> Initializing this variable seems correct at face value, but the > fact that it > wasn't generating any errors from the test script makes me > suspicious that it > wasn't actually being used.
Well, the global is used, and it does brake things. See for example client.c:3618, which started the whole issue. > Also, it shouldn't be using a global > in the first > place - the errors should be returned properly up the function call change > though this doesn't necessarily need to be done now. I agree that globals should never be used, but the whole cvs client side seems to be coded that way, (which in my opinion is really ugly.) > Lastly, I > don't think the > variable needs to be initialized with every call to > start_server() since it is > setting the process return code and if it a single call to CVS > was supposed to > exit with an error code for any reason, that shouldn't change > with additional > server starts. I agree. I am always using the cvs using the command line, so the exit value does not usually matter. It would be better if the whole failure_exit would just go away... > Can you create a script that duplicates this error? One that > fits in sanity.sh > would be best if you can handle it. It would be as simple as this: Create a file that can not be updated at the first pass, then create a file that is "on the way" for some other file, then do cvs update. A file on the way sets the failure_exit flags, and the next pass will never start. What are the conditions, that make a file unpatchabe at the first update pass? > I created an IssueZilla issue to track this. If you haven't > already registered > and requested an observer role on the ccvs project, do so and you can add > yourself to the CC list: > http://ccvs.cvshome.org/issues/show_bug.cgi?id=36 It seems that the web side has a bug. When I am trying to add me as cc for t he issue, it is complaining thet "Login Reguired" even if I have logged in... Vesa > Derek > > -- > Derek Price CVS Solutions Architect ( > http://CVSHome.org ) > mailto:[EMAIL PROTECTED] CollabNet ( http://collab.net ) > -- > I find that the harder I work, the more luck I seem to have. > > - Thomas Jefferson > > Vesa Suontama wrote: > > > Hello, > > > > I noticed that the variable failure_exit in src/client.c > (version 1.1) is > > never cleared, when doing cvs updates. > > > > It sometimes happens that the failure_exit is set around the > line 1790. (For > > example the case that you have a local file that is "on the way"). > > _______________________________________________ Bug-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/bug-cvs
