Re: parsing loginfo
The question sounds familiar, though I'm not sure since I don't pay much
attention to 'space in filename' problems. Maybe
http://mail.gnu.org/pipermail/info-cvs/2001-August/018969.html has what you
need.
dtayl
Rob Helmer wrote:
On Fri, Feb 15, 2002 at 03:58:03PM -0500, Larry Jones wrote:
Rob Helmer writes:
All the info I want is actually in %s. Anyone have examples where
they have successfully parsed this data if it contains spaces
in the subdirs leading to the filename and/or the filename itself?
Just one more reason why rational people don't use filenames with spaces
:)
in them. ;-) If you haven't also used commas in your filenames, you
can use %{sV} or %{sv} and look for the comma between the filename and
the version number to determine the end of the filename.
Thanks for the info, I'm already doing that part.
I've noticed the main problem though is that I can't tell the
directory apart from the filename, see this example of %s output
where the dir is named blah test and the file is named space file.txt :
sandbox/blah test space file.txt
Nasty. I don't really see a way around that.
What I'd like to do is this : when a file is committed, I want
to get the full path w/ filename, current revision, previous revision
and run a diff to put up on a webpage.
This works great except for a space issue like the one above.
I don't think I can get people to stop using spaces in filenames
for this particular repository, but I guess I could workaround by
not allowing spaces in dir names and in my loginfo parsing script
split on the first space I see. Ack.
Thanks,
Rob Helmer
Namodn
___
Info-cvs mailing list
[EMAIL PROTECTED]
http://mail.gnu.org/mailman/listinfo/info-cvs
___
Info-cvs mailing list
[EMAIL PROTECTED]
http://mail.gnu.org/mailman/listinfo/info-cvs
Re: cvs [commit aborted]:
Larry, thanks for your patience. Like I said: When I SSH to the CVS server I see many CVS SERVER processes on the box (via ps -ef ). So, clearly my server processes are not leaving core files (since they are not dieing). The /tmp/cvs-serv* directories do exist, and they're empty. It appears as if the both the server and client process is just sitting there. When I run the client I can use the -t option to show a client side trace. I was wondering if there was someway to turn on some server side verbose logging to help resolve my problems. From your previous reply it sounds as if my only choice at this time is run the developement version of the server. Is there anything I need to do to see the recently added log messages? Where do the messages go? Standard out? - Dave [EMAIL PROTECTED] (Larry Jones) wrote in message news:[EMAIL PROTECTED]... David Hoag writes: Im guessing its related to a slow internet connection (110kb - faster than dialup) or something about my provider. Any ideas where to look? Like I said: Look in the server's TempDir (usually /tmp) for leftover cvs-serv* directories and core files -- if there aren't any, the server isn't crashing. That will at least let us know whether the server is crashing or simply ending without sending any error message. If the server isn't crashing, you might want to try running the current development version of CVS on the server -- I recently made a change to send any pending output to the client before shutting down the server. -Larry Jones It's not denial. I'm just very selective about the reality I accept. -- Calvin ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: Tagging
I would keep one repository, for starters, and use tags to identify releases. There are many different ways of managing them, and I would refer you to http://www.enteract.com/~bradapp/acme/branching/ and Karl Fogel's book (relevant chapters of which can be found at http://cvsbook.red-bean.com/, particularly the section on working with branches.) One approach would be to do minor changes directly on the mainline but to have branches for major rewrites, remembering to merge early and often and labeling as necessary. Dave -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] Sent: Thursday, February 14, 2002 9:21 AM To: [EMAIL PROTECTED] Subject: Tagging Question: My production system is set to be released/baselined at the annual level. However, changes after the annual release are frequent. Some of these changes are minor(cosmetic) but some may require deletion or addition of code or even whole re-writes of code. Am I correct in thinking that I should have the developers use tags for minor changes(excluding cosmetic changes) and then use rtag for my release version? This scares me..typically the way we have done things in the past was to have multiple copies for each year(i.e. dir for 2000, dir for 2001,etc). Should I just set up a main repo and rtag it for each year instead of having creating multiple repos for every year? ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: Automatting cvs retrieval using Windows NT/2000 at command
This is based on my dealing with AT under NT 4, so caveat emptor... Check the permission of the user ID under which the AT scheduler is running; if it is system (presuming you repo is across the wire) then change it to something else as the system user doesn't have network access. Note future password changes to this user id must be changed manually. To test your environment under AT you can use unix style redirection: at 10:00 c:\path\batfile c:\path\file.log 21 where c:\path\batfile.bat contains: set Note you cannot put dos primatives (built-in commands) directly in the AT command. -Original Message- From: Michael Stopper [mailto:[EMAIL PROTECTED]] Sent: Thursday, February 14, 2002 7:57 AM To: [EMAIL PROTECTED] Subject: Automatting cvs retrieval using Windows NT/2000 at command A search of existing threads didn't reveal anything useful on this so here we go: I've got my build process completely automated via Ant (including checkout, tagging, etc.) and can kick that off at will from a command prompt on NT/2000. However, if I put the call to start ant in a batch file and have the at scheduler kick it off, I get the following back from cvs on the client: cvs [checkout aborted]: could not find out home directory Which home directory is it looking for and why doesn't it need it when the batch file is running normally under DOS?? Any help would be appreciated. Thanks. -- - Michael Stopper Senior Principal / SPS Technical Architect American Management Systems, Inc. 4114 Legato Road Fairfax, Virginia 22033 703.227.6646 Office 703.227.4696 FAX mailto:[EMAIL PROTECTED] ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: cvs [commit aborted]:
But you also said, many people use this CVS server. Do the server processes you see come and go (which implies that they belong to other users) or do some of them stay around forever? The processes are owned by my user id. Started when I initiate a cvs update(or commit). They linger until I manually 'kill' them. The traffic on the CVS server is not 'a lot' by most standards. About 5 people using it daily from various locations. Sorry for the confusion. I'm confused -- originally you said the client was ending with connection reset by peer and end of file from server error messages, now it appears that you're saying the client is hanging. Is this the same problem or a different one? Broken connections and hung connections are usually not caused by the same things. The original error is the only error I've received. But what is not visible by the original post is the fact that the client doesn't die immediately ( I would guess there's a timeout value triggering the broken pipe). Again, I apologize if I confused the issue. The trace is unified -- some messages come from the client, others come from the server. I've built the latest cvs from a ccvs checkout. I'll try again tonight and report if I have any additional information. Thanks again, - Dave ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: using files with .xls and .doc in CVS
Eric,
I floated a suggestion a while back that -kb should be a bit
stickier; basically, it should trump sources of -k options that
would otherwise be higher priority. The idea seemed to meet with
general (though not universal) approval, if I recall, but
nobody's written the code. Neither have I, of course, so this
isn't a gripe, just an observation -- and perhaps a gentle nudge,
should the idea happen to appeal to you :-)
Yep - you might be referring to the thread Proposal to fix CVS
binary file implementation from December, 2000. I included
somewhere in that flame war a patch which I've used to make -kb
sticky, even in the presence of -kk on checkout or update.
A link to the archived email that includes the patch is:
http://mail.gnu.org/pipermail/info-cvs/2000-December/011718.html
but I noticed that every instance of = got transformed to =3D, so
here's a cut-and-paste of the correct patch:
*** cvs-1.11/vers_ts.c Thu Dec 21 18:46:35 2000
--- vers_ts.c Fri Dec 22 12:00:45 2000
***
*** 108,115
--- 108,130
* -k options specified on the command line override (and overwrite)
* options stored in the entries file
*/
+ /* DLM start: Patch to disallow override of -kb from archive
specification */
if (options *options != '\0')
+ {
vers_ts-options = xstrdup (options);
+ if (finfo-rcs != NULL)
+ {
+ char *rcsexpand = RCS_getexpand (finfo-rcs);
+ if ((rcsexpand != NULL) (rcsexpand[0]=='b'))
+ {
+ if (vers_ts-options != NULL)
+ free (vers_ts-options);
+ vers_ts-options = xmalloc (strlen (rcsexpand) + 3);
+ strcpy (vers_ts-options, -kb);
+ }
+ }
+ }
+ /* DLM end: Patch to disallow override of -kb from archive specification
*/
else if (!vers_ts-options || *vers_ts-options == '\0')
{
if (finfo-rcs != NULL)
___
Info-cvs mailing list
[EMAIL PROTECTED]
http://mail.gnu.org/mailman/listinfo/info-cvs
Re: cvs [commit aborted]:
[EMAIL PROTECTED] (Larry Jones) wrote in message news:[EMAIL PROTECTED]... David Hoag writes: $ cvs -version Concurrent Versions System (CVS) 1.11 (client/server) ... more version info $ ssh -l dhoag 192.168.0.1 cvs -version Concurrent Versions System (CVS) 1.11 (client/server) ... more version info With CVS 1.11, you can just do cvs version (that's a command, not an option, so no -) to get both the client and server version information in one shot. Read from remote host 192.168.0.1: Connection reset by peer cvs [commit aborted]: end of file from server (consult above messages if any) That indicates that the server is closing the connection. That could be because it is crashing, or it could be because you've got some kind of access control software (e.g., tcpd) that's denying access to the server. Look in the server's TempDir (usually /tmp) for leftover cvs-serv* directories and core files -- if there aren't any, the server isn't crashing. -Larry Jones Philistines. -- Calvin Im guessing its related to a slow internet connection (110kb - faster than dialup) or something about my provider. Any ideas where to look? ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: using head revision in branch after add on branch
-Original Message- From: C. Wienberg [mailto:[EMAIL PROTECTED]] Sent: Wednesday, February 13, 2002 4:07 AM To: [EMAIL PROTECTED] Subject: Re: using head revision in branch after add on branch Hi Greg, You can only check them in into _one_ module. Of course -- that's about the only way you'll ever make any kind of change management system make sense! I think there is a misunderstanding: I do not intend to check one file into several modules! I just want to check in different files, that are in a common _folder_, into different repositories! Nope; not in CVS. It would take a major redesign of CVS to make this possible. If we have one module for common stuff and one with specialties, that go into the same directory, we have a problem. Huh? Nope. Not a problem. The source modules are each checked out into their own working directories. Your build system combines them into the common target directory. Well, so far we have very successfully been working using branches, without the need to have an additional build system, like ANT. Also, I do not think that what we do is an abuse of the branch concept, or that my request is perverted ;-). I wouldn't call the desire perverted, but CVS alone isn't going to do what you want, and probably never will. This is unfortunate, since the easiest way to test things is often to change them in the working directory rather than to change them elsewhere and run make/ANT/whatever, but it is a characteristic of CVS. All files managed by CVS that are checked out into one directory must be in the same directory in the same repository. If you don't like this, your choices are basically: 1. Check the files out into separate directories and use some sort of build process to put them in the working directories. 2. Bash your repository until all the files are in the same repository directories (which obviously may not be possible, such as if you're using other people's repositories). 3. Bash your working directories so that they work the way the repository is set up. 4. Use another version control system (which, if you're using other people's repositories, may also not be possible). 5. Modify CVS extensively on your own (bearing in mind that your changes are almost certainly not going to become part of mainstream CVS). There is no guarantee that any of these options are attractive, but I don't see any alternatives. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: Hierarchical team integration in CVS
-Original Message- From: Earl Williams [mailto:[EMAIL PROTECTED]] Pierre Asselin [EMAIL PROTECTED] wrote in message [EMAIL PROTECTED]">news:[EMAIL PROTECTED]... Adapt your tree structure. Do *not* mix files from teams A and B in a single directory, CVS won't handle that well. Hi Pierre, Thanks very much for your detailed suggestions. It seems that in order to use CVS, the directory structure of my project must be driven by short-term, division-of-labor-during-development concerns (Team A, Team B, Shared) rather than by the conceptual structure of the classes in my project. Too bad. I guess I need to learn more about Software Configuration Management. To be specific, that is the case if you are determined to use CVS and no other tools for software configuration management. If you have been taught that no more than one tool should ever be necessary, then indeed you need to learn more. The Unix philosophy is to create tools that do one thing very well, and to combine those tools as needed. In accordance with that, CVS does some things very well and some other things not at all. This is sometimes inconvenient, but such things as working with the exact specific tools somebody else thought was exactly what you needed (or what you were going to get, anyway) can also be inconvenient. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: HOW TO ---- UNSUBSCRIBED
