error :-( More ideas? Tom That's not a syntax I've seen before.. No ideas beyond that, as I've never used SLES. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Vanilla install on SLES
Tom Halliley wrote: I'm trying to get a vanilla install of cvsnt running on openSUSE/SLES. What I'm getting is an authorization error: cvsnt [login aborted]: authorization failed: server source1 rejected access to /export/local/home/toolsadm/CVS_REPO for user toolsadm What you have looks OK. Do you have cvsnt defined in PAM? It'll probably be /etc/pam.d/cvsnt (not sure on SLES). You should see the login on your auth logs (possibly /var/log/auth.log but again I'm not sure which name SLES uses). Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] CVSNT server tries to write non UTF-8 characters infileattr.xml
Arthur Barrett wrote: 1) fileattr filenames are not UTF8 when CVSNT Server is not running in UTF8 mode That'll probably work from cvsnt clients, but not others. The problem is a deficiency in the old CVS client/server protocol. It predates any widespread use of multilingual character sets so has no way for the client to tell the server what character set it's using, and no way for the server to tell the client. cvsnt adds that functionality but it can't address the problem of what happens when a legacy client tries to send an 8 bit filename to the server - it has no way of knowing what character it is seeing, so it assumes it's the same character set as the server. This works more or less when they're both the same OS, but not when they're not, as the defaults are often different. The stock answer is to put both client and server into utf8, so they can both handle the full range of characters that can be sent (evs operates exclusively in utf8 to avoid this problem as much as possible). It should be possible for the java client to expect utf8 filenames.. after all, even on legacy cvs you're going to need to know what character set the filenames are being send and received in otherwise it's all going to get very messy.. A similar problem exists for log messages - the character set never got recorded by RCS so old log messages could be a mixture of ISO8859, Shift-JIS and Martian for all that cvsnt knows.. that particular problem never got solved satisfactorily, but luckily it hasn't been a major problem. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Message form list come with reply to the individual sender
Peter Crowther wrote: Hi, The messages come with reply to: set to the individual sender. Please keep this the same! Sometimes I forget on other lists and send critical and private information to a mailing list rather than just to the person I intended. Don't worry it isn't changing.. the last time this came up was quite a while ago (years, IIRC) and the reasons were gone into then. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] RCS format question
Fabio D'Alfonso wrote: Hi, I tried first yesterday to use a native cvsnt repository with bonsai and I got the following message: Unexpected parsing error in RCS file . Expected token: date, but saw: text Before starting any investigation I would know if the RCS format is different in native repositories, respect to the cvs 1.11 (it worked on a cvs 1.11 repository). We do stick to the RCS spec except with respect to binary deltas and extra expansion options (the latter shouldn't really affect a parser, but may affect the rendering of the final file). However some RCS parsers are very inflexible with regards to the exact format, especially the parsing of new keywords, of which we use several. As Arthur said in this case it's best to use the cvsnt rcs wrappers to parse the file, which will always give you the correct result. Ideally don't use tools that try to parse the RCS files directly as these are prone to breakage. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] How to invoke internal mailer
Bart S wrote: If I only tag a revision, I see in the trace that the mailer is invoked only once. The [to_user] is not expanded correctly in the TO field, and because upon tag I see in the trace that the mail is sent only once, the SMTP session ends with an error after the to field (invalid address) thus no mail is sent. Could this be the reason that it looks that the internal mailer only works on notify mails ? to_user is for notify so that mails get sent to the watch list. It doesn't make sense on tag emails because they don't have any kind of destination user to send to. If you want it to send to multiple users setup a list on the mailserver and send to that list. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] add watcher automatically(cvsnt2.5.04-3236)
kom...@gmail.com wrote: I do so - cvs watch off then add file. Again: file name=testwatch watcher name=user/ That is a disabled watch. For it to be enabled it must both have a 'watched' tag and a list of items to be watched. Looks like it's working correctly to me. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] cvs init problem
Tom Halliley wrote: bash-3.00$ cvsnt -f -d :pserver:[EMAIL PROTECTED]:/export/local/home/toolsadm/CVS_REPO -n init cvsnt [server aborted]: Must specify root for remote init command The error is corrrect. You need to specify a (relative) root to the init command with -r. Personally I'd forget about remote init. Apart from the security issues of people being able to create repositories on a remote machine, it's difficult to get working correctly anyway. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] cvs init problem
Tom Halliley wrote: Permissions look good on PServer file: bash-3.00$ ls -l /etc/cvsnt/ total 304 -rw-r--r-- 1 toolsadm toolsadm6954 Dec 5 09:02 PServer -rw-r--r-- 1 root bin 6545 Nov 3 14:24 PServer.example -rw-r--r-- 1 toolsadm toolsadm 630 Dec 4 18:55 Plugins -rw-r--r-- 1 root bin 558 Nov 3 14:24 Plugins.example -rw-r--r-- 1 root bin 124609 Nov 3 14:24 ca.pem bash-3.00$ who am i toolsadm pts/1Dec 5 08:50 (troy1hl1485535.corp.fsroot.flagstar.com) What about write access to /etc/cvsnt? Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Problem with / in $Id$ keyword replacement in Windows domain
Tim Chippington Derrick wrote: public class OptimisationScope { public static String versionInfo = $Id: OptimisationScope.cs,v 1.18 2008/10/24 14:42:12 Company\TimDerr Exp $ $Name: $; IIRC in C# you can switch off escaping by using @ in front of the string. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Does gserver support impersonation (Windows Active Directory)?
