Hello, On Thursday 14 June 2007 21:16, Eric Y. Kow wrote: > ... > On Thu, Jun 14, 2007 at 10:18:26 +0200, Thorkil Naur wrote: > > I believe that my ability to do this was important to convince my ISP to look > > hard at this problem. So I would suggest such events to be taken into > > consideration when deciding whether alternative means of communicating should > > be retained. > > Would the loss of this functionality be acceptable if darcs had some > kind of extra verbose mode, where it printed out all the patches it was > trying to get? Or would calling wget still be useful/more flexible for > other reasons? > ...
In case of something still as brittle as Internet connectivity, I would always be careful not to limit my options needlessly. Clearly, in this case, darcs was fortunate enough to be able to play a major role when debugging my Internet connection, but I certainly don't mean to imply that darcs (or any other program that simply uses the Internet for its purposes) should include all sorts of network debugging mechanisms. Printing the names of patches (and other files downloaded) would be useful if it was clear and known that what was happening was a combined operation of downloading a sequence of files. This may be the case, I don't know. The difference between darcs using libcurl and darcs using wget was that I was able (using the error message from darcs on my killing the process that was running wget) to see the name of the file involved. With darcs using libcurl, I had to kill darcs itself, generating no messages or other indications of what it was doing. Summarizing, I don't think you should let this particular problem enter your considerations with any significant weight. But I do believe that you should retain working and useful options in this area. This is also what seems to be the direction that this matter is taking. Thanks and best regards Thorkil _______________________________________________ darcs-devel mailing list [email protected] http://lists.osuosl.org/mailman/listinfo/darcs-devel