I'm also interested in unsubscribing. First it is not clear how to do this from the URL pointed to at the end of the e-mail. But also to unsubscribe I need my password which I have forgotten. It is apparently is mailed to me once a month. I think I must have missed getting it. Does anyone know what the subject line is on such a mail - or who it looks like it comes from? Okay, this isn't rocket science: http://mail.gnu.org/mailman/listinfo/info-cvs Then down at the bottom where it says To change your subscription (set options like digest and delivery modes, get a reminder of your password, or unsubscribe from Info-cvs), enter your subscription email address: (note: it says unsubscribe), you enter your subscription email address. Then, for anyone who's forgotten their password, note the large letters Forgotten Your Password? Then Click this button to have your password emailed to your list delivery address. and the actual button right there Email My Password To Me You guessed it; you click that button, and it will email your password to you. Then on that page there's the Unsubscribing from Info-cvs To unsubscribe, enter your password and hit the button. (If you've lost your password, see just below to have it emailed to you.) So even if the big letters didn't attract your attention, the smaller ones there in the directions should. So now that you've gotten your password emailed to you, you put it in that little box (Password:) and click the button (Unsubscribe) And voila. PLEASE READ THINGS LIKE THIS. The text is all right there in front of you, and you're just wasting a lot of people's time to read your message and quote the page that's sitting on your screen. Perhaps you should file a bug with mailman and request that they add a big blinking UNSUBSCRIBE animated gif to attract people's attention, since reading is too difficult. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Readonly
are you by chance using WinCVS? Or perhaps someone has set up a .cvsrc file somewhere? vvor wrote: Help. I have tried for several eons to turn off read-only permissions. CVSREAD is set to no. there don't seem to be locks or watches. I tried cvs -w co modulename. I tried cvs -w -f co modulename. I tried cvs watch off. I tried cvs watch remove. There are no lock files. This started suddenly, and now everyone has to do attrib -R /S at their cvsroots to work, because everyone is used to having read-write on everything. Everybody trusts everyone here. Also, importing modules naturally doesn't add modules to the modules file. I'm sure there is a good reason for that. Help. Please. Vora ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
cvs [commit aborted]:
I can not commit - Here's an example session: CVS server is on solaris, CVS client is on RedHat, Win98, or Win2k. I am behind a Nat router when accessing the cvs server (I've change the IP of the CVS server to protect the innocent). $ cvs -version Concurrent Versions System (CVS) 1.11 (client/server) ... more version info $ ssh -l dhoag 192.168.0.1 cvs -version Concurrent Versions System (CVS) 1.11 (client/server) ... more version info $ cvs -t -z 9 commit -m Updates to property holders BrokerProperty*.java cvs commit: notice: main loop with [EMAIL PROTECTED]:/space/cvsroot - Starting server: ssh 192.168.0.1 -l dhoag cvs server - Sending file `BrokerPropertyDetail.java' to server - Sending file `BrokerPropertySource.java' to server Read from remote host 192.168.0.1: Connection reset by peer cvs [commit aborted]: end of file from server (consult above messages if any) $ cvs -z 9 status BrokerPropertySource.java Read from remote host 192.168.0.1: Connection reset by peer cvs [status aborted]: end of file from server (consult above messages if any) Sometimes CVS update/status works. Cvs checkout always works. Any suggestions? ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: cvs advice
First, get Karl Fogel's excellent 'Open Source Development with CVS'. Create a test repository with some dummy modules and do some experimentation. Concentrate on the concepts and the syntax will come. It's really not too hard once you get the hang of it. Dave -Original Message- From: Alvarez Lorencio, Maria Jesus [mailto:[EMAIL PROTECTED]] Sent: Tuesday, February 05, 2002 7:18 AM To: [EMAIL PROTECTED] Subject: cvs advice Hi all, I´m totally new to CVS and I wonder if someone who is not a programmer, engineer... with a little knowledge about RCS should be able to manage this tool. I´m talking about me, obviously. I read the info-cvs e-mails every day, the cederqvist and everything that I find but I´m not very sure. The program has not be installed yet (don´t know which version is going to run) but the day is arriving...and I´m terrified. I´m suppose to keep the files in the repository (checkin, checkout), do backups and that´s all. I hope so Any advice to start? Thanks a lot ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Problem with recursive cvs edit command
There are a lot of weird behaviors with ampersand modules because they aren't fully implemented. This appears to be one of them. On Wed, 6 Feb 2002, Dan wrote: OK, I'm new to cvs, and think I've got everything working except for one peculiarity. I've got a modules file that looks like: moduleA -d moduleA path1/moduleA moduleB -d moduleB path1/moduleB AllModules moduleA moduleB I then checkout AllModules and end up with a directory tree locally like: AllModules moduleA moduleB which is what I would expect. I can commit, get status, get logging, etc. at any point in the tree and everything works just fine. If however I do a 'cvs edit' on moduleA, and moduleA has sub-directories (let's say suba for example - which has a file in it called filea), then I get the following error: cvs -z9 edit (in directory C:\AllModules\moduleA) cvs [edit aborted]: cannot find suba/filea: No such file or directory *CVS exited normally with code 1* If I do a 'cvs edit' on 'suba' directly however it works fine, and all the files in suba are marked edit: cvs -z9 edit (in directory C:\AllModules\moduleA\suba\) *CVS exited normally with code 0* Any help for this novice cvs user would be greatly appreciated. Dan ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: cvs [commit aborted]:
More information: When I SSH to the CVS server I see many CVS SERVER processes on the box (via ps -ef ). This problem appears to happen only from one set of IPs. While many people use this CVS server, I'm the only one experiencing problems ( and from only 1 location ). I have no problems with different machines on a couple of different networks. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: how to change a rev number
-Original Message- From: Shane McDaniel [mailto:[EMAIL PROTECTED]] I accidently checked in a file with the wrong rev number. How do I go about changing the rev number in the DB? Is there an easier way than checking the rev out, removing it from the DB and putting it back with the new rev? The right answer is to stop thinking about the rev number and just apply a tag when you want something you can refer to. Leave the revision numbers to CVS. point taken. but then what is the use of having a hierchal rev number if tags are what one should use? wouldn't cvs just use an incrementing number ie 1,2,3,4 instead of 1.0.1,1.0.2,etc.. The reason is historical. Originally, CVS was a set of wrapper scripts over RCS, and CVS continues to use the RCS format of save files. (There's advantages there.) One intended successor to CVS, Subversion (http://subversion.tigris.org/index.html), does use sequential numbers for its revisions. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
username in CVSROOT going unheeded in winCVS
I'm using the .rhost method for CVS with winCVS (W2K). Problem: The username seems to be unheeded, i.e. when I examine the log portion of the modified file it indicates that ipmds* (see below) was the author of the change rather than davec%. My CVSROOT within the winCVS tool is shown as :ext:davec@pmdbprod:/home/ipmds/cvsroot I've also tried remove the :ext: and there is no difference. Under unix (HP/UX) it seems to work correctly i.e. the authorship tracks the username specifed in the CVSROOT environment variable. *ipmds: is the username under which the CVS server is running i.e. the login shell %davec: is my user id under W2K Does any have any idea how I can rectify this? --@@ ~ DavidC 'The Biggest Game In Town' Finally, America will begin to see the staggering wealth our own city, county, state, and federal governments hold in hidden and secret accounts and assets. If these assests - that the AMERICAN PEOPLE own - can be liberated from government agencies, we can see a virtual end to property and income tax. Sound impossible? Then you haven't heard Walter Burien exposing the Comprehensive Annual Financial Report scam. http://www.wces.org/html_files/burien.html ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
branch deleted but still needed
We had a developer accidentally delete a branch from our repository today. Using WinCVS, we're able to produce a graph that shows the branch apparently still exists, but is not named. Is there a way to restore the branch (i.e., restoring the name) so we can continue to reference it as before? Dave ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: cvs (or something!) on very large scales with non-source code objects
I think CVS would probably do quite well for the system you're describing. You're already doing a primitive form of versioning when you rename the files to FILE..mm.dd.hh.mm. -Original Message- From: Nigel Kerr [mailto:[EMAIL PROTECTED]] Sent: Friday, February 01, 2002 9:25 AM To: [EMAIL PROTECTED] Subject: cvs (or something!) on very large scales with non-source code objects good folk, i ask this forum because i'm not at all sure where start looking for ideas on how to address my problems. cvs may not be the right tool for what i have, but any ideas or suggestions or redirections to other fora are welcome and desired. i have several million objects (very large scales): roughly half of them are bitonal TIFF files, scanned page images of printed material; the other half are OCR'd text of those same TIFF files. there are a relatively small number of other kinds of files: metadata about chunks of these data, and auxilliary images of parts of some of the pages. right now the top level chunks of this corpus number about 3,000, with sub-chunks inside those top-level chunks. at any moment, it might be discovered that there is an error or problem with any of these objects, that will need to be fixed: the TIFF file might be bad/corrupt/unclear the ocr'd text might be bad/corrupt/unclear the metadata might be found to be wrong the auxilliary images might be bad/corrupt/unclear we might make a change to a small number of things at a time, we might also make a batch change to thousands of things at a time. back when we had less than 500 top-level chunks, our life was relatively easy: we had a home-grown edit-history-type system that basically: moved the old file FILE to FILE..mm.dd.hh.mm moved the new version of FILE into place wrote in a date-stamped log file a message meaning i changed this!, where the message phrased differently depending on what got changed. used the doughty mirror perl script on our different machines to get the changed data from the master to the slave machines. we're still using that system. we get about 400,000 new items a month in between 30-50 new top-level chunks (a top-level varies in size considerably). the increases in size of our corpus will never slow down. our stated *goals* for using this system are two-fold: a method for communicating from the master to the slave machines about what has changed, and what they should try to update. a record of what all has changed ever, so that if we had to start from original source media (the cd-roms the data arrive to us on), we could, and only update what needed updating. i don't have much problem with the first goal: we need some communication method from master to slave. i am increasingly nervous about the second goal as we get larger and larger, and am looking for other ways to address or consider that problem. it might be that we: give up on record of what all has changed ever, and try to go for record of what all has changed since the last time we had a complete checkpoint of our corpus, and keep using our change system, and give up on the restore from original media idea. use a version control system that can handle millions of things (which would be?!) changing, and the master-to-slave transport of changes efficiently. keep going about things as we have, and just hope we never have to restore from scratch. something else? anyone here approached this kind of problem, know someone who has, or have any ideas about it? people/places i can seek advice from? anything is appreciated, thank you. cheers, nigel kerr [EMAIL PROTECTED] ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: info needed for 3 way merging
JAKramer writes: I have used ediff in the past for merging a small set of files and also for resolving conflicts from a cvs merge. Here's what I'm really after. The project I'm working on consists of a total of 2,700 cpp, h, and idl files. We also frequently have 2 or 3 development branches going on concurrently. In about a week I'll start the merge process for the branch that my team has been working on. Presently with a test cvs merge I've found that there are 94 files with merge conflicts. Unless I find a better way to handle this, I'll end up using xemacs/ediff to perform a three way merge on each of the conflicting files after a cvs merge is completed. This involves identifying the conflicting file (easy), identifying each of the 3 files for the merge in xemacs (tedious), and resolving the actual conflicts (downright painful). There's no getting around the final step. However, life would be much better if the previous two steps could be automated. E.g. perform the merge and have xemacs cycle through each of the conflicting files so that the conflicts could be resolved. This would at least make a difficult situation less painful. This is why I'm seeking more information on running emacs in batch mode. Any URLs to that effect or other suggestions would be greatly appreciated! I don't see how you could get around doing the last step interactively (if it could've been done in batch, then CVS would do it). Therefore, I don't see how you could run emacs in batch to do this process. I haven't used it, but doesn't run-ediff-from-cvs-buffer in the ediff package provide what you want? -- David Mastersondmaster AT synopsys DOT com Sr. RD Engineer Synopsys, Inc. Software Engineering Sunnyvale, CA ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: reserved checkout
pardon my abject ignorance on all of this... But could you use uuencode and uudecode? The deltas may be *HUGE* but it might be worth a lookit. --@@ ~ DavidC The real truth of the matter is, and you and I know, that a financial element in the large centers has owned the government of the U.S. since the days of Andrew Jackson. History depicts Andrew Jackson as the last truly honorable and incorruptible American president. ~ President Franklin Delano Roosevelt, November 23, 1933 in a letter to Colonel Edward Mandell House -Original Message- From: Matthias Kienle [mailto:[EMAIL PROTECTED]] Sent: Friday, February 01, 2002 12:04 AM To: Greg A. Woods Cc: [EMAIL PROTECTED] Subject: Re: reserved checkout [ On Thursday, January 31, 2002 at 10:34:52 (+0100), Matthias Kienle wrote: ] Subject: reserved checkout What do I wrong? Don't use reserved checkouts with CVS -- CVS is the Concurrent Versions System! I know cvs is a concurrent version system, but a new requirement comes up in our department to use in future only one version system. The other version system administrated binary files like word, excel and powerpoint. CVS (and other version systems too) can not merge binary files after more than one edit from several editors of the same version. The other version system can do reserved checkouts but it runs on a windows server and has no command line. In a first test we used the editors and watcher function of cvs, but the test shows us it is not enough. I must often merge by hand two versions and I can say I hate it. I am a linuxer and I would hate it to explain my manager that cvs does not support reserved checkouts for binary files. I control all web pages, server configuartions and the total source code in a cvs repository. I love my command line for cvs and I love my CVS. I know I am sitting between two chairs. Matthias ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: DIfferent workspace directories than repository...
-Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] Sent: Tuesday, January 29, 2002 10:50 PM To: [EMAIL PROTECTED] Subject: DIfferent workspace directories than repository... Hope I can get some help here... I've used CVS for awhile now, but I have never really had to deal with different workspace locations for repository directories. But I actually need this now. Here is an example of the repository layout I need : + myproject | + bin | + lib | + html | + src | + docs Um, why do you need this specific layout in the repository? It seems to me that the structure of the repository should be determined by the structure of what must be checked out. (Not to mention that I'm always at least a bit suspicious about CVS-controlled directories called bin and lib.) Here is an example of the workspace layout I need : + /home/myhome | + mydirectory | | | + bin | | | + lib | + html | + src | + docs Is there anyway that I can map a CVS directory and assign it a specific workspace directory? I'm really hoping I can do this, otherwise I may have to dump CVS in favor of another SCM tool. And I really don't want to. But this one is a show stopper. CVS will use the identity map when checking out, but afterwards each directory has its own metadata (in the CVS subdirectory) so that it knows where it came from. This means that it will retain the mapping while you move the directories around on your own. Therefore, if you do something like cvs co bin; mkdir mydirectory; mv bin mydirectory/ you will still be able to do updates and checkins from mydirectory/bin. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: DIfferent workspace directories than repository...
-Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] Sent: Wednesday, January 30, 2002 10:19 AM To: [EMAIL PROTECTED]; [EMAIL PROTECTED] Subject: RE: DIfferent workspace directories than repository... Um, why do you need this specific layout in the repository? It seems to me that the structure of the repository should be determined by the structure of what must be checked out. Some would disagree with you on that... myself included. I was hoping that I could get CVS to treat a directory as a true project/first-class object. Not just as a directory. Nope; CVS does not treat directories as first-class objects, and is never likely to. Doing that would require a thorough redesign, and would not be likely to be called CVS afterwards. Check out the Tigris project (www.tigris.org) which is addressing these kinds of issues that CVS apparently doesn't meet. Subversion (the SCM project at Tigris) is indeed intended as a thorough redesign of CVS, taking advantage of years of experience with CVS and its quirks, and is intended as a CVS replacement. Last I looked, it looked promising, but not something I have immediate use for. Well, that answers my question. It can be done, but with manual kludges involved. I was hoping to avoid such things. I do lots of things with Perl to make CVS easier to use. I can get lots of the quirks papered over that way. Overall, I find that CVS's reliability, support of branching and concurrent development, and cost make it an extremely useful tool, although far from ideal. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: URGENT: Initial revision number for CVS.
-Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] Sent: Monday, January 28, 2002 8:46 AM To: [EMAIL PROTECTED] Subject: Re: URGENT: Initial revision number for CVS. I have a similar problem. At our company we have Sources that we want to put under CVS control. Some of the sources are already RCS controlled, so copying the *,v file directly into the repository seems to be the right approach for us. The history is preserved and we get the revision numbers that have been created by RCS (which all have the syntax x.y, let's call this depth 2). OK, no problem with that. No, importing the other, non RCS-controlled sources has the same effect that you describe. All files get the revision number 1.1.1.1., so the depth here is 4. OK, no problem with that. We really do not need the depth 4 revision numbers and want a default revision number depth of 2. I have searched all the CVS manuals, Google and other written resources, but found no answer to this question. Correction: if you're doing a cvs import you do need the depth 4 revision numbers, and that's not just because CVS does it that way. In imports, CVS maintains its own import branch, and since it's a branch it needs depth 4. Depth 2 revision numbers are only on the head branch, and in an import situation are for your own changes. Let me be more specific. It may happen that you import a codebase and maintain local changes to it. We do it with Gnats. In that case, CVS has to have a way of keeping track of the changes in the imported source versus the local changes. CVS keeps track of such things with branches, and so you need branches, and so you need the revision numbers you're complaining about. If you aren't importing somebody else's source, but you're bringing in source that will be developed from the CVS system, then you can use cvs add instead of cvs import. Has anybody an answer that is other than don't care about the CVS revision numbers, use your own scheme with the use of tags ? What other answer is needed or accurate? ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: merge issue
Schwenk, Jeanie wrote: I have read and reread the documentation. I must be missing something. Even when he follows my directions, the merge is not what he expects. He does not want to ask the list for help because (and this is a quote) it is like asking for directions. I often ask for directions, but have never gotten the response Read The F- Map!. ;-) dtayl ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
listing available modules from a server
Sorry if this is a FAQ (which I couldn't find in the FAQ): How can I get a list of modules available for checkout from a cvs server? Is there a cvs command for that? TIA -- PANOPE: Déjà, de sa présence avec honte chassée, Dans la profonde mer OEnone s'est lancée. (Phèdre, J-B Racine, acte 5, scène 5) ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
race condition?
One of our developers pointed out that cvs commit foo bar followed by cvs rtag I_just_uploaded_this foo bar contains a race condition. Someone else might have committed between the two tags. Assuming we want to use tags as a way of assigning blame, is there a way round this? david hj ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: race condition?
-Original Message- From: David Hugh-Jones [mailto:[EMAIL PROTECTED]] Sent: Tuesday, January 22, 2002 11:38 AM To: [EMAIL PROTECTED] Subject: race condition? One of our developers pointed out that cvs commit foo bar followed by cvs rtag I_just_uploaded_this foo bar contains a race condition. Someone else might have committed between the two tags. Assuming we want to use tags as a way of assigning blame, is there a way round this? I'd use cvs tag, not cvs rtag, since tag tags the revisions in the sandbox, not the repository. As long as your developer doesn't update between the commit and the tag, the tag should get exactly what he or she had then. If your developer wants to tag things that existed at that moment but which were not in the sandbox, you can always use rtag followed by cvs tag -F to force the tag to move to what's in the sandbox. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: Basic usage question
-Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] Sent: Sunday, January 20, 2002 9:01 PM To: [EMAIL PROTECTED] Subject: Re: Basic usage question In article [EMAIL PROTECTED], Wade Williams wrote: Myself and another programmer are working on a project. We're working mainly on different sections of the code. Day 1: I checkout the project Day 2: I make changes and commit them, and then continue working on my working copy. Day 3: Programmer B makes changes and commits them If B does a module-level commit, then B's commit attept should fail with the ``up to date check failed'' diagnostic on files that you commited on day 2, assuming that B did a checkout before day 2, and has not updated since then. Thus B is forced to update to incorporate your changes via cvs update, resolve any conflicts and try the commit again. Right. I understood this to mean that there were cvs updates as needed, although the original poster may not have been aware of these details. Basically, you can only commit changes if your files have been updated to the head or branch tip or something, and if changed have been checked in you'll have to run an update. If there is a conflict, that'll be mentioned in the output and kept as the file status, and you won't be able to commit your changes until you've dealt with the conflicts somehow. (There was an exchange a few months ago about what does and should constitute dealing with the conflicts, which need not be repeated now.) It is possible to merge changes to create an incorrect source file without causing a conflict; one easy example would be one programmer removing a member function from a class, with another one introducing a new use of it. In practice, this turns out to be unimportant, since such problems are readily found. If people cheat by committing only the files that they modified, they can get around the up to date check. But that is a bad idea because the changes you make in one set of files can semantically conflict with changes in another set of files. Do you mean updating only the files that they modified? If so, that's a bad idea as you say, but it seems to be somewhat self-correcting. People get themselves into trouble that way, and soon find that it's easier in the long run to update everything. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: About playing with sticky dates, the art with no cvs dir
I think the only way is to do it bu hand. I explain : Even If I make a diff beetween : the result of merge of copy1 n cvs and copy2 the code substract issue to the diff beetween copy1+cvs and copy2 can correspond to : the code wich was added when copy1 merge to cvs or ths code which was surely to remove Or we can t do it automatically only a human can do, a person wich now about the project. I can start form copy1 not marged with cvs and copy2 then make a diff. But copy2 has conflit that copy1 had which had been resolve by a human to, perhaps it easy like that because these conflicts were already know. a+. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
howto rename a dir ?
