F. Troubleshooting

If you are having trouble with CVS, this appendix may help. If there is a particular error message which you are seeing, then you can look up the message alphabetically. If not, you can look through the section on other problems to see if your problem is mentioned there.

F.1 Partial list of error messages

Here is a partial list of error messages that you may see from CVS. It is not a complete list---CVS is capable of printing many, many error messages, often with parts of them supplied by the operating system, but the intention is to list the common and/or potentially confusing error messages.

The messages are alphabetical, but introductory text such as `cvs update: ' is not considered in ordering them.

In some cases the list includes messages printed by old versions of CVS (partly because users may not be sure which version of CVS they are using at any particular moment).

cvs command: authorization failed: server host rejected access

This is a generic response when trying to connect to a pserver server which chooses not to provide a specific reason for denying authorization. Check that the username and password specified are correct and that the CVSROOT specified is allowed by `--allow-root' in `inetd.conf'. See 2.9.3 Direct connection with password authentication.

file:line: Assertion 'text' failed

The exact format of this message may vary depending on your system. It indicates a bug in CVS, which can be handled as described in H. Dealing with bugs in CVS or this manual.

cvs command: conflict: removed file was modified by second party

This message indicates that you removed a file, and someone else modified it. To resolve the conflict, first run `cvs add file'. If desired, look at the other party's modification to decide whether you still want to remove it. If you don't want to remove it, stop here. If you do want to remove it, proceed with `cvs remove file' and commit your removal.

cannot change permissions on temporary directory

Operation not permitted

This message has been happening in a non-reproducible, occasional way when we run the client/server testsuite, both on Red Hat Linux 3.0.3 and 4.1. We haven't been able to figure out what causes it, nor is it known whether it is specific to linux (or even to this particular machine!). If the problem does occur on other unices, `Operation not permitted' would be likely to read `Not owner' or whatever the system in question uses for the unix EPERM error. If you have any information to add, please let us know as described in H. Dealing with bugs in CVS or this manual. If you experience this error while using CVS, retrying the operation which produced it should work fine.

cvs [server aborted]: Cannot check out files into the repository itself

The obvious cause for this message (especially for non-client/server CVS) is that the CVS root is, for example, `/usr/local/cvsroot' and you try to check out files when you are in a subdirectory, such as `/usr/local/cvsroot/test'. However, there is a more subtle cause, which is that the temporary directory on the server is set to a subdirectory of the root (which is also not allowed). If this is the problem, set the temporary directory to somewhere else, for example `/var/tmp'; see TMPDIR in D. All environment variables which affect CVS, for how to set the temporary directory.

cannot open CVS/Entries for reading: No such file or directory

This generally indicates a CVS internal error, and can be handled as with other CVS bugs (see section H. Dealing with bugs in CVS or this manual). Usually there is a workaround--the exact nature of which would depend on the situation but which hopefully could be figured out.

cvs [init aborted]: cannot open CVS/Root: No such file or directory

This message is harmless. Provided it is not accompanied by other errors, the operation has completed successfully. This message should not occur with current versions of CVS, but it is documented here for the benefit of CVS 1.9 and older.

cvs [checkout aborted]: cannot rename file file to CVS/,,file: Invalid argument

This message has been reported as intermittently happening with CVS 1.9 on Solaris 2.5. The cause is unknown; if you know more about what causes it, let us know as described in H. Dealing with bugs in CVS or this manual.

cvs [command aborted]: cannot start server via rcmd

This, unfortunately, is a rather nonspecific error message which CVS 1.9 will print if you are running the CVS client and it is having trouble connecting to the server. Current versions of CVS should print a much more specific error message. If you get this message when you didn't mean to run the client at all, you probably forgot to specify :local:, as described in 2. The Repository.

ci: file,v: bad diff output line: Binary files - and /tmp/T2a22651 differ

CVS 1.9 and older will print this message when trying to check in a binary file if RCS is not correctly installed. Re-read the instructions that came with your RCS distribution and the INSTALL file in the CVS distribution. Alternately, upgrade to a current version of CVS, which checks in files itself rather than via RCS.