Dirk Weinhardt wrote: Hi! I'm using CVSNT 2.5.04 with Windows XP (server and client). I have managed to set up a CVSNT server that authenticates users against a Windows Server 2003 AD domain (I'm using the gserver protocol). cvsservice.exe is run as mydomain\cvs, SPN cvs/myclient.mydomain is mapped to mydomain\cvs. Run as user is set to (client user). I'd expect cvsservice.exe to spawn an instance of cvs.exe as the user that connects to the server (e.g. mydomain\dirk). But instead cvs.exe is started as mydomain\cvs. cvs.exe is always started as LocalSystem, then it changes just after authenication. Don't change the user for cvsservice.exe unless you're sure you know the consequences - and when setting up definately don't as it'll just introduce other issues. cvsservice.exe creates its own SPN - I'm not sure it'll even authenticate unless it's mapped to the machine account - kerberos is very picky about what it allows. Unfortunately that prevents me from using NTFS permissions to control who may access the repository. If it's not impersonating it's probably a result of changing to a nonstandard configuration - get it working first, then start changing things one at a time. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] CVSNT on Win2003
Bo Berglund wrote: What *can* be done, however, is to check out the top of the repository to create the sandbox copy, which will include CVSROOT as a submodule. The name of the top of the repository for checkout purposes is . (that is a single period). Example from the command line: One thing that I've done in the past (but don't know if it works on all client/server combinations) is to specify the repository in an add eg. mkdir mymodule cvs -d :pserver:foo.bar.com:/cvs add mymodule Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] creating a repository on Windows Xp Sp3
Sanjeev, Neeraja (GE Infra, Energy) wrote: Hi, I was trying to create a repository in temp location on Windows Xp Sp3 PC for logging in and it keeps failing . When I see a repository folder on PC, it shows only the CVSROOT folder but not the CMS_USER_DATA. That is the reason for the login failure. But same works on Sp2. There's no real difference between xp sp2 and sp3, so it won't be the problem. When you create a repository you would only expect to see CVSROOT. Any extra modules you will have to import yourself. The existence or not of any imported modules has nothing to do with your ability to login (in fact to import you'd probably have to login in the first place, unless you were doing everything locally which isn't recommended). Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Unable to create sub-directories on a branch even with ACL set to all or read, write, create
Terry Beavers wrote: I am trying to grant a user permissions to create a sub-directories under on a branch using the following ACL and CVS is not allowing the operation. I had submitted something similar in regard to creating new files on a branch back in early October, and have since resolved that issue, but I am still unable to create/add new directories on a branch. In CVS directories exist on all branches so you can't restrict them to a single branch. You can restrict creation of files within that directory (so a co -dP or update -d will not create the unwanted directory). cvs chacl -R -r SKYWAY_BRANCH -u terry -a all skyway/ Avoid using -R if you can - permissions are inherited anyway and putting them on all subdirectories just duplicates them. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Permissions issue
Chris Delaney wrote: Hello, When I add a directory to a module via WinCvs client, the directory does not inherit the permissions of the parent directory. The new directory has no permission settings. It does, but cvsnt can't show it in lsacl (an artefact of the way permissions are calculated unfortunately). evs can show effective permissions as it uses a completely different method. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Update with non-existing tag deletes sandbox files
Arthur Barrett wrote: Jørgen, This bug applies to CVSNT version 2.5.04 build 3236. If a module is checked out or updated, with the -r option, (for example: CVS up -r this-tag-does-not-exist mymodule) and the tag does not exist (maybe because you missplelled it) CVS will report every file in the module as no longer in the repository and DELETE the file from the sandbox! Previous releases of CVSNT - correctly - reports the non-existing tag as, well - non-existing: no such tag this-tag-does-not-exist. To me the behaviour with 2.5.04 looks correct and the 2.5.03 behaviour broken. What do other users think? 2.5.03 and before used to have to go away and search the entire repository just to give that message... it led to a number of posts on here as it could take many minutes to complete. Even after that if the tag existed anywhere - even a completely different module - you'd get exactly the same behaviour as 2.5.04. So the time spent was essentially wasted. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] CVSNT(2.5.04.3236) do not work with chown command
voozoo wrote: Arthur, Thanks you for answer my question. I am a newcomer of CVSNT user, so I do not understand the development of CVSNT. What is the ['to test' list], is it means [chown] command is not be supported in 2.5.04.3236 version? Where can I find the ['to test' list], is it in the source of CVSNT? Which old version is support [chown] command? It's actually lsacl that's not displaying it.. chown is working. It's not that useful a function, which is why it can break with nobody noticing. It just sets a name on the directory gives control access to the owner - the second part is probably better done with an ACL anyway. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] CVS Server Side Reporting Tools?
Arthur Barrett wrote: Wheras if you wanted to call audit.dll the sequence would be something like this: // This creates the SessionLog entry initaudit(NULL, commit, 2000-11-01 10:42:15, localhost, fred, /cvsrepo, d:\cvsrepo, 01234, NULL, 0, NULL, NULL, '2.5.04', NULL); Nice idea but those functions aren't exported.. in fact they're defined within a c++ anonymous namespace so they're not even called that.. they'll have some mangled name internally (not sure why at some point they got 'audit' stuck in their name). To call it you need to go via the api (CLibraryAccess) or you can do it manually starting with get_plugin_interface in audit.dll - something like: plugin_interface *p = get_plugin_interface(); p-init(p); trigger_interface *t = p-get_interface(p, pitTrigger, NULL); Then you call as before: t-init(NULL, commit, 2000-11-01 10:42:15, localhost, fred, /cvsrepo, d:\cvsrepo, 01234, NULL, 0, NULL, NULL, '2.5.04', NULL); t-pretag(NULL, , /module/directory, 2, {file1.txt,file2.txt}, {1.2,1.3}, 'T', , Rel_1_2_0); And to close down: t-close(t); p-destroy(p) As that's entirely a C interface it should be possible to reproduce in delphi. Cross platform note.. get_plugin_interface may be renamed by libtool, since on some platforms (eg. HPUX) you can't export functions with the same name from two different ibraries. This won't be an issue provided you use the libtool functions to find the name. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] CVS Server Side Reporting Tools?