Hi, The first import of a module countains a dir with a misspell, so I search a method for renaming a dir in the repo. The only way is to delete in the fs, cvs delete, rename in the os , and cvs add, cvs commit ? Is there a cvs admin way ? thanks. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
About playing with sticky dates, the art with no cvs dir
Hello, I have a problem, I ask for advise in order not do serious mistake. I make a working copy for a website for a designer, he works on it, but remove the CVS dir which are important information for the communication with cvs server. I know the date when i give it to him. So I apply the method some of them told me. do a cvs checkout with -D option (for the 1st december ...) copy into hos files make a cvs -nq update to see what files are added. add them with a cvs add ... do a cvs update -A verify all do a cvs commit. This work well, I realize to merge my work (in the cvs repo) with this copy wich we call copy1. But this designer works very fast, that I did nt manage to give him a working revision with CVS dir ... In fact I have a copy1 and copy2, before knowing how to do with the cvs repo, before knowing how all this system works, I stock them apart, and manage my personnal work with cvs. He developp from his copy1 a copy2. And now I ask me how to do to merge this final copy (for him) into the cvs ? Making a copy of the copy2 over an uptodate working directory will cause me to lost things, surely, I think of result of merge that i arrive to do with copy1 abd the cvs that are surely not into the copy2? I think about this method, told me if it was wrong : I copy the copy2 over a working dir wich have the date just before realize the merge beetween copy1 and the cvs repo. Them as usual cvs -nq update, cvs add ..., cvs update -A, n commit. Perhaps the date for the first copy can do the affair for copy2 (1str december), I have to cvs add everything, the addon from copy1 n the addon from copy2 ? ouf. Help me, i nearly finish, and I will fire the ears of the designer. thanks. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: howto change into bin file, file that are text in the repository
I found a way is to copy files in a temp dir, then remove in the fs cvs remove them, committing and cvs add them with the correct stiky option, it s very long i hope somrthing better, because I hav many files. I had binaries into the repository as text. It's web pictures. And I want them to be in binaries, furtunetly I keep them in my working dir. How can i tell the repo to consider them as bin... For not making the error again I make a cvswrapper but it was to late. thanks. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: howto change into bin file, file that are text in the repository
thanks again larry, yes I see this command cvs admin -kb ... on the web very useful for my pb. I notice a think I have several files scheduled to be added, i make a cvswrapper file and when I commit them they receive what is called the sticky option -kb, so no admin command to pass. thanks again. I had binaries into the repository as text. It's web pictures. And I want them to be in binaries, furtunetly I keep them in my working dir. How can i tell the repo to consider them as bin... For not making the error again I make a cvswrapper but it was to late. thanks. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
how to list all conflits into a huge dir struture
Hi, I have a project with many dir and with deep levels. I have resolve some conflcts that I can see just in default dir. I make a commit and he finds me other that I didn't see. The commit took a lot of time, so I want to no longer forget eventual conflits n correct them before launching commit. SO if u have command to make this it will be great. thanks. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
howto change into bin file, file that are text in the repository
I had binaries into the repository as text. It's web pictures. And I want them to be in binaries, furtunetly I keep them in my working dir. How can i tell the repo to consider them as bin... For not making the error again I make a cvswrapper but it was to late. thanks. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: how to list all conflits into a huge dir struture
Yes I have it wincvs, I see the flat mode, i see a menu called macro with some useful fonctions. But where hav u find the list of conflicts ? It s a great tools with a lot of features. thanks. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Accidental -kb option used when adding a text file
I have the reversed problems, I had binaries into the repository as text. It's web pictures. And I want them to be in binaries, furtunetly I keep them in my working dir. How can i tell the repo to consider them as bin... For not making the error again I make a cvswrapper but it was to late. thanks. Duncan Sommerville writes: cvs admin -kv myfile.txt That should be -kkv. Typically you need to do an 'update' afterwards for the change to be reflected in the workspace, however in my version of CVS (1.10) changing from a -kb to -kv needed me to redo the 'checkout' operation before the change was evident in the workspace... A simple update isn't sufficient because CVS doesn't distinguish between -k options that came from the repository and -k options that came from the command line. You either need to do update -A or update -kkv. -Larry Jones I kind of resent the manufacturer's implicit assumption that this would amuse me. -- Calvin ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Repository directory/file restructuring (delete/move)
I've been reading up on the various methods to move files and
directories around in the repository, and I haven't yet found a workable
solution to what I'm trying to achieve with one of my public projects.
I have a repository which contains a structure similar to the following:
-TOPDIR
|
|-tools
| `-Perl5
| `-Python
| `-Java
|
|-libsock
|
|-libcc
|
|-man
|
|-doc
|
|-include
|
` *.c README, ChangeLog, etc. in `pwd`
...and so on. I am about to make a pretty monumental release, and would
like to restructure this a bit to resemble something like the following:
-TOPDIR
|
|-src
| |-include
| |
| |-pix
| |
| |-prc (was previously two files in `pwd`)
| |
| `-*.c
|
|-libpisock (was previously libsock)
|
|-libpisock++ (was previously libcc)
|
|-doc (which now contains README, ChangeLog, etc.)
| |
| `-man
|
`-bindings
|
|-Perl5
|-Python
`-Java
Basically, a few renames at the top level (libcc becomes libsock++ and
libsock becomes libpisock), and a few tucks of the parent under another
directory ('man' gets tucked under 'doc', 'pix' and 'include' gets tucked
under the new 'libpisock', and so on).
Right now, HEAD contains about 40 tags of this project, from the first
import, through pre-releases, all the way to current. There is also a tag
off of this which is a very alpha development branch (call it an internal
fork). This must remain under HEAD, intact.
My question is, what is the best approach to doing this? I've heard some
people tell me that doing direct cvs-repository surgery is the only way,
by copying directories around, but that leaves two copies of versions,
one in HEAD, and one in the new directory. It also orphans off tags for
files that aren't in the new directory.
I've also heard people mention using 'cvs remove; cvs add;' for the
files I want to move around, but that doesn't preserve the history or the
versions.
Can someone give me a pointer to a good info file or manpagr or
cvshome.org reference on this? I've found the following to be
interesting, but don't quite cover my issues:
http://www.cs.utah.edu/dept/old/texinfo/cvs/cvs_15.html
http://www-es.fernuni-hagen.de/cgi-bin/info2html?(cvs)Outside
http://www-es.fernuni-hagen.de/cgi-bin/info2html?(cvs)Rename%20by%20copying
Any help anyone can give would be immensely appreciated. Direct replies
would be preferred, but I'll check back here often as well. The release
is only days away.
/d
___
Info-cvs mailing list
[EMAIL PROTECTED]
http://mail.gnu.org/mailman/listinfo/info-cvs
error checking out codes
Hi, I have an error which only occurs when using the pserver access method. We have a project which is stored in CVS, and it has been working fine for a month or so now, and so the problem has mysteriously appeared today. If we type cvs -d :pserver:[EMAIL PROTECTED]:/usr/local/cvs co package then it checks out about half of the code, and then terminates, with the error mesage Terminated with fatal signal 11 However, if I do it on the machine on which CVS is installed and I type cvs -d /usr/local/cvs co package then it works fine. Does anybody have any ideas about what is going on? Thanks in advance, David Churches. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: Validate html file before to commit.
-Original Message- From: Franck Marchand [mailto:[EMAIL PROTECTED]] Sent: Thursday, January 03, 2002 5:14 AM To: [EMAIL PROTECTED] Subject: Validate html file before to commit. Hi all Someone has a script to check a file before to commit it in cvs ? I would like to do something like that : CommitInfo file : DEFAULT $CVSROOT/validate-html I need a file which be able to read the new file and check it if it contains correct html tag. Do you have a validation program? You can use that, with some setup. One thing I ran into when I tried adding complicated validation to the CommitInfo file was that the time needed to commit slowed to what would be unacceptable limits if committing in multiple directories. If validating the file is going to take inconveniently long, you might want to use LogInfo to validate it later, and then send out some sort of email, such as a notification to the manager or a termination notice or something like that. Seriously, if you can't trust your employees to follow process, you should get new employees. Establish that employees validate first, then commit, and test if you like after the commit (because people do make mistakes), and that should be enough. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: About merging when no CVS dir are present (David D)
Yeah it's cvs update -A (the most recent version of files), here I have conflicts : One of them was about I have a production and a dev verson of a file. I was adviced to make a makefile : my file config.php no longer belong to cvs, I would hava a config.template n make a : make prod : sed /prod-dir/tpl-dir or make dev : ... I thinks it s a great idea, dont think ? Another question is what was doing the cvs update (without -A) i make ? Does it have any consequences ? thanks. But I update no conflits where found cool, But when I commit I hava a lot of messages like this : cvs commit : connot commit with sticky date for file 'toto/toto.php' I dont know what to do ? It looks like you need to pass the '-A' argument to the update command in the first instance - see the following section of the Cederqvist: http://www.cvshome.org/docs/manual/cvs_4.html#SEC53 I suspect that without the '-A' argument, your initial update is doing very little (hence no conflicts)... You may also find the following link useful: http://cvsbook.red-bean.com/cvsbook.html#update Kind Regards, Duncan. _ Do You Yahoo!? Get your free @yahoo.com address at http://mail.yahoo.com ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
About merging when no CVS dir are present
Hello, I make a post a few weeks ago, some of u respond me with : I work with a designer fot my site, I give him a version of my site not tagged I think. He remove all the cvs dire he s got on it an d gave me his job. 1. Checkout a copy of the module based on date/time using -D. Use the date that you gave him the files. 2. Copy his changed files over to the checked out copy. 3. Do a cvs update to bring in any changes that you have commited since you gave the files to the designer, resolve any conflicts and ensure everything merged smoothly. 4. cvs commit the changes. 5. If the designer is doing more work, give the designer the up to date files, including cvs directories, and tag the appropriate files in the repository in case he loses the cvs/ directories again. But I update no conflits where found cool, But when I commit I hava a lot of messages like this : cvs commit : connot commit with sticky date for file 'toto/toto.php' I dont know what to do ? thanks. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
viewing all tags
hi all Is there a simple way to view all the tags that have been created in a repository? David ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: 1.9MB attachments (was Re: Wincvs download)
-Original Message- From: Marcel van der Boom [mailto:[EMAIL PROTECTED]] Sent: Thursday, December 20, 2001 10:29 AM To: Scott McDermott Cc: [EMAIL PROTECTED] Subject: Re: 1.9MB attachments (was Re: Wincvs download) Configuring the mailing-list to limit or forbid attachments would be better IMHO. time for a blacklist entry :) On the other hand, it's sometimes useful to attach small things, like perl scripts. Unfortunately, some anti-virus filters seem to filter them out and complain to the list. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
diff after update
sorry if this is an obvious question, but after doing a CVS update, how do I diff files? Here's the scenario: I am working on a file. I haven't committed any changes. I do a general cvs update. I notice that that file shows M. How do I find out what the cvs update has changed, between my old uncommitted version, and the new version created by the update? David ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: Possible modifications to CVS.
-Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] Sent: Tuesday, December 18, 2001 12:45 PM To: [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Subject: Re: Possible modifications to CVS. [EMAIL PROTECTED] writes: Is it possible to checkout multiple files from different directories within the repository and put them into a single directory? Can CVS keep track of where each file came from and update the repository as necessary? No. CVS tracks metadata at the directory level, so all files in a single working directory have to come from the same repository directory. It doesn't appear to be a problem to have sandbox subdirectories that don't correspond to repository subdirectories, since the subdirectories maintain their own metadata, but it's not going to work for files. If this isn't possible, I would suggest it be added to CVS. What I would like to see is a way to check out files from anywhere in the repository, and from any repository, and be able to put the files anywhere I want in my working directory structure, and have CVS maintain the files. That would require completely redesigning and reimplementing CVS. Anybody know how common this ability is in such systems? Subversion, in particular? In any case, you can always manipulate files after you've checked them out, so you can put varied files into one directory. You can't check back in from that directory, though, which is something of a pain. So, I figure that that would be a nice capability that isn't going to happen in CVS. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: significance of -n option in rtag
Sangeetha Parthasarathy wrote: Hi, could any of you please explain to me the significance of '-n' option in rtag cvs -H command displays a usage message explaining all the options of command. (To get a list of all CVS help commands: cvs --help) For example cvs -H rtag displays: -n No execution of 'tag program'. For example what os the difference between cvs rtag VersionNumberOne SManagerand cvs rtag -n VersionNumberOne SManager The first command tags, the second doesn't. dtayl ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: significance of -n option in rtag
David Taylor wrote: For example what os the difference between cvs rtag VersionNumberOne SManagerand cvs rtag -n VersionNumberOne SManager The first command tags, the second doesn't. dtayl Wrong. That should be The first should execute taginfo, the second shouldn't. I say should because I've never used either rtag -n or taginfo, so I can't say with any certainty. sorry, dtayl ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: recursive add
I m not an expert of nix command but : What the purpose of xargs, it redistribute the output of the command before | to the commande after ? a+ Karl E. Jorgensen [EMAIL PROTECTED] wrote in message news:[EMAIL PROTECTED]... ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: could cvs make auto delete/add file judging from my revision ?