cvs checkout: could not check out file

With CVS 1.9, this can mean that the co program (part of RCS) returned a failure. It should be preceded by another error message, however it has been observed without another error message and the cause is not well-understood. With the current version of CVS, which does not run co, if this message occurs without another error message, it is definitely a CVS bug (see section H. Dealing with bugs in CVS or this manual).

cvs [login aborted]: could not find out home directory

This means that you need to set the environment variables that CVS uses to locate your home directory. See the discussion of HOME, HOMEDRIVE, and HOMEPATH in D. All environment variables which affect CVS.

cvs update: could not merge revision rev of file: No such file or directory

CVS 1.9 and older will print this message if there was a problem finding the rcsmerge program. Make sure that it is in your PATH, or upgrade to a current version of CVS, which does not require an external rcsmerge program.

cvs [update aborted]: could not patch file: No such file or directory

This means that there was a problem finding the patch program. Make sure that it is in your PATH. Note that despite appearances the message is not referring to whether it can find file. If both the client and the server are running a current version of CVS, then there is no need for an external patch program and you should not see this message. But if either client or server is running CVS 1.9, then you need patch.

cvs update: could not patch file; will refetch

This means that for whatever reason the client was unable to apply a patch that the server sent. The message is nothing to be concerned about, because inability to apply the patch only slows things down and has no effect on what CVS does.

dying gasps from server unexpected

There is a known bug in the server for CVS 1.9.18 and older which can cause this. For me, this was reproducible if I used the `-t' global option. It was fixed by Andy Piper's 14 Nov 1997 change to src/filesubr.c, if anyone is curious. If you see the message, you probably can just retry the operation which failed, or if you have discovered information concerning its cause, please let us know as described in H. Dealing with bugs in CVS or this manual.

end of file from server (consult above messages if any)

The most common cause for this message is if you are using an external rsh program and it exited with an error. In this case the rsh program should have printed a message, which will appear before the above message. For more information on setting up a CVS client and server, see 2.9 Remote repositories.

cvs [update aborted]: EOF in key in RCS file file,v

cvs [checkout aborted]: EOF while looking for end of string in RCS file file,v

This means that there is a syntax error in the given RCS file. Note that this might be true even if RCS can read the file OK; CVS does more error checking of errors in the RCS file. That is why you may see this message when upgrading from CVS 1.9 to CVS 1.10. The likely cause for the original corruption is hardware, the operating system, or the like. Of course, if you find a case in which CVS seems to corrupting the file, by all means report it, (see section H. Dealing with bugs in CVS or this manual). There are quite a few variations of this error message, depending on exactly where in the RCS file CVS finds the syntax error.

cvs commit: Executing 'mkmodules'

This means that your repository is set up for a version of CVS prior to CVS 1.8. When using CVS 1.8 or later, the above message will be preceded by

cvs commit: Rebuilding administrative file database

If you see both messages, the database is being rebuilt twice, which is unnecessary but harmless. If you wish to avoid the duplication, and you have no versions of CVS 1.7 or earlier in use, remove -i mkmodules every place it appears in your modules file. For more information on the modules file, see C.1 The modules file.

missing author

Typically this can happen if you created an RCS file with your username set to empty. CVS will, bogusly, create an illegal RCS file with no value for the author field. The solution is to make sure your username is set to a non-empty value and re-create the RCS file.

cvs [checkout aborted]: no such tag tag

This message means that CVS isn't familiar with the tag tag. Usually this means that you have mistyped a tag name; however there are (relatively obscure) cases in which CVS will require you to try a few other CVS commands involving that tag, before you find one which will cause CVS to update the `val-tags' file; see discussion of val-tags in 2.2.2 File permissions. You only need to worry about this once for a given tag; when a tag is listed in `val-tags', it stays there. Note that using `-f' to not require tag matches does not override this check; see A.5 Common command options.

