Consistency check of repository & knowing whether files are being watched
Hi, is it possible to check the consistency of a repository? Sometimes I modify my repository by hand... something forbidden, I know... Also it would be interesting to know whether there is a way to check whether all files of a given repository are actually being watched... Moving directories in the repository around isn't a big deal, but moving files and forgetting to modify the CVS/fileattr entries can become troublesome, I guess. That is why I am asking. Marko ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Consistency check of repository & knowing whether files are being watched
SORRY IN CASE THIS MESSAGE APPEARED MORE THAN ONCE, LOOKS LIKE THERE IS A PROBLEM WITH SUBMITTING POSTS... - Hi, is it possible to check the consistency of a repository? Sometimes I modify my repository by hand... something forbidden, I know... Also it would be interesting to know whether there is a way to check whether all files of a given repository are actually being watched... Moving directories in the repository around isn't a big deal, but moving files and forgetting to modify the CVS/fileattr entries can become troublesome, I guess. That is why I am asking. Marko ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Checkout from a Linux-based repository with DOS line-endings
Hi, I have a repository located on a Linux cvs server. My source code files are coded in DOS (CR+LF) style. My cvs client is WinCVS. -- Sometimes I take the whole repository home and check the actual source code out under Linux, to later copy it to my Windows partition. This sounds a bit odd, but I don't have a 2nd computer at home to make it the server. The problem: line endings are now of course Unix style (LF). This is a pain, since you'd have to dosify the whole tree, before you can actually work with the sources... -- How to solve this? 2 Possibilities: a) Let WinCVS handle the repository under Windows directly. Checking out might be ok, but committing something could screw everything?! b) Use some magic optional switch unter Linux while checking the tree out... (Question is which one?) Regards, Marko ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Checkout from a Linux-based repository with DOS line-endings
Hi Spiro & Doug, > > It is no problem to move the repository around, if you don't go across > > the border LF/CR+LF. > But it IS important, when transferring a repository, to avoid using an > ASCII transfer that could convert line endings in ,v files... That's exactly what I was afraid of! So I see there is no easy way to overcome this trouble. I guess the only way is which got described by Doug previously. THanks for your responses! Marko ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Checkout from a Linux-based repository with DOS line-endings
Hello Spiro, > So, the line has to be: > > "It is no problem to move the checked out working copy as long as you > don't cross the border LF/CR+LF". > Well, but, perhaps I misunderstand you, I do cross that border! And that's the trouble. My repository (created and maintained on linux' cvs) uses LF, but I'd like to check the source out in CR+LF, which obviously doesn't work so easily... Need to try what Doug recommended. Let WinCVS work on a copy of the repository directly and hope it would take care of the LF->CR+LF conversion. But, as stated before, I am afraid that WinCVS might screw things up when I try to add files or commit changes to that repository since it doesn't know anything about the LF-origin of it... And, of course, as Doug noted, also the filenames might be changed uncontrollably by Windows as well... Tricky task. So, looks like I need either Linux or Windows running on an OS-emulator so that I could use the true linux cvs pserver... OVERKILL... :) marko ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Consistency check of repository & knowing whether files are being watched
Hi Jim & Mark, if I try to write my own script for checking the consistency of a repository, which things I should take into account? I mean, would it be enough to read the content of each dir recursively and then check whether all files got really included in the fileattr file like ---snip--- FThe1stFileName _watched= FThe2ndFileName _watched= ... ---snip--- and check that every subdir-containing directory has the entry ---snip--- D _watched= ---snip--- or are there still other quircks which have to be taken care of. Of course this assumes that the repository is currently not in use, meaning that there are no edited files. Well, one could think about what you once mentioned that there might be entries like this ---snip--- Fwbl804.FIS _watched=;_editors=kaening_flmpc25>Tue May 18 14:14:01 2004 GMT+flmpc25+P:\MIRROR;_watchers=kaening_flmpc25>tedit+tunedit+tcommit ---snip--- So one might detect also editors and watchers optionally... Or even throw them out as someone liked to do... Question: Do you know anything else which might cross my way? No quircks? Regards, Marko ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Checkout from a Linux-based repository with DOS line-endings
Hi Spiro, thanks for your detailed explanations. I once did a similar routine, also using diff and patch. Though, I wanted to avoid this procedure, because it doesn't allow me to do intermediate commits to the repository. I have to work at home with a local copy which gets checked in only at work. This is not really an option if one prefers to work step by step... But anyway, thanks for the info, I guess that could be of value also for other readers of this list! What concerns cygwin: that might be a good try! I have to test that as well. But what troubles me is the mangling of CaSeS in file names. Windows behaves here kind sort of unpredictable... Regards, Marko ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Consistency check of repository & knowing whether files are being watched
Hi Mark, > Nothing comes to mind, but it has been a long day... have you considered > a file that has only been added to a branch or removed from the mainline > (it will be in the Attic/ directory). well, I considerd the Attic/ as a subdir of the repository as all the others. The script could iterate through it as usual. But I see that I was wrong. The attic doesn't have a fileattr! So it's not so easy. That's the kind of snare I expected! :) So, I guess, the processing of the Attics would require a parsing of the ,v-files, since I assume the watch or edit information there. But perhaps I am mistaken. Is it possible to remove a file which is still watched or edited. I guess at least the latter should be impossible. Or is the information about watches/edits lost in such a case? Anyway, those files wouldn't be so important anyway, since they belong in principle to trunks which have state 'dead', right?! That brings me to the question how can I get information via cvs about files being held in attics? Also I do not know what happens if you have several watches on more than one revision of a file... Have to test, but expect that to be simply appended to the according line in the fileattr file together with information which revision is actually in focus. OK, still collecting all these thoughts to get an idea what I have to take care of, since I do not really want to screw up my repositories with a bad designed tool... Looks like it's not so easy to set it up. Greetings, Marko ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
binary files bad idea? why?
