Community technical support mailing list was retired 2010 and replaced with a professional technical support team. For assistance please contact: Pre-sales Technical support via email to sales@march-hare.com.
Bryce, I've just tried rename half a dozen different times on files with different ages and it always works for me. If you want us to look at this further you will need to provide specific instructions "from scratch" using only cvsnt client 2.5.03.2382. What I did was (I'm using several files in a directory called "asn"): cvs tag -b dev_1_0 asn cvs co -r dev_1_0 dev_1_0_asn sample/asn cd asn cvs edit sample.ini cvs rename sample.ini sample_config.ini cvs edit sample_config.ini notepad sample_config.ini *** change a line/add a comment cd .. cvs ci -m "Changes" asn cvs co -r dev_1_0 dev_1_0b_asn sample/asn The file is checked out in "dev_1_0b_asn" as sample.ini, but in "asn" it is correctly sample_config.ini. I also tried creating revisions on the branch before the rename on trunk, files with revisions on the trunk before and after the branch, and tried with and without the "notepad" step to force a revision - all variations I can find work OK. I am using 2.5.03.2750 on windows xp (client and server) with SSPI protocol. This is a commercial build, but I am unaware of any changes since 2.5.03.2382 that would effect this except perhaps this one (but that is unlikely): http://customer.march-hare.com/webtools/bugzilla/ttshow_bug.cgi?id=4991 Here are the "cvs log" results for a file renamed but no revision (no "notepad" step): sample\dev_1_0e_asn>cvs log ud6params.txt RCS file: /test/sample/asn/ud6params.txt,v Working file: ud6params.txt head: 1.1 branch: locks: strict access list: symbolic names: dev_1_0: 1.1.0.2 keyword substitution: kvx total revisions: 1; selected revisions: 1 description: ---------------------------- revision 1.1 date: 2006/12/14 03:45:48; author: Arthur Barrett; state: Exp; kopt: kvx; co mmitid: e5c4580c8e13fa3; filename: ud6params.txt; Add first edition of Sample Project for u8404 ======================================================================== ===== Here are the "cvs log" results for a file renamed with "notepad" step: sample\dev_1_0e_asn>cvs log sample.ini RCS file: /test/sample/asn/sample.ini,v Working file: sample.ini head: 1.2 branch: locks: strict access list: symbolic names: dev_1_0: 1.1.0.2 keyword substitution: kvx total revisions: 2; selected revisions: 2 description: ---------------------------- revision 1.2 date: 2007/07/10 00:21:56; author: Arthur Barrett; state: Exp; lines: +1 -0; kopt: kvx; commitid: a104692d1241f67; filename: sample_config.ini; changes ---------------------------- revision 1.1 date: 2006/12/14 03:45:48; author: Arthur Barrett; state: Exp; kopt: kvx; co mmitid: e5c4580c8e13fa3; filename: sample.ini; Add first edition of Sample Project for u8404 ======================================================================== ===== Regards, Arthur Barrett -----Original Message----- From: bryce.schober at gmail.com [mailto:bryce.schober at gmail.com] On Behalf Of Bryce Schober Sent: Tuesday, 10 July 2007 9:54 AM To: Arthur Barrett Cc: CVSNT Subject: Re: [cvsnt] rename breakage In an attempt to work around this problem, I re-named the file back to its original file name. After that, client updates received the error "cvs server: nothing known about <original_filename>". Panic ensued. I looked at the repo via viewvc, which showed it still there (with the original filename), as well as the interesting .directory_history file. Looking inside of that file, I didn't think that the differences between revisions made sense. Looking in the revision log of the twice-renamed file didn't show any record of the renames, so I took a random stab in the dark and renamed the .directory_history,v file in the server's repository, pre-pending "BAK". Voila! Everything magically worked! Though I now so no record of either rename (as I expected). So, a couple questions: - Should I run screaming in horror from this kind of workaround, or is it a reasonable solution to my current problem? - What additional data would you like from me to help fix this serious breakage in advertised-working functionality? -- Bryce Schober