Arthur Barrett wrote: I've threatened for years to write a 'C' layer on top of the existing API (meaning a laywe without the need for pointers to functions and structures), and I've only avoided it because noone was really showing any interest in using the API. If I wasn't so busy right now with EVS (and Suite and CVSNT) I'd spend a couple of days on it right now so that Bo has something to work with. It's a lot more than a couple of days work.. most of the API is C++ and has all sorts of dependencies. You would be better off with a COM API which would work with anything, at least on Win32. I tried to get swig (www.swig.org) to work but it doesn't support enough of C++ to use our headers unmodified (that was some months ago though, and I tried it on evs which is a much more complex API). Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] March-Hare message into my commit messages
Luigi D. Sandon wrote: But of course it won't work if you use a GUI on top of CVS that doesn't support it - unless one recompiles them too. And as I noted in a previous post, the lack of a good GUI (IMHO Tortoise isn't, saw too many users did the wrong thing due to the shell integration, nor Workspace Manager interface looks up to the task) is one of the reasons I have issues to build a case to buy CVS. Each version of GUI will use the cvsnt version it is written for, generally. By the time a Tortoise is released for 2.5.04 it'll be using either switch depending on what Torsten wants to do with it. Clients which use their own code like eclipse don't even have the issue. I've never seen a completely seamless GUI for any system.. SmartCVS is pretty good if you haven't tried it. Really, there are so many clients out there that it's not something that really factors into it.. and there are as many different ways of interacting with a repository as there are clients. Everyone has their favourites - the key to making it all work is at the server side. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Betr.: Re: Betr.: Re: March-Hare message into my commit messages
Gerhard wrote: I also don't know. One thing I observe though is that there are many, really many people who don't seem to know that (and how) CVSNT is different from CVS. Most online resources that compare version control systems seem to include CVS and SVN and others but not CVSNT (and if so, often only as a sidekick of CVS). That's a huge issue that we've often talked about, hence the way CM Server is an entirely different product. OTOH that's a different audience.. it's bigger, more complex, *way* more powerful (we haven't even implemented many of the things the evs core is capable of yet) and hence a more complex install. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] March-Hare message into my commit messages
Luigi D. Sandon wrote: it's free and they already paid the GUI to make it work. Most managers who sign the bill don't understand very well why we should by a VCS and then buy a GUI separately. Most peopple don't buy the GUI separately, they generally use what they have.. eg. Visual Studio or Eclipse. We also ship Tortoise which is a very good GUI that beginners have no problems getting used to, and of course Workspace Manager that handles the more advanced features. IMHO if CVSNT Pro offered a very good GUI which can take advantage of the advanced features - developed internally or licensed - the perceived value would increase a lot - company would have to buy one tool to get everything, like most commercial VCS offer. What's inside the box now, customized version of open-source projects, looks not enough, IMHO. You *do* only have to buy one tool. CVS Suite contains everything you need. You make it sound like it's a bunch of free stuff thrown together. Nothing could be further from the truth. It contains extra features - such as bug tracking and visual studio integration. The ebook is worth the price alone IMO. Even the packaged apps are customised to work better with our system have more features that aren't in the free releases (we do contribute changes back to the projects but some of it is too specific to cvsnt to make it into any generic releases). Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] March-Hare message into my commit messages
Fabio D'Alfonso wrote: Hi, just to say that I agree completely with Luigi. You should go proprietary, if you are no longer satified with open source model. We really like the open source model.. evs is also mostly open source and will likely become more so over time. Adding a small (removable) advert to cvsnt does not change this one bit. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] CVS Server Side Reporting Tools?
Bo Berglund wrote: In order not to double up the database access methods maybe something like this could be arranged: - a CVSNT API through which one could call the database write functions to create the entries CVSNT Audit normally does for CVS operations. - a CVSNT API through which one could get the data needed to be used for the above calls. Then the CVSNT settings for the Audit database would be used without the need for the external application to access the database directly and the data would be retrieved without needing to parse the RCS files directly. In theory you could load the audit library yourself and do that - in cvsnt it's even a C interface so should be easy enough from delphi. You could access the data with a combination of rlog and rdiff.. although not ideal as it involves text parsing it has the advantage you could even do it against an old cvs server. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] ipv6 handling in cvsnt is broken
Arkadiusz Miskiewicz wrote: A couple of points: This rule only applies to ::, not ::1. This is entirely irrelevant as there are only two possible cases. The code is written to handle both cases transparently, and does. $ netstat -an | grep 2402 tcp0 0 127.0.0.1:2402 0.0.0.0:* LISTEN tcp6 0 0 ::1:2402:::* LISTEN $ netstat -an | grep 2401 tcp6 0 0 :::2401 :::* LISTEN cvsnt doesn't just cater for a couple of OSs - the commercial builds run on all sorts of environments and the code to handle these cases has to be (and is) robust. You are wrong here. Tell me how to specify LockServer being ::1 with port 432 in PServer config file? Parsing is done in cvsnt, I see strchr used there for example. ip6-localhost:432 I can see absolutely no reason why anyone would want to do this. The config is there to change the port in case there's a conflicting application. Changing the address only makes sense for multi-server configurations and those are pretty rare in themselves. The only thing missing from cvsnt is the -4 and -6 command line options that are in evs. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] ipv6 handling in cvsnt is broken
Arkadiusz Miskiewicz wrote: Again. You're trying to claim something is broken that I use every day. I even posted examples of it working correctly. You can check yourself by connecting to cvs.cvsnt.org over ipv6. Please read http://www.cvsnt.org/wiki/BugReporting and provide all the information requested. Unless enough information is given to reproduce a problem then it's premature to start posting a 'fix' (especially an OS specific one, and *especially* when you haven't even said which OS its for). Somehow cvslockd is bound only to ::1 and not 127.0.0.1 here. Which is weird since getaddrinfo(NULL seems to be used). If getaddrinfo is broken then no software is going to work correctly. This is an issue for your OS vendor to fix. Attempting to modify all the software on your computer to work around it is futile. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] update -f option documentation (TortoiseCVS too)
Gerhard wrote: Hello, In the CVSNT manual it says about the update -f option: If no matching revision is found, retrieve the most recent revision (instead of ignoring the file). In the TCVS Update Special dialog, there is a checkbox with a label that sounds similar. However, I think both are wrong. When I use this option, it seems that if no matching revision is found, it retrieves the current revision on HEAD -- even though there may be a more recent revision on a branch. Is this correct? The command line help says: -f Force a head revision match if tag/date not found. Which is what appears to be happening when you do it. Sounds like a documentation error. The text you quote appears in Cederqvist dating back to 1992.. you're the first to notice! Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] March-Hare message into my commit messages