Hi, just read that handling of binary files with cvs is a bad idea (or something similar). Well, I guess one has to add that cvs might not be that suitable for the case of binary files, but, nevertheless, you won't run into trouble if you properly define them as of type 'binary' at initial import of add to a repository. Marko ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
RE: binary files bad idea? why?
> Just use the "-kb" option when adding files. That was my point, I just didn't mention that option, sorry. > I don't think CVS handles diff between binary files well... Well, it doesn't, certainly. That can only be done with additional software, of course. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Consistency check of repository & knowing whether files are being watched
forgot this: > Possibly, or it may assume that you have to get tricky in the patterns > used to account for files that are in use. Well, that's the question! Looks like it would be better to consider the tricky case, since most repositories might be in use ;) ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Getting files by CVS version
> Hi, how do I get an older file if I didn't tag it? So say I have > version 1.1 and i just commited version 1.2, how do I checkout version > 1.1 again? update -r 1.1 yourfile You should read the corresponding chapter on "update" in the manual. ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
FAQ-o-matic
Hi, the "FAQ-o-matic" looks quite messy. There are many empty entries, why is that? Marko ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Consistency check of repository & knowing whether files are being watched
Hi Mark, > The CVS/fileattr directory in the parent may list a foo,v file that is > found in Attic/foo,v however, it should not be possible to have > both a ./foo,v and an ./Attic/foo,v in the same tree, so the added > expense is mostly in processing the CVS/fileattr files using a search > path to find the real ,v-file. OK, that's good to know. > > That brings me to the question how can I get information via cvs about > > files being held in attics? > > Shortly, on feature the 'cvs ls' or 'cvs rls' command will be useful to > you... Otherwise, using something like a 'cvs log -r1' may be useful. Strange, my cvs 1.11.5 doesn't have a (r)ls command... But well, by parsing the output of the log you'd be able to find the string "Attic" in the RCS file path... ok > > Also I do not know what happens if you have several watches on more > > than one revision of a file... Have to test, but expect that to be > > simply appended to the according line in the fileattr file together > > with information which revision is actually in focus. > To be honest, I have not tried it. I found in one of my fileattr's a line which contained a "_watchers" and an "_editors" entry just appended to each other separated with a ';'. A further test showed that additional watchers or editors get simply appended to the concerning entry with a separator ','. To determine which watchers and editors are active on a repository shouldn't be too complicated, since you simply would need to issue the "cvs editors" and "cvs watchers" commands and parse the output, and filter out the user names. These should be compared with the actual users in CVSROOT/users. If somebody wouldn't be in that file you could trow a warning. This of course requires a users file which is kept up to date. > Well, the cvs advisory locking ('cvs watch', 'cvs edit', and 'cvs > watchers') commands are not necessarily as well integrated as might be > desirable. If you find bugs, feel free to send patches... I am getting a bit surprised right now... I thought these commands are of crucial value for environments where many people work on one and the same repository... By the way, I found in one of my repositories the following line in a fileattr file Fremove _watchers=kaening_flmpc23>edit+unedit+commit What does this mean? A file named "remove" doesn't exist, apparently. The 'F' should tell me that it is a file, or not?! But there is another file in the corresponding Attic folder, which got removed from the HEAD at a certain time, but its name is not "remove" either! Marko ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: Consistency check of repository & knowing whether files are
Hi Larry, > He said "on feature". 1.11 is the "stable" branch, 1.12 is "feature". I see! I am new to this list, so I didn't know about this terminology! :) > > I am getting a bit surprised right now... I thought these commands are of > > crucial value for environments where many people work on one and the same > > repository... > > No, they're a sop to people who don't really believe in the CVS > philosophy. ;) Thanks, I got the message. Of course you'd be able to live without these commands, I know. But you are right: a sop - that might be the best description... Haha. But I think they are of use if you don't like to cope with conflicts created by working in parallel with a file. ;) But looking at the title of the thread they might also lead to a certain pollution in your repository if you have to many people playing around with watches... :) Regards, Marko ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
To developers: Is there a CVS server incompatibility with CVSNT since last security update?