*PANIC* administration files missing

This typically means that there is a directory named CVS but it does not contain the administrative files which CVS puts in a CVS directory. If the problem is that you created a CVS directory via some mechanism other than CVS, then the answer is simple, use a name other than CVS. If not, it indicates a CVS bug (see section H. Dealing with bugs in CVS or this manual).

rcs error: Unknown option: -x,v/

This message will be followed by a usage message for RCS. It means that you have an old version of RCS (probably supplied with your operating system), as well as an old version of CVS. CVS 1.9.18 and earlier only work with RCS version 5 and later; current versions of CVS do not run RCS programs.

cvs [server aborted]: received broken pipe signal

This message seems to be caused by a hard-to-track-down bug in CVS or the systems it runs on (we don't know--we haven't tracked it down yet!). It seems to happen only after a CVS command has completed, and you should be able to just ignore the message. However, if you have discovered information concerning its cause, please let us know as described in H. Dealing with bugs in CVS or this manual.

Too many arguments!

This message is typically printed by the `log.pl' script which is in the `contrib' directory in the CVS source distribution. In some versions of CVS, `log.pl' has been part of the default CVS installation. The `log.pl' script gets called from the `loginfo' administrative file. Check that the arguments passed in `loginfo' match what your version of `log.pl' expects. In particular, the `log.pl' from CVS 1.3 and older expects the logfile as an argument whereas the `log.pl' from CVS 1.5 and newer expects the logfile to be specified with a `-f' option. Of course, if you don't need `log.pl' you can just comment it out of `loginfo'.

cvs [update aborted]: unexpected EOF reading file,v

See `EOF in key in RCS file'.

cvs [login aborted]: unrecognized auth response from server

This message typically means that the server is not set up properly. For example, if `inetd.conf' points to a nonexistent cvs executable. To debug it further, find the log file which inetd writes (`/var/log/messages' or whatever inetd uses on your system). For details, see F.2 Trouble making a connection to a CVS server, and 2.9.3.1 Setting up the server for password authentication.

cvs server: cannot open /root/.cvsignore: Permission denied

cvs [server aborted]: can't chdir(/root): Permission denied

See F.2 Trouble making a connection to a CVS server.

cvs commit: Up-to-date check failed for `file'

This means that someone else has committed a change to that file since the last time that you did a cvs update. So before proceeding with your cvs commit you need to cvs update. CVS will merge the changes that you made and the changes that the other person made. If it does not detect any conflicts it will report `M file' and you are ready to cvs commit. If it detects conflicts it will print a message saying so, will report `C file', and you need to manually resolve the conflict. For more details on this process see 10.3 Conflicts example.

Usage: diff3 [-exEX3 [-i | -m] [-L label1 -L label3]] file1 file2 file3

Only one of [exEX3] allowed

This indicates a problem with the installation of diff3 and rcsmerge. Specifically rcsmerge was compiled to look for GNU diff3, but it is finding unix diff3 instead. The exact text of the message will vary depending on the system. The simplest solution is to upgrade to a current version of CVS, which does not rely on external rcsmerge or diff3 programs.

warning: unrecognized response `text' from cvs server

If text contains a valid response (such as `ok') followed by an extra carriage return character (on many systems this will cause the second part of the message to overwrite the first part), then it probably means that you are using the `:ext:' access method with a version of rsh, such as most non-unix rsh versions, which does not by default provide a transparent data stream. In such cases you probably want to try `:server:' instead of `:ext:'. If text is something else, this may signify a problem with your CVS server. Double-check your installation against the instructions for setting up the CVS server.

cvs commit: [time] waiting for user's lock in directory

This is a normal message, not an error. See 10.5 Several developers simultaneously attempting to run CVS, for more details.

cvs commit: warning: editor session failed

This means that the editor which CVS is using exits with a nonzero exit status. Some versions of vi will do this even when there was not a problem editing the file. If so, point the CVSEDITOR environment variable to a small script such as:

#!/bin/sh vi $* exit 0

F.2 Trouble making a connection to a CVS server

This section concerns what to do if you are having trouble making a connection to a CVS server. If you are running the CVS command line client running on Windows, first upgrade the client to CVS 1.9.12 or later. The error reporting in earlier versions provided much less information about what the problem was. If the client is non-Windows, CVS 1.9 should be fine.

