"CVR" == Chuq Von Rospach chuqui@plaidworks.com writes:
CVR> I just updated to the current CVS, and it's broken: CVR> Traceback (innermost last): CVR> File "/home/mailman/cron/qrunner", line 85, in ? from CVR> Mailman import MailList File CVR> "/home/mailman/Mailman/MailList.py", line 45, in ? from CVR> Mailman.Archiver import Archiver File CVR> "/home/mailman/Mailman/Archiver/__init__.py", line 17, in ? CVR> from Archiver import * File CVR> "/home/mailman/Mailman/Archiver/Archiver.py", line 33, in ? CVR> from Mailman import Mailbox File CVR> "/home/mailman/Mailman/Mailbox.py", line 43 <<<<<<< CVR> Mailbox.py ^ CVR> SyntaxError: invalid syntax
When you updated, did you see a file that got checked out with a `C' status letter? That means that something in your local version conflicts with some changes in the master copy, and CVS can't resolve it. You have to resolve the conflicts manually.
Yeah, I hate that CVS is so silent about that. I've taken to doing "find . -type f | xargs grep '<<<<<'" every time I do "cvs update" (in a script). The stuff it leaves behind is *clearly* not usable, so I think there should be a big huge warning when it happens, and even an "invoke $EDITOR on the manual-merge-needy". But that's me.
At 3:20 PM -0700 10/3/00, Dan Mick wrote:
Yeah, I hate that CVS is so silent about that. I've taken to doing "find . -type f | xargs grep '<<<<<'" every time I do "cvs update" (in a script).
good idea. Thanks, Dan. I made one other mistake -- I'm so used to having no problems that I dind't checkpoint either my lists or the source before updating. Oops. Fortunately, I *also* keep a generic, unpatched version of the source, so I was able to checkpoint that, do the CVS update, and then compare changes, which made the problem obvious.
Always, always, keep backups when hacking without a net. Or with a net. Nets fail, too. (grin)
-- Chuq Von Rospach - Plaidworks Consulting (mailto:chuqui@plaidworks.com) Apple Mail List Gnome (mailto:chuq@apple.com)
You seem a decent fellow. I hate to die.
"CVR" == Chuq Von Rospach chuqui@plaidworks.com writes:
CVR> Always, always, keep backups when hacking without a net. Or
CVR> with a net. Nets fail, too. (grin)
Heh. I almost wasn't going to mention it, but I did a stupid thing yesterday. I accidently deleted my local working directory. :( Now, I don't think I lost much because of course, there's a baseline in CVS. Fortunately the files I had changed but not yet checked in were still in my XEmacs buffers, so I went to each and saved them out just to be sure. Like a big dummy, I didn't have the directories backed up, so the one thing I lost is my extensive bounce collection, which I use as a regression test. :(
Oh well. Now if I could just get my zip drive working under Linux, I'd be set.
-Barry
On Tue, Oct 03, 2000 at 03:20:16PM -0700, Dan Mick wrote:
CVR> "/home/mailman/Mailman/Mailbox.py", line 43 <<<<<<<
Yeah, I hate that CVS is so silent about that. I've taken to doing "find . -type f | xargs grep '<<<<<'" every time I do "cvs update" (in a script). The stuff it leaves behind is *clearly* not usable, so I think there should be a big huge warning when it happens, and even an "invoke $EDITOR on the manual-merge-needy". But that's me.
It is pretty loud about it, actually. My cvs client shows 3 or 4 lines for each failed merge ;) What I usually do, is 'cvs update' twice, so the 'C' stands out more the second time. Alternatively, you can look for .#-files, which CVS creates for conflicts.
-- Thomas Wouters thomas@xs4all.net
Hi! I'm a .signature virus! copy me into your .signature file to help me spread!
participants (4)
-
bwarsaw@beopen.com
-
Chuq Von Rospach
-
Dan Mick
-
Thomas Wouters