Hi CVS developers, a few days ago, when SuSE put the latest security update on the web, we, of course, did that crucial update immediately. But, strangely, with our WinCVS and TortoiseCVS clients problems started. It turned out that suddenly everything checked out was "in the way" during next update! I joined the discussion concerning strange behaviour of the "SF cvs servers" (but I guess they are cvs servers on *n*x systems, so perhaps this should be addressed to you?!) on 2004-05-19 which can be found at: http://sourceforge.net/tracker/?func=detail&aid=949583&group_id=10072&atid=110072 where it is eventually claimed that there is a server problem, not a client problem. What do you say to this? Now I use with WinCVS the latest CVSNT 2.0.41a, which is said to have a workaround hotfix to overcome the server troubles. With that it works fine. TortoiseCVS obviously also released a new version. I don't know what is going on here, perhaps you have more specific information (probably from SF) or discuss this already on the developer mailing list... ... I just wanted to let you know about it. Regards, Marko ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: To developers: Is there a CVS server incompatibility with CVSNT since last security update?
Hi Boris, > Hallo marko, > > Wednesday, May 26, 2004, 10:09:10 AM, you wrote: > > m> Hi CVS developers, > > ich würd die aktuellste CVS Version vom Hersteller runterladen und > installieren. Well, that's true. Perhaps I should not use the security-patched cvs of SuSE 8.2, which is actually 1.11.5 (Release 112) but instead go for the latest version which is somewhere above 1.11.16, right? (Can't check it since cvshome.org is still down.) ;) Well, true, might be it's all my own fault then... Thanks for the hint! Marko ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
How safely "move" files from one branch to the other?
Hi, I've a directory with a few files. There exist 2 branches for this directory, say A and B. I checked out the branch A and committed the changes. Only then I noticed that I should have modified branch B instead. How do I safely "move" the files from branch A to branch B? Marko ___ Info-cvs mailing list [EMAIL PROTECTED] http://lists.gnu.org/mailman/listinfo/info-cvs
Re: How safely "move" files from one branch to the other?
Hi Tyler, On Fri, 18 Jun 2004, Tyler wrote: > > How do I safely "move" the files from branch A to branch B? > > For a small change like this, it's easiest to: > > - back out the changes from branch A > > (in A sandbox) > cvs up -j[new rev] -j[previous rev] [list of files] > (verify changes and commit) > > - merge changes into branch B > > (in B sandbox) > cvs up -j[previous rev] -j[new rev] [list of files] > (verify changes and commit) Well, ok, that's a possibility, of course, but it has the drawback that you have all changes which are caused by the simple swapping of branch names reflected in the repository files in a very inefficient manner. I understand that this is the standard way of doing changes like this, though I hoped that there might be another way to force a modification of the branch tag names instead... Regards, Marko ___ Info-cvs mailing list [EMAIL PROTECTED] http://lists.gnu.org/mailman/listinfo/info-cvs
How to tell CVS that it should use the proper date/time
Hi, I wonder how CVS can be convinced to set the original date/time of a file at checkin time when I do a checkout! Marko ___ Info-cvs mailing list [EMAIL PROTECTED] http://lists.gnu.org/mailman/listinfo/info-cvs
Re: How to tell CVS that it should use the proper date/time
Hi Larry, > I'm having a hard time parsing that sentence, but I think that's what it :) sorry for my poor english. > If you mean you want checkout to record the current timestamp of the > file rather than the time of the checkin, that would also break things > like make. Import, however, has the -d option to do just that. that's what I wanted, right. I don't want the checkin time, I really want the actual timestamp of the file! Will check that option -d. Thanks for your help! Marko ___ Info-cvs mailing list [EMAIL PROTECTED] http://lists.gnu.org/mailman/listinfo/info-cvs
Re: How to tell CVS that it should use the proper date/time
Hi Larry, On Thu, 8 Jul 2004, Larry Jones wrote: > If you mean you want checkout to record the current timestamp of the > file rather than the time of the checkin, that would also break things > like make. Import, however, has the -d option to do just that. oops, I couldn't find a -d option in cvs' info docus. Perhaps it's still not supported in 1.11.5?! Marko ___ Info-cvs mailing list [EMAIL PROTECTED] http://lists.gnu.org/mailman/listinfo/info-cvs