Clóvis Garcia Marcondes wrote: Hi, After installed the new 2.5.04 CVSNT Server (Free) the text below is been inserted into my commit messages: Committed on the Free edition of March Hare Software CVSNT Server. Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/ Is there anyway to disable that or I will have to go back to the 2.5.03? You could just buy a copy. The development, testing and support of cvsnt (and evs) takes up a lot of time and money. The software is then given away free, no strings attached. For this all you have to do is cope with an unobtrusive message on commit. There are even multiple ways to disable that if you dig around the source a bit. Arthur once calculated that if 1% of the people who download cvsnt each year bought 1 copy, we'd be able to hire 3 programmers. That would mean a much better product, better support, and it would still be free for the 99%. It pains me to have to do something even so minor.. I'd love everything to be free but unfortunately I haven't yet managed to perfect a way to live without food or a roof over my head. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] March-Hare message into my commit messages
Clóvis Garcia Marcondes wrote: Tony, I got your point but should be nice at least to advise when you change the way the product works. I just asked because I´m testing the new version together with the CVS Suite Trial exactly to see if it is interesting for me to buy the comercial version and thought it could be a side effect of that. But if it is a new March Hare policy you should at least warned about that. It was mentioned in the release notes for the release candidates and the final release notes. It's really not a change in the way the product works.. it still does exactly the same things as always - I'm against any kind of crippling or time limiting in software where people may be relying on it continuing to work. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] March-Hare message into my commit messages
Fabio D'Alfonso wrote: Hi, I appreciate very much the effort you made to enhance cvs. This shareware like change in the policy to add messages in the software, will not get expected results, in my opinion. I unfortunately expect that this will reinforce migration to other tools like, but not only svn that your efforts rightly contained in the past years, adding missing features to make cvs better. Also the most of people will stay on previous versions, unless you'll remove from downloadables. Well it's a tradeoff.. losing some to gain a few who can fund the project. It only affects the builds we make, so the various linux versions etc. will not have the message. We already (I think) told Torsten how to disable it in TortoiseCVS as well. We already posted how to switch it off once on this list already, for the more observant :) Ultimately evs is the answer - its versions are differentiated and the free version is clearly a trial. OTOH it serves a different audience. There's a need for both products to continue, and making people aware that there even *is* a commercial version of cvsnt (given that only about .01% of the users actually frequent this list) is tricker. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Using bugid with WinCVS
Andy Southby wrote: Arthur, I would really like to use this option. Could you let me know if its available in the open source code or is it proprietary to March Hare? (Obviously if I wanted to compile it). Nearly all features are in the public repository, but I'm not sure about that one - it may be in the bug trigger which is our own bug integration library. Arthur is on a plane for about the next 18 hours and he's the one that knows for sure. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] keywords admin file syntax?
Andy Southby wrote: Arthur, When I add and commit files any ware within my 'ThirdParty' module I wish the keywords to be ignored. Just add them with -ko. I realise that would do the trick but I have configured the server to use the 'Ignore client-side force -k options' to try and achieve some consistency. The developers all use their own choice of clients some of which force these options and the defaults cannot be changed without user intervention which invariably is a bit hit or miss. This is my attempt to enforce server side control. But with the configuration setting I have made (above) the -ko option is understandably ignored. I did not really want to kick everyone off the server and change this option whenever ThirdParty code needs adding. The manual seemed to suggest that the keywords file was the way to go. Is this keywords file problematic or have I missed the point of its function? It should work... the idea is to be able to import 3rd party sources and keep their keywords as they are whilst using your own. The module name it matches is the real directory name after translation by the modules file so you might be matching on the wrong name. I suspect ^ThirdParty$ will only match the ThirdParty directory and nothing below it, and you might want to remove the $ off the end (looks like a documentation error). If that still fails look at the traces as it's reasonably verbose should tell you what's wrong. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Latest Updates - CVSNT 2.5.04 Build 3236 (stable)
Luigi D. Sandon wrote: It's an MSI bug - we don't provide the dialogs for the uninstaller, or Maybe, but I never saw it in any other msi based installer- which tool are you using to generate the setups? It looks as if it is using some Asian fonts or the like. It's Wix - which is a Microsoft tool, so should generally be OK. The uninstall dialogs aren't in the MSI database AFAIK, they're part of the MSI uninstaller itself which ships with the OS. As it's uninstall it's never really been considered important. It's been doing the same thing for years... I first noticed it testing NT4 builds and that was a long time ago. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] CVS Server Side Reporting Tools?
Fabio D'Alfonso wrote: I would know if the feature supports importing data from an existing repository or works only on data subsequent the activation. Audit records every action, whether it succeeds or not, including things that don't modify the repository. There's no way of getting that kind of detail retroactively unfortunately. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] 2.5.04.3125 (RC8) Very slow to start processing!
Arthur Barrett wrote: No valid ipv6 address found for host I'm pretty sure this is the reason. Your PC has IPv6 installed but it is not configured. This error comes from mdns_mini and means ipv6 is not installed (if I'm reading the code right, anyway). It probably needs re-wording or removing as an message as the lack of ipv6 is generally uninteresting. OS vendors are shipping IPv6 to install by default now (including MS, RedHat etc) however very few organisations configure it or even when it is configured it's dodgy, I seem to be constantly having to specify rsync -4 or ssh -4 or scp -4 because of IPv6 network problems. CVSNT probably should have a 'use IPv4 only' setting (both the client and the server) but it'll have to wait. It takes more than having it enabled to cause something like that.. it would both a machine running router advertisement to both give it a global address and advertise a default route, but have that router unable to actually respond. That's quite hard to do accidentally... the equivalent of 'accidentally' installing a DHCP server, then stuffing up the config. That said, in common with most other apps we should probably have a -4 switch to cover these cases (I must admit it's never occurred to me before as our ipv6 works :p). Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] how to disable line ending translations?