You'll get a warning message for files that are still present. You'll get a warning message for files that are already in CVS. warning means that the command doesnt apply to these files ? So I have to confirm it by an no ambiguous command for each of these files in case. like cvs remove t.php Benediction, like in my native language (french), babylon told me it exists with the same meaning in english, isn t it in american ? thanks Larry Jones [EMAIL PROTECTED] wrote in message news:[EMAIL PROTECTED]... David D writes: How to auto add these files ? cvs add *. You'll get a warning message for files that are already in CVS. How to auto delete files that are not present any more ? cvs rm. You'll get a warning message for files that are still present. I can understand that developers think it s a bad idea cause of mustakes it can bring, but in my case, having sevral files at lot of directories, it's a benedecition Benedecition? I think the word you're looking for is blessing. -Larry Jones Pitiful. Just pitiful. -- Calvin ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
could cvs make auto delete/add file judging from my revision ?
Euh ? indicates something added. How to auto add these files ? How to auto delete files that are not present any more ? Is there an option in the cvs commit ? I can understand that developers think it s a bad idea cause of mustakes it can bring, but in my case, having sevral files at lot of directories, it's a benedecition Thanks. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
cvsweb download problem
I have the latest cgi script from http://www.idaemons.org/~knu/cvsweb/ and am running it with cvs 1.11.1p1 and rcs 5.7 on sun solaris 2.8 with iplanet enterprise server 6.0SP1. Everything seems to work (including annotate), with the exception of clicking on a revision or download link. When accessing these links, the following message occurs: Error: Unexpected output from cvs co:, followed by the contents of the file in one long string, with no formatting. This is displayed in the browser window. How can I fix this? Thanks, Dave. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
How to merge with my cvs
Hello, I've got a problem taht I don t know how to resolve ! I work with a designer fot my site, I give him a version of my site not tagged I think. He remove all the cvs dire he s got on it an d gave me his job. I dont know how to merge difference know, because the repository dont know for wich file he has to compare ? How can i do know. thanks. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: ScanMail Message: To Recipient virus found and action taken.
-Original Message- From: System Attendant [mailto:[EMAIL PROTECTED]] Sent: Tuesday, December 11, 2001 8:49 AM To: '[EMAIL PROTECTED]' Subject: ScanMail Message: To Recipient virus found and action taken. ScanMail for Microsoft Exchange has detected virus-infected attachment(s). Sender = [EMAIL PROTECTED] Recipient(s) = [EMAIL PROTECTED];[EMAIL PROTECTED] Subject = Re: Unix/Windows CR/LF Problems Scanning Time = 12/11/2001 06:48:31 Action on virus found: The attachment commit_prep.pl matched file blocking settings. ScanMail has Deleted it. Warning to recipient. ScanMail detected a virus in an email attachment. [Thornley, David] This is not called for. The info-cvs list frequently has perl scripts sent around, and it is not reasonable to complain about it. It is reasonable to delete perl scripts from incoming mail if that is your desire. However, issuing email in HTML claiming to have found a virus is wrong and inconsiderate. (BTW, for whomever decided on the wording ScanMail detected a virus in an email attachment, I'd suggest consulting a legal professional. That looks to me like it could be taken to be libel.) ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: cvs co error: No space left on device
-Original Message- From: Vinh Cao [mailto:[EMAIL PROTECTED]] Sent: Friday, December 07, 2001 11:59 AM To: '[EMAIL PROTECTED]' Subject: cvs co error: No space left on device Hello All, Our Linux cvs server is working great for the pass two years. For the pass couple week it has the error: No space left on device when my users are checking out or do cvs update. CVS server has 1GB RAM and below is it partitions: FilesystemSize Used Avail Use% Mounted on /dev/sda9 2.0G 91M 1.8G 5% / /dev/sda1 23M 3.0k 22M 0% /boot /dev/sda5 28G 5.2G 21G 20% /cvs /dev/sda6 5.8G 213M 5.3G 4% /home /dev/sda10197M 11M 176M 6% /tmp /dev/sda7 2.0G 1.3G 517M 73% /usr I need your advice how to resolve this problem or what to do next. Thank you in advanced. What caused that where I work is that the main source tree, which is a very large module, got close to the size of /tmp. CVS builds a source tree in /tmp, and if there wasn't enough room in /tmp it didn't work. What is the checked-out size of your largest module? If it's getting close to the 176M size of /tmp (which seems plausible, given the 5.2G size of /cvs), you may just be running out of /tmp space. What we did was to increase the size of /tmp and put in a cron job to remove anything in /tmp that was over a couple of days old. Since your /tmp is only 6% full, it looks like you don't have the slow filling-up problem we had, but you may need to enlarge /tmp. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Is this possible? More Info
Don Weeks wrote: I need to pass this info over to another system, perhaps the problem reporting tool or the project management tool: PR# 321 was solved by adding foo.c version 3.2.1 and foo.h version 3.3.1. These now have the tag REL3.2 (This implies that if I checkout all the files using tag REL3.2, I can build the entire product.) Sometimes called change sets. If the problem reporting tool is Bugzilla, then you may find cvszilla to be useful: http://homepages.kcbbs.gen.nz/~tonyg It contains a Perl script, that can be run from loginfo, that passes loginfo (name of file, old and new version, log message) to the Bugzilla record for the bug cited in the log message. If you're not using Bugzilla, you still may want to view cvszilla as one example of interfacing CVS with a problem tracker. dtayl ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
[ANNOUNCE] CVSps - patchsets for cvs (new tool)
Hi all, I just wanted to announce the release of a new tool called CVSps. CVSps is a GPL'ed program, written it C, that generates 'patchset' like information out of a CVS repository. A quote from the README: -- quote CVSps is a program for generating 'patchset' information from a CVS repository. A patchset is defined as a set of changes made to a collection of files, and all committed at the same time (using a single 'cvs commit' command). This helps you see the big picture of the evolution of a cvs project. You can see the history of committed patchsets, restrict by author, date range, files affected, branches affected. The program can also generate a diff of a given patchset. It essentially gives you the equivalent of tagging before and after each commit. -- end quote CVSps is available at: http://www.cobite.com/cvsps/ At this time, version 1.2 is the latest version. It is quite stable and has a feature set derived from users requests. If you try it, and have either success or failure, like it or hate it, I'd love feedback! Thanks. David -- /==\ | David Mansfield | | [EMAIL PROTECTED] | \==/ ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Is this possible? More Info
David Taylor wrote: Don Weeks wrote: I need to pass this info over to another system, perhaps the problem reporting tool or the project management tool: PR# 321 was solved by adding foo.c version 3.2.1 and foo.h version 3.3.1. These now have the tag REL3.2 (This implies that if I checkout all the files using tag REL3.2, I can build the entire product.) Sometimes called change sets. If the problem reporting tool is Bugzilla, then you may find cvszilla to be useful: http://homepages.kcbbs.gen.nz/~tonyg It contains a Perl script, that can be run from loginfo, that passes loginfo (name of file, old and new version, log message) to the Bugzilla record for the bug cited in the log message. If you're not using Bugzilla, you still may want to view cvszilla as one example of interfacing CVS with a problem tracker. dtayl Looks like the new tool CVSps may also provide what you need: its patchset == change set. From http://www.cobite.com/cvsps/README : CVSps's output is information about patchsets. A patchset looks like: - PatchSet 1701 Date: 2001/11/06 19:49:04 Author: joe Log: this release completes line summary find Members: Makefile:1.3.4.9-1.3.4.10 [v4_1-production-patches] apache_mod/lineSummary.C:1.66.2.2-1.66.2.3 [v4_1-production-patches] apache_mod/tbill_sql.C:1.59.2.5-1.59.2.6 [v4_1-production-patches] - This patchset is taken from an internal project. It shows the date, the author, log message and each file that was modified. For each file the pre-commit and post-commit revisions are given. In this case, you can see that the files are on a branch, and the branch tag is shown (for each file) inside square brackets. dtayl ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: copy repository to a new cvs-server
Wolfgang Kormann wrote: hi, I have a new cvs-server and I want to copy the whole repository from the old to the new server. The newest cvs-version is installed and running on the new one. What is the best way to do this? See http://mail.gnu.org/pipermail/info-cvs/2001-June/016556.html ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
cvswrappers for text files with no merging allowed
I'm struggling to know what to put in cvswrappers: I have text files, but they should never be merged. Should I then use * -k 'o' -m 'COPY' or should I use * -k 'b' -m 'COPY' and does the -m 'COPY' actually have an effect? The files have already been checked in, so how do I use cvs admin to correct the mistake with respect to the -m 'COPY'? cvs admin -m appears to have a different meaning. Thanks, Dave. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
merging branch into trunk while keeping trunk fixes
I have branch MY_BRANCH and HEAD development. There have been fixes to HEAD and new development that does not have these fixes on MY_BRANCH. It seems that the fixes to HEAD are not retained when, in my HEAD work area, I do cvs update -j MY_BRANCH. Instead, files which seem to have not changed replace the files that have been changed and commited to HEAD. Am I doing something wrong here? Do I have to merge HEAD fixes into MY_BRANCH first? Thanks, Dave. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: merging branch into trunk while keeping trunk fixes
Please disregard this, I think I fat-fingered something here, because now it works for me the way I think it should. On Mon, Dec 03, 2001 at 11:09:02AM -0700, David Everly wrote: I have branch MY_BRANCH and HEAD development. There have been fixes to HEAD and new development that does not have these fixes on MY_BRANCH. It seems that the fixes to HEAD are not retained when, in my HEAD work area, I do cvs update -j MY_BRANCH. Instead, files which seem to have not changed replace the files that have been changed and commited to HEAD. Am I doing something wrong here? Do I have to merge HEAD fixes into MY_BRANCH first? Thanks, Dave. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: cvswrappers for text files with no merging allowed