Re: How to tell CVS that it should use the proper date/time
hi larry, > > oops, I couldn't find a -d option in cvs' info docus. Perhaps it's still > > not supported in 1.11.5?! > It's mentioned in the Quick Reference, but not in the detailed guide. > It's not new, it's been there "forever". thanks for the hint. Strange that it isn't visible in the info manual! Look like that needs some refinement or update... Anyway, as I see now there doesn't seem to be a comparable option for the add command, though. Only import would allow this, add won't! Am I right? If so, it would be quite strange... Marko ___ Info-cvs mailing list [EMAIL PROTECTED] http://lists.gnu.org/mailman/listinfo/info-cvs
Re: coaxing scientists to use cvs
Hi, I also use cvs for LaTeX sources and have the same problem with the diffs. It always depends on how you format your sources. With code it's easier, much easier. Every line contains a certain piece of code. But text source can't be handled in this simple way. I used to write one-line-paragraphs in Emacs to be able to have automatic line wrapping, but doing so leads to huge ,v-files eventually, because a tiny change in a paragraph triggers a complete substitution of the whole thing... If you do the word wrap yourself you have similar problems, just because of a single letter you can still end up with many diffs all over the place... I'd guess CVS is not so well-suited for this job. Of course keeping the revisions is essential, but for the diffs you'd need something else in that case. I recall a certain discussion on this list about other ways of diffing and custom diff programs, though I have to admit I didn't follow that thread. I guess textual content would be such a case where this would come in handy. Right?! Marko ___ Info-cvs mailing list [EMAIL PROTECTED] http://lists.gnu.org/mailman/listinfo/info-cvs
Re: coaxing scientists to use cvs
> my strategy is: > never refill/reformat existing paragraphs. > This strategy has the advantage > that the diffs reflect better the logical units > that I'm changing. True. I did the same also, but you may end up with pretty scattered pieces of text if you change your sources very often over time. The sources readability itself deminishes... But I guess that's the only chance though... > Emac's 'ediff' is very helpful here. It let's you go through diffs I know. I used that also quite a few times. > merge CONFLICTS. It displays the 2 old versions and the merge result Didn't do merges up to now with this. > 'pcl-cvs' package--a Emacs frontend for CVS--is really wonderful. haven't used pcl-cvs very often up to now, since am mainly working with Windows now... But, well, just to add that: I fully support your ideas concerning keeping sources of LaTeX under CVS' control. I just wanted to underline that there are some tripwires in your path if you do that. I understand your boss, my bosses acted exactly the same way, so that I had to do all the updating myself and find their changes and handle the cases they didn't use the most up-to-date versions... ;) neverending story ;) Happens all the time! :) Marko ___ Info-cvs mailing list [EMAIL PROTECTED] http://lists.gnu.org/mailman/listinfo/info-cvs
how to un-add an added but uncommitted file?
Hi, say I 've added wrongly a file named FILE to a repository. I still haven't done a commit on that one. If I do a cvs remove FILE the files get's thrown away and doesn't belong any longer to the module. If I simply delete the file and do an update on the module I get the warning warning: new-born FILE has disappeared So far so good, both methods seem to "remove the file from my repository", well I know, actually the file is not there yet, but anyway... BUT, I'd like to preserve the file itself! Is there a way of un-adding a file which got by accident added without deleting it?? Marko ___ Info-cvs mailing list [EMAIL PROTECTED] http://lists.gnu.org/mailman/listinfo/info-cvs
RE: how to un-add an added but uncommitted file?
Hi Edwin, > Maybe this is an obviously stupid suggestion but how about in your second > method instead of deleting, you rename it... Then rename it back after you > do the update? OK, that would solve it, but yes, it doesn't really sound like a perfect solution, because I'd have to rename, update and rename again... Strange! :) Marko ___ Info-cvs mailing list [EMAIL PROTECTED] http://lists.gnu.org/mailman/listinfo/info-cvs
RE: how to un-add an added but uncommitted file?