Arkadiusz Miskiewicz wrote: On Thursday 23 of October 2008, Arthur Barrett wrote: Arkadiusz, Yes there is it's -kb, I wrote that in a later message but maybe you thought it was a duplicate since it was mostly a cutpaste, I just changed -kD to -kb. Yeah, -kb is a solution. Not sure how cvs stores such data in rcs files. Still as diffs or each revision keeps complete file? -kb doesn't change the storage mechanism. -kB uses a binary diff to store the files which is generally more efficient for things like executable files (at the cost of an increase in memory use, and a slight speed hit). Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Difference between exclusive edit and resrved edit?
Bo Berglund wrote: I looked at the dsecription inside the cvswrappers file itself and got confused # Format of wrapper file ($CVSROOT/CVSROOT/cvswrappers or .cvswrappers) # # wildcard [option value][option value]... # # where option is one of # -f from cvs filter value: path to filter # -t to cvs filter value: path to filter # -m update methodology value: MERGE or COPY # -k expansion mode value: b, o, kkv, c # # and value is a single-quote delimited value. # For example: #*.gif -k 'b' Does this mean that *only* the values b, o, kkv and c are allowed as arguments? No, they're just examples. The -f -t and -m stuff has no meaning any more - I assume that's an old repository has the comments current at the time it was created. Maybe it should be like this instead? *.package -k 'b' -k 'x' *.padstack -k 'b' -k 'x' *.part -k 'b' -k 'x' *.symbol -k 'b' -k 'x' Or *.package -kbx *.padstack -kbx *.part -kbx *.symbol -kbx Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt Upgrade to CVS Suite for more features and support: http://march-hare.com/cvsnt/
Re: [cvsnt] Switch to user failed due to configuration error
Arthur Barrett wrote: It may be easier to bypass all this token stuff if we switched from using setuid.dll to use cyglsa64.dll and modify the calling codein trysuid() of cvsnt/Windows-NT/win32.cpp: http://cygwin.com/cgi-bin/cvsweb.cgi/src/winsup/lsaauth/?cvsroot=src Interesting.. cygwin wrote one eventually then. Looks like sections of it are based on setuid but they've missed out quite a bit for some reason, eg. domain logins are completely absent. I'm not sure what we'd gain by introducing a dependency on cygwin and using a less functional library.. Note that you do not need create token privilege to do any of this. You need 'act as part of the operating system', which is enforced by the Win32 system to even call LogonUserEx and invoke LSA - without that check any random user could instantly become any other user, including administrator. XP SP2 and later have an impersonation privilege too that might need to be granted (aka. the right to actually switch users given an impersonation token). cvsnt does have a legacy API - NTCreateToken - that is redundant unless it's run on NT4. That does need create token rights, but is pretty irrelevant nowadays.. it's gone from evs and should probably be gone from cvsnt by now (unless we still want NT4 compatibility). Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
Re: [cvsnt] Switch to user failed due to configuration error
Arthur Barrett wrote: 21:43:18: S - SuidGetImpersonationToken returned 0554 btw. That error means that cvsnt is not correctly installed (the helper is not registered with the system). Not a lot is going to work without that.. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
Re: [cvsnt] Using -r and -D together while applying a tag
Tony Eva wrote: I'm probably being thick here but I can't see how to get the equivalent of the hypothetical command cvs update -rSymbolic_Tag:Date_Specifier file Anyone? (This is CVSNT 2.5.03 2382 on Linux, BTW.) cvs has simply never supported that... it's a limitation of the entries file format - you can be on a branch/tag *or* a date, never both. evs might at some point - there's code to do all sorts of things with tags that isn't being called right now, due to lack of time to implement it. In practice it's rarely needed.. it's standard to tag releases so you can get back to a known state without having to remember the exact time of release. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
Re: [cvsnt] Betr.: Re: Using -r and -D together while applying a tag
Jan Keirse wrote: I don't understand why the entries file format is a problem here. If last week I checked out the TEST branch of my system and didn't touch the sandbox it would still be here today. So why wouldn't it be possible to create that same sandbox with the same contents and all the revisions from last week today? The entries file records the sticky tag which can be either a tag or a date. Currently no syntax exists to hold both at the same time. There don't seem to be many actual checks to stop you doing it.. rtag is the only explicit one I could find - although it's possible I removed them sometime.. the code I work with is subsantially different to the cvsnt release. For cvsnt I suspect it's too entrenched in the old code to make it work, although it's Arthur's call if he wants to do it. In EVS as far as I can tell it'll just work anyway, since date is a separate attribute (and the entries file isn't used normally)... OTOH I've not had occasion to test that. It could be usefull if you have messed up your system and want to go back to a week ago because you're sure it worked last week (but didn't tag because there was no release or anything.) For that you would normally do a merge to remove changes from the interim period (or more often a diff followed possibly by a merge of the parts that don't work). Indeed I have had occasion to do this.. it's easy to break things and need to roll back a change or two. Tony ___ cvsnt mailing list cvsnt@cvsnt.org http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
Re: [Cvsnt] Question: renaming modules