Thanks Actually, we do have files that need to be merged, but there are others that shouldn't and should have no keyword expansion either. I didn't want to confuse the issue, since I already knew how to deal with the others. On Mon, Dec 03, 2001 at 01:55:49PM -0500, Larry Jones wrote: David Everly writes: I have text files, but they should never be merged. Then why are you using CVS? Should I then use * -k 'o' -m 'COPY' or should I use * -k 'b' -m 'COPY' and does the -m 'COPY' actually have an effect? You don't want keyword expansion, either? You should use -ko -- -kb uses binary I/O, which isn't appropriate for text files. The files have already been checked in, so how do I use cvs admin to correct the mistake with respect to the -m 'COPY'? cvs admin -m appears to have a different meaning. There's no need -- the -m status is stored only in the cvswrappers file. -Larry Jones Somebody's always running my life. I never get to do what I want to do. -- Calvin -- [EMAIL PROTECTED] V-Net: 622-3286 Phone: 1-719-535-3286 Pager: 1-800-724-3624 # 140-1311 ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: How to bring a branch back on a file when the tag has been moved
I don't like moving or deleting tags, so I have the following taginfo script: #!/bin/sh ## ## do not allow anyone to move or delete tags ## if [ $2 = mov ] ; then echo ---Please do not try to move or change existing tags. exit 1 fi if [ $2 = del ] ; then echo ---Please do not try to delete tags. exit 1 fi On Fri, Nov 30, 2001 at 12:03:22PM -0500, [EMAIL PROTECTED] wrote: Subject: Re: How to bring a branch back on a file when the tag has been moved To: [EMAIL PROTECTED] (Prabhu Ram) Date: Fri, 30 Nov 2001 11:01:52 -0500 (EST) Cc: [EMAIL PROTECTED] From: [EMAIL PROTECTED] (Larry Jones) Prabhu Ram writes: We have a branch on which dev. was taking place and someone accidentally moved the tag (which is also the branch name) to the tip of the branch. Is there a nice way of setting the branch back to the branch point ? No. CVS won't move an existing tag unless you force it to using the -F flag, so you may want to beat up any users that always use -F to prevent future mistakes of that sort. The current development version of CVS goes one step farther and refuses to convert a branch tag to a revision tag unless you also specify the -B flag, which should prevent such accidents completely. -Larry Jones Sometimes I think the surest sign that intelligent life exists elsewhere in the universe is that none of it has tried to contact us. -- Calvin ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: nonmergeable file needs merge
By the way, the reason I'm using -k 'b' is so that there is no substitutionbut these should be all mergeable text files. On Thu, Nov 29, 2001 at 04:12:18PM -0700, David Everly wrote: My cvswrappers has: * -k 'b' Then I started getting cvs update: nonmergeable file needs merge. So I tried: * -k 'b' -m 'MERGE' However, the problem still happens...possibly because all the files were checked in on the earlier configuration. What is the setting I really want here, and how do I make sure everything in my repsitory now has it? Thanks, Dave. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
nonmergeable file needs merge
My cvswrappers has: * -k 'b' Then I started getting cvs update: nonmergeable file needs merge. So I tried: * -k 'b' -m 'MERGE' However, the problem still happens...possibly because all the files were checked in on the earlier configuration. What is the setting I really want here, and how do I make sure everything in my repsitory now has it? Thanks, Dave. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: nonmergeable file needs merge
Thanks...this fixed it... On Thu, Nov 29, 2001 at 04:12:05PM -0800, Paul Sander wrote: Have you looked into using -ko for this purpose? --- Forwarded mail from [EMAIL PROTECTED] By the way, the reason I'm using -k 'b' is so that there is no substitutionbut these should be all mergeable text files. --- End of forwarded message from [EMAIL PROTECTED] -- [EMAIL PROTECTED] V-Net: 622-3286 Phone: 1-719-535-3286 Pager: 1-800-724-3624 # 140-1311 ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
CVS problem
Hi, I have just set up a repository on a machine and I intend for all users to connect through the pserver. I have followed all of the instructions, and I can verify that all users can log in and log out of CVS. However, I get problems when I try to import code (although it does seem to work). The main problem is I cannot check out code. The error I get is cvs server: cannot open /root/.cvsignore: Permission denied cvs [server aborted]: can't chdir(/root): Permission denied after I try to do a checkout. It looks like an obvious permissions problem, but I can't see where. Does anybody know where I an going wrong? Thanks very much, David Churches. - David Churches Department of Physics and Astronomy Cardiff University, PO Box 913, Cardiff, CF2 3YB, U.K. Phone: + 44-29-20874785, 20875121 (direct line) Fax: + 44-29-20874056 [EMAIL PROTECTED] ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: Revision on a Branch
-Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] Sent: Wednesday, November 28, 2001 1:07 PM To: [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Subject: Re: Revision on a Branch [EMAIL PROTECTED] writes: ex . I would like to update to ver: 1.4 on a branch 'br1' That doesn't make any sense -- revision 1.4 is on the trunk, it's not on a branch. Ah, but it can be on a branch, provided that no revisions have ever been committed to that branch. In that case, cvs update -r br1 will bring up 1.4. If that's not the case, if for example the tip of br1 is 1.4.2.3, then it doesn't make sense. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: Can we find all the branches on a module through some script?
-Original Message-
From: Jerry Nairn [mailto:[EMAIL PROTECTED]]
Sent: Monday, November 26, 2001 4:34 PM
To: '[EMAIL PROTECTED]'; [EMAIL PROTECTED]
Subject: RE: Can we find all the branches on a module through some
script?
From: jsk-intoto [mailto:[EMAIL PROTECTED]]
Sent: Wednesday, November 21, 2001 9:25 PM
based on the branch the user is working on. Can we obtain
the branches
existing on the repository thru some script mechanism?
cvs -nq log -h something | \
sed -n -e '/^symbolic/,/^keyword/ {
/[1-9][0-9]*\.[0-9][0-9]*\.0\.[0-9][0-9]*$/ p
}'
Will list all of the branches on something. The key thing
here is tags which
have at least four numbers, and the next to last number is 0.
It occurs to me that this will not get those special branches
created by
imports.
Jerry
It's also possible to parse the output of cvs stat -v, which has
lines in the form of
TAGNAME (branch: 1.2.14)
TAGNAME1 (revision: 1.2.14.1)
In Perl, this would be something like
\s*(\w*)\s+\(branch: ([0-9.]*)
for branch names, with corresponding for revisions, with $1 being
the tag name and $2 being the rev number.
___
Info-cvs mailing list
[EMAIL PROTECTED]
http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Need cvs binaries for sun solaris...
Thirumal Raj wrote: hey.. I need binaries of cvs to install in sun solaris system .PLS help me where can i get that binaries or installable files so that i can build it. http://www.cvshome.org/dev/codeunix.html ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: A newbie question about branches
-Original Message- From: Elbert Andrés Messa Díaz [mailto:[EMAIL PROTECTED]] Sent: Tuesday, November 20, 2001 7:23 AM To: [EMAIL PROTECTED] Subject: A newbie question about branches I have a project under CVS version control. Right now I am planning to make a dramatic change on a previous release of the project, but I don't want to affect the main trunk. I would think that it is appropiate to create a new branch on that previous release, and make modifications to the branch. I completely agree. Since I won't incorporate the modifications on the main trunk, because both branches would be irreconciliable, should I just keep the branch forever and never merge? or is it worth it to create a complete new project starting on that previous release? You don't have to merge branches for them to be useful. We split off a release branch for each release, and, while we will merge bugfixes to later branches and head, there's stuff we don't merge. You could create a complete project, but that would lose some history and probably increase the storage requirements (if you care about that). In addition, you'd lose the ability to merge bugfixes you find. In favor of a separate project, if you're going to be working on both branches indefinitely, it will take longer and longer to generate the tip of the branch. In my case, given a slow server and a 20,000 + line file with about thirty changes since the branch was cut and about thirty changes on the branch, it's noticeable. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Using Common files across multiple projects
Thomas Frasher wrote: Hi all, I'm trying to change the repository that we are using here. There are several projects that use some of the same files, headers and .cpp files. Ideally I would like the checkout to get the files automatically, whether they are in a different repository (i.e. a different module) If the files are in a different module, but same repository you can create aliases to get what you want, with something like: checkoutForProject1 -a project1 commonHeaderDir commonCPPDir checkoutForProject2 -a project2 commonHeaderDir commonCPPDir If the common files are in a different repository from the project, this won't work. You'll need to create a simple script to do the checkout that does something like: cvs -d /cvsrootCommon checkout commonHeaderDir commonCPPDir cvs -d /cvsrootProjects checkout project1 or whether they are local. I don't know what you mean by local. This will save me many hours explaining this. Also I need to put them back to the home repository when the changes are committed (all this without the need to commit me at the end of it). This is done automatically on commit: the workspace files called CVS/Root take care of that: see http://www.cvshome.org/docs/manual/cvs_2.html#SEC10 dtayl ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Request to add ability to list files in cvs repository
I find viewcvs ( http://viewcvs.sourceforge.net ) does a fine job of displaying this info. dtayl Kerry Keal wrote: Hi, What I would like is a cvs command to list files in a cvs repository, maybe something like: cvs ls . The command should support the common command flags (-D -l -r ) such as: cvs ls -r rev1 . Another nice feature would to be able to list all the tag names for a particular file with something like -t. Output might be something like This would list all the files in the cvs repository cvs ls -t . module1/file1.c module1/file2.c module1/file3.c module2/fileA.c module2/fileB.c module2/fileC.c Here all revision names in module1 are listed. cvs ls -t module1 module1/file1.c:rev1,rev2,rev3 module1/file2.c:rev1,rev2,rev3 module1/file3.c:rev3 In this case all revisions up to rev2 are listed. cvs ls -t -r rev2 module1 module1/file1.c:rev1,rev2 module1/file2.c:rev1,rev2 Is there any chance of getting this type of functionality into cvs? Any comments about the proposed format. I'm willing to help do some coding if that is what it would take to get this functionality into cvs. Regards, Kerry ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
A problem configing Hosts.sam in ME?