Dear Jim and folks, > Strange is in the eye of the beholder. true! :) > To me, it seems very strange that you want to preserve a file, but not > have it in CVS. Surely if it's that important it should be under version > control? Well, say it's obj-file or an exe which you added just accidentally while going through a bunch of files. You don't want to restart the compiler to recreate them, perhaps... That would be my arguments! :) OK, I see, I have to live with the renaming thing. :) Not that I am totally against it, it's only that I thought there might have been a cleaner way to achieve that! :| Thanks for your responses! Marko ___ Info-cvs mailing list [EMAIL PROTECTED] http://lists.gnu.org/mailman/listinfo/info-cvs
Re: Problem to delete CVS directories...
Dear Mr. Sanchez, we had the same problem and came to the conclusion that the only way to delete directories from a modul as well as deleting a whole modul is to be done via filesystem on the unix machine, i.e. log in on your unix box and change to the repository directory, then delete the respective subdirs you want to get rid of. Fortunately, the rest of the CVS system is not affected. In advance, -- Marko Faldix M+R Infosysteme Hubert-Wienen-Str. 24 52070 Aachen Tel.: 0241-93878-16 Fax.:0241-875095 E-Mail: [EMAIL PROTECTED] "Hector Sanchez-Villeda" <[EMAIL PROTECTED]> schrieb im Newsbeitrag news:[EMAIL PROTECTED]... > > Dear sirs, > > > > I have installed CVS in one of my UNIX boxes, then I included several > > directories for testing purposes and I tryed to delete them > permanently > > according to CVS documentation, but it appears that the directories > > and files went to the ATIC directory. > > > > How can I get rid of them? > > > > Any help is appreciated. > > > > PD. My E-Mail is [EMAIL PROTECTED] > > > > Sincerely > > > > Hector Sanchez > > University of Missouri > > > > ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
Re: HELP! Howto disable binary flags on files
"Josh Baudhuin" <[EMAIL PROTECTED]> schrieb im Newsbeitrag news:[EMAIL PROTECTED]... > Well, you can use cvs admin -k on the existing files, but is it > possible there's some systematic problem which caused the files to get > created -kb? (Or was it just a user error?) > > cvs admin -kkv > > This is one of those commands that can be done recursively, so be > carefule not to undo the files... Hello, where do you type in the cvs admin commands. As user on clients? I tried in unix command line in the repository directory on the server, but there was no file "entries" found, which normally resides on the client after checkout. But on the client, we have NT 4.0 with jCVS an the cvs admin command did not change anything. So where is the command line for this commands? -- Marko Faldix M+R Infosysteme Hubert-Wienen-Str. 24 52070 Aachen Tel.: 0241-93878-16 Fax.:0241-875095 E-Mail: [EMAIL PROTECTED] ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
giving up CVS
Hello, we tried to use jCVS for * LARGE * directory trees consisting of html files and binary files like .gif, .jpg and so on. We consider giving up cvs for our web projects because of the number of problems with large directory trees with mixed files (binary and text). We had binaries which occured as text and so they couldn't be repaired anymore. I studied several days cvs and found out, that binaries can only while importing handled as binary. If forgotten to type in all binary types during import you've lost. Adding a file is a problem for us. If one of us works a day, he will have to add whole parts of new directory trees or many new files - binary and text - in different subdirs. Adding as binary and adding as text is for our purpose very uncomfortable and here is again loss of data in cvs possible because you have to handle cvs which so much care. It was recommended to us to use cvs, but I think by an old unix C programmer who only has to deal with one single text file. What we are looking for is something like CVS but combined with saving data. Using cvs became a safety problem to us. Everything that is stored in cvs has to be saved on third media before. What do you think about that? Are there any web programmers here using cvs with success and how to you save data? -- Marko Faldix M+R Infosysteme Hubert-Wienen-Str. 24 52070 Aachen Tel.: 0241-93878-16 Fax.:0241-875095 E-Mail: [EMAIL PROTECTED] ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs
error writing to lock file (IBM AIX)
I am maintaining a large cvs server with a few repositories on there. Each repository is on it's own logical filesystem (IBM AIX box) In one of the repositories i keep getting the following problem No locks are set. cvs [server aborted]: error writing to lock file /batcvs/impl/brio/,filename.bin, This is a binary file that is < 1MB but the server version is about 20 MB with approx 20 revisions at a time. This is the only file this happens on and before what i had to do is rename the file to filename_old.bin and then do a fresh checkin of the new version It is not a diskspace issuelots of room all accross the systemincluding the /tmp. Any help is greatelly appreciated!!! ___ Info-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/info-cvs