On Thu, 1 Aug 2002 14:11:37 + (UTC), [EMAIL PROTECTED] wrote: Hello, a question just occurred in our company, and I dont want to crash our project, so I ask you before: We are developing a project called aaa. Now we realized, that the correct name of aaa should have been bbb. Is it possible to rename aaa in the repository-dir on the cvsnt-server to bbb? What do I have to change on the server? And what on the clients (beside changing the entry in file Repository)? You can rename the module dir but you have to be careful how you do it: 1. Get everyone to do their final checkins under the old name 2. Everyone now should delete their sandboxes with the old name in it 2. Shut down the cvsnt service 3. Rename the directory 4. Start the cvsnt server 5. Everyone checks out the module under the new name This should be safe enough. The biggest problem is if people keep their repositories under the old name - things can get very confused (someone manually renamed a directory here 6 months ago without telling anyone and it still bites us occasionally, because they renamed it when there was active development on it). Tony ___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
Re: [Cvsnt] Usage of cvs passwd
On Mon, 29 Jul 2002 14:32:57 + (UTC), [EMAIL PROTECTED] (Jonas Bergvall) wrote: Hi all, I get this error when I try to do a cvs passwd -a new_user. cvs [server aborted]: Only administrators can add or change another's password My domain account on the server belongs to the Administrators group. Server uses only pserver mode. Do you have to specify CVS-administrators somewhere in the repository? I've missed something about this command, so can someone help me? It should work, depending on the SystemAuth setting in the config file. If SystemAuth=No then it doesn't check the system account privilege, only the CVSROOT/admins file. Tony ___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
Re: [Cvsnt] Usage of cvs passwd
On Mon, 29 Jul 2002 17:03:11 + (UTC), [EMAIL PROTECTED] (Jonas Bergvall) wrote: Ok. I have SystemAuth=no. So what should the file CVSROOT/admins contain? Usually, just a list of usernames (one per line). Depending on your network setup it'll either be just the username or DOMAIN\username. It's CVSROOT/admin (no 's'). My mistake... (I'm always doing that, I'm tempted to make it open both filenames just in case). Tony ___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
Re: [Cvsnt] Can't get repository prefix to work
On Sat, 27 Jul 2002 22:44:58 +0100, Bill Jones wrote: In the CVSNT control panel, I set my prefix to D:/, my repository to /cvstools. In WinCVS (b8 build 1), I set my repository to /cvstools, and try to check out module flash. I get: cvs checkout -P flash (in directory C:\flash) Cannot access /cvstools/CVSROOT No such file or directory You're accessing in local mode - repository prefixes only work in server mode (you can specify them on the command line with '-D' but it's a but pointless). Tony ___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
Re: [Cvsnt] Couldn't connect to named pipe... error
On Fri, 26 Jul 2002 22:57:23 +0100, Bo Berglund wrote: What happens if you try the sspi protocol instead? Supposedly this will not use the named pipes so the shared drive workaround will not be needed. I am rally interested to hear what comes out of this test so please send me all details about the CVSNT server config and the client (WinCvs) setup if you are successful with sspi. It should work with SSPI, which only needs enough to perform a remote logon to the server. Overstrict domain security might break it though, in which case you'd (ironically) be down to an insecure protocol like pserver. Tony ___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
Re: [Cvsnt] Couldn't connect to named pipe... error
On Fri, 26 Jul 2002 21:10:50 +0100, P DoubleYa wrote: However, the company we're working at does not allow us to share directories between computers--it's a rule applied across the domain. We can map to specified servers under the IT group's control, but this is local workstation. If you don't have trust with the server then you cannot use ntserver. sspi should be able to do it, otherwise pserver. Tony ___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
Re: [Cvsnt] Couldn't connect to named pipe... error
On Sat, 27 Jul 2002 00:48:34 +0100, P DoubleYa wrote: As I pointed out in my other message, it is not possible for pserver or sspi to generate this error message, as they don't use named pipes at all. Something odd is happening. If I change the Authentication (Admin/Preferences) to pserver, I get the following in the console NEW CVSROOT: [EMAIL PROTECTED]:/project (password authentication) When I then try Admin/Login, I get cvs login cvs [login aborted]: The :ntserver: protocol does not support the login command You have some CVS/Root files hanging around I expect. CVS uses those in preference if it finds them. You can either (a) Delete the sandbox and check it out clean again, or (b) Change the contents of the CVS/Root files (wincvs 1.3 comes with a script which will do this for you). Tony ___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
Re: [Cvsnt] :ext: on Windows 98
On Wed, 24 Jul 2002 17:18:27 + (UTC), [EMAIL PROTECTED] (Francis Irving) wrote: Hi, I'm having trouble with :ext: (used for SSH) with version 1.11.1.3 (build 57f) of CVSNT. The error message that I get is: In C:\testeroo\TortoiseCVS: cvs95 -q update -d -P cvs95 [update aborted]: Connection to server failed Error, CVS operation failed Drat. I'll have to try to setup a Win98 machine again... That'll mean wiping the coffee stains off my setup CD :-) Can you file a bug report on it? I'm currently going through the release bugs to get 57g out of the door. Tony ___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
[Cvsnt] test
___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
Re: [Cvsnt] How to disable pserver but have SSPI working on port 2401???