I am having problems getting ME to recognize the new server added to Hosts.Sam. I believe the syntax is right IPSPACEALIAS I can ping the localhost obviously under same exact format. But the added server will not ping by its alias only the direct ip. Its not recognizing the added host. Therefore, login is denied under alias. Does anyone know why or how to fix it? ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: cvs-1.11.1p1 binary installer
http://www.cvshome.org/dev/codeunix.html has links to binaries for Solaris 2.5.1 and 2.6. I've built a Solaris package under 2.7 if you need it. dtayl Li, Jerry wrote: Hi, All: Could anybody in this list tell me where I can download a binary installer of cvs-1.11.1p1 for Solaris 7? I have downloaded the source of this version, but got some compiling errors. thanks, Jerry -- Note: The information contained in this message may be privileged and confidential and protected from disclosure. If the reader of this message is not the intended recipient, or an employee or agent responsible for delivering this message to the intended recipient, you are hereby notified that any dissemination, distribution or copying of this communication is strictly prohibited. If you have received this communication in error, please notify us immediately by replying to the message and deleting it from your computer. Thank you. -- ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: CVS - setup reserved checkout
Kaz Kylheku writes: [...with respect to CVS...] Tell the manager to shed his or her superstitions, and work with the facts. The facts are: - Concurrent development works just fine. - Your team already likes it. - Strict locking does not prevent concurrency, it only reduces it to a coarse granularity: coarse enough to interfere with productivity, but not coarse enough to eradicate conflicts. To eliminate conflicts, you have to lock the entire repository so that only one developer at a time can do anything on the software base as a whole. Since it is already working for you, you can invite the manager to witness, or participate in, some of your day to day version control activities. The point is that the development policy should fit the configuration management tool and the CM tool should fit the development policy. If the two don't get along, then the development environment is broken (well, if not broken, certainly very hampered). Brad Appleton's papers on SCM patterns provide a good start at understanding how to setup your policies and patterns: http://www.enteract.com/~bradapp/acme -- David Mastersondmaster AT synopsys DOT com Sr. RD Engineer Synopsys, Inc. Software Engineering Sunnyvale, CA ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: CVS - setup reserved checkout
[EMAIL PROTECTED] (Kaz Kylheku) wrote: Tell the manager to shed his or her superstitions, and work with the facts. The facts are: - Concurrent development works just fine. - Your team already likes it. - Strict locking does not prevent concurrency, it only reduces it to a coarse granularity: coarse enough to interfere with productivity, but not coarse enough to eradicate conflicts. To eliminate conflicts, you have to lock the entire repository so that only one developer at a time can do anything on the software base as a whole. Well said. May I add, Concurrency works best with good communication among the developers. Responsibility of certain sections of code is usually divvied among just a few people. Strict locking might hurt the need for good communication among a group. -- David Gravereaux [EMAIL PROTECTED] $ make war make: *** No rule to make target `war'. Stop. Try `love' instead. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: pvcs2cvs.pl
David, - Original Message - From: David Everly [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Wednesday, October 24, 2001 6:12 PM Subject: pvcs2cvs.pl I ran the following on a copy of my pvcs archives: pvcs2cvs.pl -r flat -p flat Then copied all the *,v files into a cvs root directory which I then did 'cvs init' on. Should I have taken a different approach here? I normally would do a cvs init first, then copy the archives over, but I don't think the order should matter. I also would recommend keeping the PVCS and RCS archives in separate directories (i.e. only use either -r flat or -p flat, but not both). There was a pvcs floating label which originally was defined as 1.* which got converted into 0.1 and it seems I can no longer check out based on that label or that revision number. It's possible a floating label 1.* might not convert properly. To PVCS, this typically would refer to the latest revision on the trunk (unless you have major revision number 1 e.g. 2.0, 2.1, etc - the script may not handle this case correctly). In CVS, the default revision checked out is the latest revision on the trunk, so for CVS usage, you could use -r HEAD or -r 1 instead, or even just use the default trunk revision. The pvcs2cvs.pl script tries to create a CVS branch tag from a PVCS floating label, erroneously in this case of 1.*. A PVCS floating label on the trunk can't be converted to an equivalent user-defined branch tag in CVS. Also, although the latest version of the files appear to have been checked out into the directory along with the ,v and -arc files, the most recent version can not be checked out of this new cvs archive. The checked out files should not appear in your repository area, only the RCS archives (,v files) should. These are likely remnants of the conversion process. You might refer to my readme file at http://pages.prodigy.net/dlmart/scm/id19.htm which provides usage detailed on how I've used the conversion script. Regards, David Martin ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
pvcs2cvs.pl
I ran the following on a copy of my pvcs archives: pvcs2cvs.pl -r flat -p flat Then copied all the *,v files into a cvs root directory which I then did 'cvs init' on. Should I have taken a different approach here? There was a pvcs floating label which originally was defined as 1.* which got converted into 0.1 and it seems I can no longer check out based on that label or that revision number. Also, although the latest version of the files appear to have been checked out into the directory along with the ,v and -arc files, the most recent version can not be checked out of this new cvs archive. Is there some work around for this? -- [EMAIL PROTECTED] V-Net: 622-3286 Phone: 1-719-535-3286 Pager: 1-800-724-3624 # 140-1311 ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: CVS Sharing
-Original Message- From: Bryon Lape [mailto:[EMAIL PROTECTED]] Ah, the typical Unix answerroll your own... If you would prefer the typical non-Unix answer, please send me a thousand dollars and I'll tell you the product doesn't do that. If you insist enough, for extra money I'll provide a patch that makes you go through additional hoops if you try anything similar. For enough extra money, I'll tell you it's a bug. (I've had experience dealing with commercial software support, and these are synopses of true stories. The names have been omitted to protect the guilty.) More seriously, if you would like somebody to write a script such as was designed, you can probably hire somebody to do it for far less than the cost of one ClearCase seat. You won't have to wait for the vendor to decide to do it, or negotiate with the vendor so you can pay an arm and a leg for the vendor to do it closer to your schedule. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
stopping 'cvs tag -b'
Since I didn't see any responses, I thought I would ask this once more: How can I keep users from doing 'cvs tag -b' and still allow 'cvs tag' I'm aware of CVSROOT/taginfo, but I don't see any arguments that indicate whether it will create a branch or not. $2 only has 'add', 'mov', 'del'. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: CVS export
-Original Message- From: Lee Sau Dan [mailto:[EMAIL PROTECTED]] Sent: Thursday, October 18, 2001 12:23 PM To: [EMAIL PROTECTED] Subject: Re: CVS export This message uses a character set that is not supported by the Internet Service. To view the original message content, open the attached message. If the text doesn't display correctly, save the attachment to disk, and then open it using a viewer that can display the original character set. This is how your message came across in Microsoft Outlook (no, it wasn't my idea to use it companywide). Posting in a form that people can't read is not going to help them respond to you. Please make sure you're posting in straight ASCII. Everybody can read that easily. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
GNU expr
Is GNU expr required for CVS or only for the 'make check' testing of CVS? ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: CVS a Web site
-Original Message- From: Jason Purdy [mailto:[EMAIL PROTECTED]] Sent: Wednesday, October 17, 2001 3:22 PM To: [EMAIL PROTECTED] Subject: CVS a Web site Hi, What is a good scheme of automating synchronization between a CVS repository and a Web site (or does a repository have to be separated from the Web site files [besides renaming the files from the ,v suffix])? More than that is required. You will need to checkout or export the files, not just rename them from the repository. A co-worker I have recently decided to work with CVS on our ever-growing Web site, and I've got CVS up going, but lack an important piece of the puzzle - getting data out of the repository and into the Web site (hopefully in an automated fashion). I have a couple of ideas, but before I go down a road that's surely been traversed before, I'd love to get some input from the more travelled on the list. I hope there's already a solution for this scenario, though I couldn't find anything on the cvshome.org Web site or in a Google search. 2) Create 'trigger' scripts which would be triggered when the developer 'commit'ed their changes, which would also affect the changes on the Web site directories. This is the correct answer, and the file you are looking for is commitinfo within CVSROOT. That file will specify programs to be run when something is checked in, and one to check out changed files and copy them to the web site should work just fine. Since you've been searching in the right place, you should be able to find the commitinfo material in the documentation. Sounds like all you needed was a pointer. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Cygwin CVS: local server: storing existing directory
Jerry == Jerry Nairn [EMAIL PROTECTED] writes: Jerry Is there something weird in your .cvsrc? Jerry I'm guessing that no one is answering this because you should not be getting Jerry this error. One explanation would be if you had something in your Jerry environment changing the behavior of cvs. Jerry Jerry Apparently that was it. I had set up a $HOME/.cvsrc with just the contents checkout -r. That's clearly bogus. I can't even remember what I was trying to achieve with that. After I moved that file out of the way, the checkout proceeded fine. Thanks. -- === David M. Karr ; Best Consulting [EMAIL PROTECTED] ; Java/Unix/XML/C++/X ; BrainBench CJ12P (#12004) ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: CVS - setup reserved checkout
-Original Message- From: Bryon Lape [mailto:[EMAIL PROTECTED]] Wrong. Kaz Kylheku wrote: In article [EMAIL PROTECTED], [EMAIL PROTECTED] wrote: You need to ask yourself why your group is experiencing so many conflicts while so many other groups (thousands?) are not. Because there is no group, and there are no conflicts. This is just another Chicken Little yelling that the sky is falling. Actually a step beneath Chicken Little, because something actually did fall on Chicken Little's head, it wasn't just pure imagination. :) Gee, that's an informative response. You burst on the scene with complaints about a feature of CVS that has given none of us serious problems, and never explain why it is a problem for you. You never answer questions about what it is that you are doing, or make any significant responses to suggestions. The only way some of us have been able to interpret your complaints about CVS is that you have such a messed-up shop that no version control system is going to work, and that strict locking is merely going to shift the mess around a bit, and probably increase it. If you are here merely as a troll, then you're getting more consideration than you deserve. If you have a legitimate problem, then your problems are far more severe than any software product can handle, and free advice on mailing lists isn't going to help either. Right now, you're asking questions akin to When I use Amoco gasoline, my company's cars always catch fire. What additive should I use? ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: Handling project documentation using CVS
-Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] Sent: Friday, October 12, 2001 4:28 PM To: [EMAIL PROTECTED] Subject: Handling project documentation using CVS Hello all, I was wondering if some of you nice people could give me some feedback on an issue I've been wrestling with. Besides the actual source code to a project, I also need the ability to version design documentation ( and maybe other stuff, but this is my immediate concern ). By design documentation, I mean things like Visio documents, with UML diagrams, etc. As opposed to TeX or ?roff documents, I guess. These are likely to be proprietary formats best represented as binary files, which means that they aren't well suited for CVS. On one level, I have a feeling that CVS isn't the best way to handle versioning these documents. It kinda feels like using the wrong tool for the job. But, on the other hand, I can't think of any really, really, solid reasons why NOT to do it. I think it would be more accurate to say that these files aren't the ones CVS handles best. I think that a systems that does file locking better than CVS would do a little better, but nobody's shown me a system that works much better than CVS. You have to remember that automatic merging isn't going to work, and in general the only way to merge is to take one of the versions and manually recreate the changes. This means that you want at least advisory locking, so that anybody starting work on such a file will know that there's likely to be a conflict that will require redoing changes. It also means that, while you can have branches, merging doc changes between branches will not in general work, and so it would probably be necessary to redo the work when applying changes to more than one branch. CVS won't store the files efficiently, so frequent small changes are likely to take lots of disk space in the repository. This may or may not be important. If you have another system that does what you need it to do and has good file locking facilities, then it may or may not be worth your while to use it to control documentation in a separate project. It is likely to do somewhat better than CVS, but introduces two version control systems and potential confusion. So, if some of you have some experience, or thoughts, on this issue, please share them with me. It works. It isn't ideal, but it works. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Cygwin CVS: local server: storing existing directory
David == David M Karr [EMAIL PROTECTED] writes: David (I sent this to the list on 10/12, but I never saw it. I have the option set David to get my posts sent to me.) David This should be a no-brainer, but I just can't get my CVS setup going. David I'm on Win2k, with Cygwin 1.3.2, CVS 1.11 (the cvs built into Cygwin). David I have a directory $HOME/java/sgs that I want to store into CVS. David I'm setting my CVSROOT to :local:/cygdrive/c/cvsroot. David I want to check out sgs into $HOME/java/cvswork/sgs. David Ok, so in the sgs directory, I did this: Davidcvs import -m Simple Grading System sgs intsoft start David This appeared to work, creating lots of N and cvs import: Importing ... David lines. It did, however, skip importing one source file, because it was in a David DIRECTORY named tags. I know why that happened. I'll get that file added David once I get everything else working. David Anyway, now in /cygdrive/c/cvsroot, there is a sgs directory that contains David all of my RCS files. David So, now I go to $HOME/java/cvswork and try to figure out how to check it David out. I'll show a little bit of shell history here, indicating what I tried. David David % pwd David /home/dmkarr/java/cvswork David % cvs checkout sgs David cvs [checkout aborted]: must specify at least one module or directory David % mkdir sgs David % cvs checkout sgs David cvs [checkout aborted]: must specify at least one module or directory David % cd sgs David % cvs checkout sgs David cvs [checkout aborted]: must specify at least one module or directory David David So what am I doing wrong? I still could use some help figuring this out. I even changed my setup slightly to make it more similar to the example in the CVS manual. I set my CVSROOT to :local:/c/cvsroot. I did cvs init. This created a CVSROOT directory in /c/cvsroot. I then went to $HOME/java/sgs and did cvs import -m Simple Grading System intsoft/sgs intsoft start. This printed lines like this: - N intsoft/sgs/build.xml N intsoft/sgs/sgs.txt N intsoft/sgs/Status cvs import: Importing /c/cvsroot/intsoft/sgs/bin N intsoft/sgs/bin/createdataset N intsoft/sgs/bin/listclasses N intsoft/sgs/bin/rawinitdb cvs import: Importing /c/cvsroot/intsoft/sgs/clientsrc cvs import: Importing /c/cvsroot/intsoft/sgs/clientsrc/com cvs import: Importing /c/cvsroot/intsoft/sgs/clientsrc/com/intsoft cvs import: Importing /c/cvsroot/intsoft/sgs/clientsrc/com/intsoft/sgs cvs import: Importing /c/cvsroot/intsoft/sgs/clientsrc/com/intsoft/sgs/client N intsoft/sgs/clientsrc/com/intsoft/sgs/client/AppCallbackHandler.java ... - I then went to $HOME/cvswork and did cvs checkout intsoft/sgs. It gave me the following response: cvs [checkout aborted]: must specify at least one module or directory I would appreciate any help in understanding what I'm doing wrong. -- === David M. Karr ; Best Consulting [EMAIL PROTECTED] ; Java/Unix/XML/C++/X ; BrainBench CJ12P (#12004) ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: cvs [login aborted]: connect to 216.254.34.156:2401 failed: Connection timed out
You seems to have the same problem as me. The firewall systematically block the ports he knows nothing about. Usually they know ftp, web, telnet, ssh, https but cvs is not on the list. Unfortunately, asking your system administrator to open the port seems to be the only solution. And i know this is very difficult. In my case, nobody seems to be able to tell me who the hell is the firewall administrator. So i can't CVS from school Hope this will disespair you, David Delbecq ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Cygwin CVS: local server: storing existing directory
(I sent this to the list on 10/12, but I never saw it. I have the option set to get my posts sent to me.) This should be a no-brainer, but I just can't get my CVS setup going. I'm on Win2k, with Cygwin 1.3.2, CVS 1.11 (the cvs built into Cygwin). I have a directory $HOME/java/sgs that I want to store into CVS. I'm setting my CVSROOT to :local:/cygdrive/c/cvsroot. I want to check out sgs into $HOME/java/cvswork/sgs. Ok, so in the sgs directory, I did this: cvs import -m Simple Grading System sgs intsoft start This appeared to work, creating lots of N and cvs import: Importing ... lines. It did, however, skip importing one source file, because it was in a DIRECTORY named tags. I know why that happened. I'll get that file added once I get everything else working. Anyway, now in /cygdrive/c/cvsroot, there is a sgs directory that contains all of my RCS files. So, now I go to $HOME/java/cvswork and try to figure out how to check it out. I'll show a little bit of shell history here, indicating what I tried. % pwd /home/dmkarr/java/cvswork % cvs checkout sgs cvs [checkout aborted]: must specify at least one module or directory % mkdir sgs % cvs checkout sgs cvs [checkout aborted]: must specify at least one module or directory % cd sgs % cvs checkout sgs cvs [checkout aborted]: must specify at least one module or directory So what am I doing wrong? -- === David M. Karr ; Best Consulting [EMAIL PROTECTED] ; Java/Unix/XML/C++/X ; BrainBench CJ12P (#12004) ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: CVS - setup reserved checkout
-Original Message- From: Bryon Lape [mailto:[EMAIL PROTECTED]] The benefits add up to zero. Now, if it did method locking, that would be helpful, protective AND productive. Without some sort of locking, having developers waste time with doing merging by hand is counterproductive. What do you mean by method locking? Locking individual parts of a file? It wouldn't do you any good. If you are getting large amounts of conflicts with CVS merging, that means that multiple people are changing the same parts of files in different ways. If the changes were localized in the files, so that different developers would be locking different member functions, then CVS would merge the changes just fine. In my experience, with some sort of locking developers waste time doing merging by hand. Developer A is adding a feature, and a bug report comes in from the field so developer B is assigned to fix it. B is now trying to hurry A up so she checks in and releases the lock, which means that A is likely to skimp on unimportant things like testing. Assuming B has not simply been playing 5,235 games of Minesweeper while waiting, B has likely figured out how to fix the bug, and then finds that A has modified that section and so he has to redo the bugfix. (Of course, if A did not modify that section, CVS would work just fine.) Alternatively, management yanks the lock away from A and gives it to B, who fixes the bug and checks in, and A now has to do the manual merging. Since merging of some sort is necessary when you have more than one person (or, in some cases, one person with more than one project) working on the same file, it's useful if the version control system actually has facilities to assist with the merge. Given that, it makes sense to allow concurrent development. You claim that the benefits are zero, in spite of the fact that many, many projects have found them to be great. It's pretty simple, really. If you have developers all over the place, changing everything in sight, then CVS isn't going to help you, but neither is anything else, because your shop is thoroughly messed up. If you have developers working on specific projects that change specific parts of the code, even if scattered among several files, then CVS is going to help you. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: CVS - setup reserved checkout
-Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] Sent: Friday, October 12, 2001 11:46 AM To: [EMAIL PROTECTED] Subject: RE: CVS - setup reserved checkout [ On Friday, October 12, 2001 at 09:35:58 (-0500), Thornley, David wrote: ] Subject: RE: CVS - setup reserved checkout What do you mean by method locking? Locking individual parts of a file? It wouldn't do you any good. Well, not with CVS anyway! :-) Maybe in a multi-user smalltalk image it might (since you only ever edit one method at a time I was apparently unclear; I meant that method locking would do no good for anybody who finds CVS unusable because of merge conflicts. If people can work on separate methods OK, then using CVS it really doesn't matter if they're parts of the same file or not, because the changes won't conflict. If, on the other hand, everybody is messing with widespread changes all the time, which is basically what you'd have to do to have that much trouble with CVS, method locking is no better than file locking, probably more likely to cause deadlocks, and certainly more of a pain to find who's using all the locks you need and why. If you *want* to use a locking version control system on files you edit in distinct segments, then I suppose locking by method is more suitable to your desires than locking by file. In that situation, though, there's no reason not to go concurrent. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: CVS - setup reserved checkout
-Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] Sent: Friday, October 12, 2001 1:34 PM To: [EMAIL PROTECTED]; [EMAIL PROTECTED] Subject: RE: CVS - setup reserved checkout One would hope that one's shop is not using the same branch for both maintenance and new features. That kind of thing is best done on separate branches (where the two schedules don't interfere with each other). The bug fix is later merged into the new development when it's appropriate to do so. The last job I had not involving the use of CVS was with SCCS, and we didn't have branches. This did make shipping bug-fixed stuff to customers interesting. Now assume the conditions where I'm working now, where the new features go on the main trunk and the bugfixes will be applied to a release branch, or maybe a patch subbranch. These need to be merged eventually, and I'd rather get them merged now before the developer forgets about them. Under those conditions, almost any version control tool provides the necessary merge tool. And locks don't matter because there's no concurrent development on the same branch. Any version control tool with branches. Of course, anything going in on the release branch probably should go into the development branch, and we're back to merging. The question, I suppose, is whether the merge will be done semi-automatically and promptly, so that the developer fixing the bug will watch it happen and have the problem fresh in his or her mind, or if it's going to be done manually and possibly at a later time, when the developer doesn't quite remember all the details, or not at all, and the developer finds a note three years later stuffed into documentation for an old version of the compiler about merging the change. I know which I prefer, but others seem to prefer cases 2 or 3. 'Course, it's a different story if multiple developers are adding their own bug fixes or features on either branch... Yup. Any time more than one developer is working on things at the same time, there's a need for merging. It is possible to design a locking protocol that obviates the need for merges or wasted work. When a developer has a project, he or she grabs all needed locks. If that developer cannot grab all of them, he or she releases the grabbed ones (to avoid deadlock, except in the case of race conditions. This can be avoided by giving each developer a different time of day to grab locks). If the developer has all locks necessary for a task, he or she works on that task. If the developer does not have all locks necessary for any assigned task, that developer surfs the web or plays bocce ball or something. Personally, I'm not convinced that this is better than having to merge. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Cygwin CVS: local server: storing existing directory
This should be a no-brainer, but I just can't get my CVS setup going. I'm on Win2k, with Cygwin 1.3.2, CVS 1.11 (the cvs built into Cygwin). I have a directory $HOME/java/sgs that I want to store into CVS. I'm setting my CVSROOT to :local:/cygdrive/c/cvsroot. I want to check out sgs into $HOME/java/cvswork/sgs. Ok, so in the sgs directory, I did this: cvs import -m Simple Grading System sgs intsoft start This appeared to work, creating lots of N and cvs import: Importing ... lines. It did, however, skip importing one source file, because it was in a DIRECTORY named tags. I know why that happened. I'll get that file added once I get everything else working. Anyway, now in /cygdrive/c/cvsroot, there is a sgs directory that contains all of my RCS files. So, now I go to $HOME/java/cvswork and try to figure out how to check it out. I'll show a little bit of shell history here, indicating what I tried. % pwd /home/dmkarr/java/cvswork % cvs checkout sgs cvs [checkout aborted]: must specify at least one module or directory % mkdir sgs % cvs checkout sgs cvs [checkout aborted]: must specify at least one module or directory % cd sgs % cvs checkout sgs cvs [checkout aborted]: must specify at least one module or directory So what am I doing wrong? -- === David M. Karr ; Best Consulting [EMAIL PROTECTED] ; Java/Unix/XML/C++/X ; BrainBench CJ12P (#12004) ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Freezing CVS tags
I'm thinking that what I do works to not allow tag changes: First, I define a 'cvsadmin' group. Second, I place this line in $CVSROOT/CVSROOT/taginfo: ALL $CVSROOT/CVSROOT/tagrules.sh Third, I place tagrules.sh in my $CVSROOT/CVSROOT/checkoutlist Fourth, my tagrules.sh looks like this: #!/bin/sh if [ $2 = mov ] ; then echo ---Please do not try to move or change tags. exit 1 fi if [ $2 = del ] ; then echo ---Please do not try to delete tags. exit 1 fi exit 0 On Fri, Oct 12, 2001 at 04:51:18PM -0400, Larry Jones wrote: [EMAIL PROTECTED] writes: Is there a way to freeze the CVS tag? After we create a CVS tag, we would like to freeze it so that no one can change it and causes problem, is this possible? No, but note that CVS won't move an existing tag unless you force it to by using the -F flag -- simply threaten to break the fingers of any user who uses it. :-) -Larry Jones What this games needs are negotiated settlements. -- Calvin ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs -- [EMAIL PROTECTED] V-Net: 622-3286 Phone: 1-719-535-3286 Pager: 1-800-724-3624 # 140-1311 ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Web interfaces that supports the CVSROOT/modules
http://www.sourceforge.com/projects/cvsbrowser/ It is in beta, and I haven't had time to work on it in a few months, but development is picking up again. It parses the modules file and builds the display based on that, including support for nested modules (module), including only specific files in a module, alias modules, and excluding specific directories from a module. It is incomplete. You can currently traverse down to the file you want, but very little of the file operations have been completed to date. It is written in PHP. If you or anyone else has interest let me know and I'll bump it up on my priority list a couple notches. -- David F. Rich Wittmer wrote: I have a developer that is interested in a web interfaces that supports the CVSROOT/modules file to display the directory representation rather than the actual set of directories. Does anyone know of a interface of this nature? ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: CVS - setup reserved checkout
Andrew writes: Has anyone setup reserved checkout in CVS (ver 1.11.1p1) in Unix (Solaris)? Or is there any documentation on this other than the manual that comes with the source code? Given the CVS model of unreserved checkouts, why do you need reserved checkouts? Also, are you talking about reserved checkouts of a file or an entire product? -- David Mastersondmaster AT synopsys DOT com Sr. RD Engineer Synopsys, Inc. Software Engineering Sunnyvale, CA ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
checkout wincvs 1.2
i have installed cvs on linux 7.2. With linux it works fine even with the checkout command. With wincvs i´m faced to the problem, that with the checkout i receive always the same error message: cannot access /cvstric/CVSROOT/CVSROOT Is there a default file or ini-file which has to be edited? i tried to checkout the module cvstric/seiten within the [create] [checkout setting] i tried the following syntax: seiten /cvstric/seiten user@servername:/cvstric/seiten all of these command are result in the same error message as mentioned above. the structure has been installed as follows: Server: /cvstric/CVSROOT /cvstric/seiten What am ie doing wrong. Where can i find help? Kind regards David Badstuebner [EMAIL PROTECTED] ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