If the error messages are not sufficient to track down the problem, the next steps depend largely on which access method you are using.

:ext:

Try running the rsh program from the command line. For example: "rsh servername cvs -v" should print CVS version information. If this doesn't work, you need to fix it before you can worry about CVS problems.

:server:

You don't need a command line rsh program to use this access method, but if you have an rsh program around, it may be useful as a debugging tool. Follow the directions given for :ext:.

:pserver:

Errors along the lines of "connection refused" typically indicate that inetd isn't even listening for connections on port 2401 whereas errors like "connection reset by peer" or "recv() from server: EOF" typically indicate that inetd is listening for connections but is unable to start CVS (this is frequently caused by having an incorrect path in `inetd.conf'). "unrecognized auth response" errors are caused by a bad command line in `inetd.conf', typically an invalid option or forgetting to put the `pserver' command at the end of the line. Another less common problem is invisible control characters that your editor "helpfully" added without you noticing.

One good debugging tool is to "telnet servername 2401". After connecting, send any text (for example "foo" followed by return). If CVS is working correctly, it will respond with

cvs [pserver aborted]: bad auth protocol start: foo

If instead you get:

Usage: cvs [cvs-options] command [command-options-and-arguments] ...

then you're missing the `pserver' command at the end of the line in `inetd.conf'; check to make sure that the entire command is on one line and that it's complete.

Likewise, if you get something like:

Unknown command: `pserved' CVS commands are: add Add a new file/directory to the repository ...

then you've misspelled `pserver' in some way. If it isn't obvious, check for invisible control characters (particularly carriage returns) in `inetd.conf'.

If it fails to work at all, then make sure inetd is working right. Change the invocation in `inetd.conf' to run the echo program instead of cvs. For example:

2401 stream tcp nowait root /bin/echo echo hello

After making that change and instructing inetd to re-read its configuration file, "telnet servername 2401" should show you the text hello and then the server should close the connection. If this doesn't work, you need to fix it before you can worry about CVS problems.

On AIX systems, the system will often have its own program trying to use port 2401. This is AIX's problem in the sense that port 2401 is registered for use with CVS. I hear that there is an AIX patch available to address this problem.

Another good debugging tool is the `-d' (debugging) option to inetd. Consult your system documentation for more information.

If you seem to be connecting but get errors like:

cvs server: cannot open /root/.cvsignore: Permission denied cvs [server aborted]: can't chdir(/root): Permission denied

then you probably haven't specified `-f' in `inetd.conf'.

If you can connect successfully for a while but then can't, you've probably hit inetd's rate limit. (If inetd receives too many requests for the same service in a short period of time, it assumes that something is wrong and temporarily disables the service.) Check your inetd documentation to find out how to adjust the rate limit (some versions of inetd have a single rate limit, others allow you to set the limit for each service separately.)

F.3 Other common problems

Here is a list of problems which do not fit into the above categories. They are in no particular order.

• On Windows, if there is a 30 second or so delay when you run a CVS command, it may mean that you have your home directory set to `C:/', for example (see HOMEDRIVE and HOMEPATH in D. All environment variables which affect CVS). CVS expects the home directory to not end in a slash, for example `C:' or `C:\cvs'.

• If you are running CVS 1.9.18 or older, and cvs update finds a conflict and tries to merge, as described in 10.3 Conflicts example, but doesn't tell you there were conflicts, then you may have an old version of RCS. The easiest solution probably is to upgrade to a current version of CVS, which does not rely on external RCS programs.