On Tue, 23 Jul 2002 15:07:45 + (UTC), [EMAIL PROTECTED] (Bo Berglund) wrote: Is it possible to have CVSNT only active on the ntserver and sspi protocols? I think that port 2401 is used by both pserver and sspi, and there is no control panel setting for enabling the various protocols, so how can I have 2401 open but pserver disabled? I want to use the added security of sspi so I can open 2401 out to the Internet, but I *must* disable pserver because of its inherent lack of security. Any thoughts? Delete the pserver_protocol.dll from the server. Tony ___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
Re: [Cvsnt] Cvsweb, annotate, and LockDir
On Thu, 18 Jul 2002 19:11:59 + (UTC), [EMAIL PROTECTED] (Jeff Loomis) wrote: I am using the LockDir parameter to avoid giving all cvsweb users write per= mission to my repository files. Cvsweb can work because its impersonation = matches the permissions on the repository. The annotate feature does not w= ork. I have determined that this is because cvsweb uses the cvs server c= ommand and streams protocol to it. When operating in this mode, the cvs pr= ogram does not parse the config file and so does not pick up the lock dir p= arameter. It then tries to create locks in the repository directory.=0D =0D Is anyone aware of a version of cvswebnt that does not work in this way, or= of any other solutions to this problem. Ideally, the cvs command would ob= ey the LockDir command.=0D CVSWeb probably needs modifying to use the rannotate command rather than trying to do it itself. I did something like this once but unfortunately have lost that version. Tony ___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
Re: [Cvsnt] CVS allowed me to commit file with conflicts
On Fri, 12 Jul 2002 11:19:05 + (UTC), [EMAIL PROTECTED] (John Hall) wrote: I have cvsnt 1.11.1.3 (57f) running on Windows 2000. I just did a commit fr= om Linux which has standard cvs 1.11. As it turned out the big merge I had = just done had a number of conflicts which I had missed (you're only notifie= d of conflicts on stderr, and I had only logged stdout). How was it possible that cvs allowed files with conflicts to be committed? = I had checked for conflicts by doing 'cvs -n up', and none of the files had= a C next to them. The conflict check isn't perfect, and can miss things - it's up to the developer to check whether a file still has conflicts before committing. It does manage to catch it 95% of the time, though. A worse problem tends to be false positives, where someone has used '' in a comment. Tony ___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
Re: [Cvsnt] CVS admin group
Gavin Kinsey wrote: I'm running version Concurrent Versions System (CVSNT) 1.11.1.3 (Build 57a) (client/server) I tried the command you suggested and got the following, cvs -t admin gavink *CVS exited normally with code 0* - main loop with CVSROOT=:sspi:Coventry1:d:/cvs What does that mean? Do I need to updgrade to a newer release of the server? 57a didn't have the trace in to show the username. There are lots of bug fixes since, too. Probably it's best to upgrade if you can. Tony ___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
Re: [Cvsnt] CVS admin group
On Mon, 8 Jul 2002 15:12:21 + (UTC), [EMAIL PROTECTED] (Gavin Kinsey) wrote: Are you sure about this, I can not get it to work. I have SystemAuth=Yes and created the admin file with my username on the first line followed by a blank line. WinCVS still tells me cvs [admin aborted]: usage is restricted to repository administrators. I've tried various formats, names, and even restarted the server once in case the service only checked for the file but can't get it to work. Do 'cvs -t admin whatever' and it'll print out what username it's checking against. Domain access can do strange things to the final username. 57d didn't work properly with admin - use a later one if you have that release. Tony ___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
Re: [Cvsnt] pserver fails on win2k
On Thu, 04 Jul 2002 21:00:15 +0100, Robert Mertens wrote: I installed 1.11.1.3 (57f) on win2k sp2. The ntserver protocol appears to be working, but all calls to pserver result in an error reading from server: -1. You're running an old version of Norton Antivirus I presume This is a known bug with NAV. I tried the latest version and they seem to have fixed it. Incidentally, in order to build using visual studio and sp6, I had to include the fake_protocol library (and edit most of the dsp files). The old Visual Studio 98 SDK is ancient can't compile modern apps such as cvsnt. Download the platform SDK from Microsoft. Tony ___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
Re: [Cvsnt] odd situation
On Tue, 2 Jul 2002 18:46:16 + (UTC), [EMAIL PROTECTED] (John Deighan) wrote: I've run into an odd situation a couple of times now. If I select a file, and select 'Remove Selection' from the menus, it marks the file as removed. Then I'm supposed to commit the removal, but if it turns out that someone else has modified the file, and committed the change, I'm stuck in a 'Update' will just produce a conflict and allow you to resolve it (by removing the file). You can resurrect the file by just doing 'add' on it it should reappear. Tony ___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
Re: [Cvsnt] Does a CVSNT server have to be a domain controller to use pserver
On Tue, 25 Jun 2002 22:53:08 +0100, Dan Castenholz wrote: I get: cvs [login aborted]: Cannot login: Server has insufficient rights to validate user account - contact your system administrator whenever I use pserver from Solaris. ntserver and sspi work fine from windows. Server is a member server in an NT4 domain that has a one way trust to the user domain (another NT4 domain). The account that the cvsnt service is running under needs 'Create a system level token' privilege to use pserver. The System account has this by default normally. I don't think the trust system makes much difference - I've run clients that are on trusted domains and it seems to work, but that's always been bidirectional trust. Tony ___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
[Cvsnt] Bugzilla is back
I've re-instated bugzilla: http://www.cvsnt.org/bugzilla I'm using it at the moment to prioritise fixes try to manage my time a bit better. If there's something that I've missed feel free to add bugs. Tony ___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
Re: [Cvsnt] Bugzilla is back
On Tue, 25 Jun 2002 19:22:23 +0100, Bo Berglund wrote: I know you have your CVSNT repository on a Linux box and probably Bugzilla is running there too, but I have to ask anyway: Is it possible to run Bugzilla on a W2000 Server (the one I have CVSNT running on) to track down our own bug reports and tasks? If so is there a good treatise on how to set it up with IIS5? Can't see any reason why not... it's just Perl/CGI connected to a MySql backend. There's docs on http://www.bugzilla.org/docs/html/win32.html. There's some mention of IIS but it of course prefers Apache. Tony ___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
Re: [Cvsnt] new feature: Unix-attributes with CVSNT ?
On Mon, 24 Jun 2002 10:02:53 + (UTC), [EMAIL PROTECTED] (Matthias Mohr) wrote: Hello, I'm using the CVSNT-server with Unix (Linux Solaris)-clients. Now I have the problem, that CVSNT forgets the Unix- file attributes of my files; especially for shell-scripts which have the executable-flag set. Would it be possible for CVSNT-server to save such attributes and give it back when any cvs-client is asking for such files ? I know, that the NT-security attributes don't have such executable-flags, but I'm asking for a new feature :-) Or does anybody know a trick how to do this ? It's not that easy, as there's currently nowhere to store such information. It would require extensions to the RCS file format... possible, but it'll probably be a while. I was under the impression that the Unix server didn't do this either any more - the The PreservePermissions stuff has been commented out for a while because of bugs in it. Tony ___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
Re: [Cvsnt] Need some information to validate the product
On Thu, 20 Jun 2002 22:19:37 +0100, Brusset, Mathieu wrote: Hi, I use CVSNT 1.11.1.3 with WinCVS 1.2. I need more information about the official testing of these releases. In order to validate the product and to use it in my company, I would like to have more information about the technical testing that was done on 1.CVSNT 1.11.1.3 2. WinCVS 1.2 First question to ask: Do you really want an opensource product or would you rather pay for suppport? Opensource is not 'free' in the sense that it doesn't cost anything - you'll probably have to have someone spend some time getting to know the product and collecting documentation, helping to sort out problems, etc. A commercial product will give you this as part of the purchase price, along with certain quality guarantees. If these are important to you opensource is not the way to go (except when purchased via a third party such as Redhat or IBM who will give these guarantees for a price). However, if you're prepared to cope with the occasional teething troubles, you'll find the support from existing users and developers (both here and on the cvsgui group) is excellent. Bugs usually get fixed within a day or so (proper releases are less frequent - up to the whims of the developer, but you can usually push for a new release if it's an important fix for you). Personally I write on Win2k test in domain and standalone configurations. I also use cvsnt every day at work and any bugs tend to get reported to my by my colleagues quite rapidly. The core functionality is used every day by hundreds (possibly thousands... never counted) of users so is about as well tested as you can get. However bugs do slip through occasionally (such as when 57d triggered a crashing bug in pserver - I couldn't replicate or produce it here so had to rely on the help of others to nail it down). You can minimise any risk by waiting a few days after a release to see if anyone is complaining. (Of course if your installation is working already there is no need to upgrade - if it aint broke don't fix it). Tony ___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
Re: [Cvsnt] Another auth issue
On Tue, 18 Jun 2002 23:18:23 +0100, Gleske, Paul 6151 DUR HWS wrote: This seems to now be related the NTFS permission that I was trying to configure. I have to allow Everyone on the CVSROOT directory to get in. The group for the user already has permission on the directory, and I have even given permission to the user directly. I think that cvs is not doing the login correctly. The initial read of CVSROOT is done as 'System' - it needs to check things like the passwd file, the SystemAuth flag etc. before it can validate the user. Tony ___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
Re: [Cvsnt] Fw: Memory error in cvsnt build 57d
On Fri, 14 Jun 2002 10:38:32 + (UTC), [EMAIL PROTECTED] wrote: Has anynone else seen this problem? I would like to debug it, but I have not been able to reproduce it myself. It sounds a bit like the kind of error that can be caused by bad RAM - especially if one one person has ever seen it. Not sure what tcvs is, though, and whether that could be doing something strange. Maybe it really is trying to add a 4GB file to CVS? Tony ___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
[Cvsnt] Latest updates
CVSNT 1.11.1.3 Build 57e Mostly just bug fixes... The C library putenv function is broken again (thanks MS) so I've bypassed it completely in this release, which should sort out some of the odd crashes that people have been seeing. Workaround for 'cvs edit' trying to do server-side things from the client. Tony ___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
Re: [Cvsnt] problems with ssh
On Fri, 14 Jun 2002 16:56:37 +0100, Schneider, Rudolf wrote: hi all! i am using the ssh protocoll (Cygwin) to access to the server. if i add a new directory to the repository the directory on the server is containing the two files .owner (contains (null)) and .perms (contains default:rwc), and i am not able to add a file to this new directory. error: ERROR: Failed to get task list from server That error message doesn't occur in cvsnt. I can only assume it's an ssh error - perhaps if you ask on the cygwin forums someone might have heard of it? if i delete the two files .owner and .perms everthing works, or if i use the pserver protocoll everthing works too. There's probably something wrong with the permissions setup on the server. ssh setup isn't something I'm that familiar with, though, so maybe someone will know more about that. Tony ___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
Re: [Cvsnt] Editor problem using CVSNT 1.11.1.3 build 57d on Win2K
On Fri, 14 Jun 2002 16:08:18 +0100, Craig Barkhouse wrote: This is a multi-part message in MIME format. -- -- [ Picked text/plain from multipart/alternative ] Hi, I'm using the console-based VIM as my preferred editor (I've got cvs -e vim in my .cvsrc). I'm having the following problem on Win2K. When I do a commit, my editor launches, but there is no output at all to the console. Using blind keystrokes, I can quit my editor. That's when I see the error message Vim: Warning: Output is not to a terminal. (Not sure why I see the error message only then, not earlier.) It works here - I don't have vim but dos edit works fine, which presumably is similar. Nothing has changed on the client side that could cause an editor to fail - the execution method is slightly different but the end result is the same. Tony ___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
Re: [Cvsnt] Error: cannot specify domain as machine is not a domain member on Windows 2000 Server
On Fri, 14 Jun 2002 15:26:48 +0100, Dan Powell wrote: cvs [login aborted]: unrecognized auth response from cvs: cvs server: error 0 Invalid username - cannot specify domain as machine is not a domain member This means what it says, basically. Only domain members have the required access to query trusted domains for their users. This server is not the primary domain controller, and does belong to the domain. I am running the service as a domain user with create a token permission. Either the server isn't actually part of the domain, or the 'use local users instead of domain' is checked. Tony ___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
RE: [Cvsnt] Possible bug in cvs export -f?
On Thu, 13 Jun 2002 18:36:47 +0100, Czarnowski, Aric wrote: The documentation states: snip So it looks as if it's behaving correctly. Thanks Tony for pointing me in the right direction so quickly. What documentation did you find this in? All I could find in the Cederqvist on cvshome.org under export is: It's in the cvs.pdf file, which is mostly the same as the contents of cvshome.org.. see http://www.cvshome.org/docs/manual/cvs_16.html#SEC118 Tony ___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt
Re: [Cvsnt] .owner and .perms: how I can rid of them?
On Thu, 13 Jun 2002 21:50:32 +0100, Gabriele Carcassi wrote: Hi all, I use CVSNT as a command line to access a repository on a UNIX server via filesystem (:local:). Whenever I add a directory, I noticed that it also adds two other files: .owner and .perms. Does anybody now what they are, and how I can make them not appear again? They're part of the access control system on cvsnt. If you delete them it'll just put them back. It's not generally a good idea to use :local: over a network share - you get locking issues. It's especially not a good idea to use cvsnt locally to a repository shared with a standard CVS server - at the moment the repositories are compatible but that can (and will, probably) change. It's much better to use pserver or ssh/rsh to the unix server, which is guaranteed to remain compatible. Tony ___ Cvsnt mailing list [EMAIL PROTECTED] http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt