Re: [Monotone-devel] [ANNOUNCE] monotone 0.22
On Wed, 2005-08-10 at 14:05 +1200, Matthew Gregan wrote: We shouldn't be able to end up in a situation where the .mt-attrs file has records in it for files that don't exist in that revision... perhaps there is not enough sanity checking being done on the .mt-attrs file. If anyone attempts to perform stronger error checking on .mt-attrs then please remember that this file is used not only to manage the execute bit but also to create empty directories when using a 'broken' build system that requires empty directories to exist. Since it bolts straight into lua it can be used for a whole bunch of other weird and confusing things as well. -- Daniel Thompson (STMicroelectronics) [EMAIL PROTECTED] 1000 Aztec West, Almondsbury, Bristol, BS32 4SQ. 01454 462659 If a car is a horseless carriage then is a motorcycle a horseless horse? ___ Monotone-devel mailing list Monotone-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/monotone-devel
Re: [Monotone-devel] [ANNOUNCE] monotone 0.22
On Wed, Aug 10, 2005 at 08:34:24AM +0100, Daniel THOMPSON wrote: On Wed, 2005-08-10 at 14:05 +1200, Matthew Gregan wrote: We shouldn't be able to end up in a situation where the .mt-attrs file has records in it for files that don't exist in that revision... perhaps there is not enough sanity checking being done on the .mt-attrs file. If anyone attempts to perform stronger error checking on .mt-attrs then please remember that this file is used not only to manage the execute bit but also to create empty directories when using a 'broken' build system that requires empty directories to exist. Since it bolts straight into lua it can be used for a whole bunch of other weird and confusing things as well. In the .rewrites.change_set branch that graydon's been working on, there is currently both empty directory support, and attributes moved from .mt-attrs into manifests/changesets. This branch doesn't even _work_ yet, but it is a hint at how things will likely end up :-). This means that ATM you won't be able to put attributes on things that don't exist, but OTOH you won't need them for this particular hack. Did you have any other hacks in mind that this would break? -- Nathaniel -- So let us espouse a less contested notion of truth and falsehood, even if it is philosophically debatable (if we listen to philosophers, we must debate everything, and there would be no end to the discussion). -- Serendipities, Umberto Eco ___ Monotone-devel mailing list Monotone-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/monotone-devel
Re: [Monotone-devel] [ANNOUNCE] monotone 0.22
I guess today is not my day :-). On Tue, Aug 09, 2005 at 01:53:17AM -0700, Nathaniel Smith wrote: Just make this: function get_netsync_read_permitted(keyid) conditionals here return false end This, of course, is missing an argument and should be: function get_netsync_read_permitted(branch, keyid) conditionals here return false end Tue May 3 00:31:37 PDT 2005 0.19 release. performance improvements, features, ui improvements, and bug fixes. And this, of course, is horribly out of date. Teach me to forget which computer my mutt session is running on... http://venge.net/monotone/NEWS is up to date, as always. -- Nathaniel -- Eternity is very long, especially towards the end. -- Woody Allen ___ Monotone-devel mailing list Monotone-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/monotone-devel
Re: [Monotone-devel] [ANNOUNCE] monotone 0.22
On 8/9/05, Nathaniel Smith [EMAIL PROTECTED] wrote: I'm please to announce that monotone 0.22 is now available at http://venge.net/monotone/. As always, more binary packages will appear mysteriously from the ether over the next day or two. Looks like someone got it updated in Gentoo already as well. -- Justin Patrin ___ Monotone-devel mailing list Monotone-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/monotone-devel
Re: [Monotone-devel] [ANNOUNCE] monotone 0.22
Hi Myself and a colleague seeing a bug with FC4 (binary* and compiled from source) using 0.22: $ monotone --db=my.db co --branch=default ... lots of lines... monotone: binding 1 parameters for SELECT id FROM files WHERE id = ? monotone: binding 1 with value '99c97df73594ad6da25efa38dbd3ecbe62236e75' monotone: binding 1 parameters for SELECT data FROM files WHERE id = ? monotone: binding 1 with value '99c97df73594ad6da25efa38dbd3ecbe62236e75' monotone: version cache expiring 2efe1d7ce337d5316d4c604068ce1feffb352fad monotone: version cache expiring 50cd2617410d3e05c20056392a2aa1f7c4193044 monotone: version cache expiring 6ee923712be02fb51081a53a327df24135f6042e monotone: work path is MT/work monotone: attribute map path is .mt-attrs monotone: unix/process.cc:61: usage constraint 'N(fd != -1)' violated terminate called after throwing an instance of 'informative_failure' monotone: fatal: signal: SIGABRT (application abort requested) monotone: this is almost certainly a bug in monotone. monotone: please send this error message, the output of 'monotone --full-version', monotone: and a description of what you were doing to [EMAIL PROTECTED]: discarding debug log (maybe you want --debug or --dump?) The error doesn't occur with 0.21. Any ideas? Thanks! Conrad * The link on the monotone web front page to the FC4 binary rpm is broken as well: http://www.venge.net/monotone/downloads/monotone-0.22-1.i386.rpm should be http://www.venge.net/monotone/downloads/monotone-0.22-1_fc4.i386.rpm On Tue, 2005-08-09 at 01:53 -0700, Nathaniel Smith wrote: I'm please to announce that monotone 0.22 is now available at http://venge.net/monotone/. As always, more binary packages will appear mysteriously from the ether over the next day or two. The biggest change in this release is the switch of our underlying crypto library from crypto++ to botan. This should have no user-visible effects, but let us know if you see anything weird. Otherwise, this is mostly a bug fix and smoothing-things-out release. Some error messages are a little more informative, 'merge' and 'propagate' now take an '--lca' option as a convenience to avoid explicit_merge until we get merging fixed properly, etc. The biggest change is that the get_netsync_anonymous_read_permitted and get_netsync_read_permitted hooks have been merged; experience showed them to be very confusing. When upgrading to 0.22, server operators should: -- remove get_netsync_anonymous_read_permitted -- if necessary, adjust get_netsync_read_permitted to take into account the possibility of receiving the Lua special value 'nil' as a key id. If you allow anonymous reads, you probably have: function get_netsync_anonymous_read_permitted(branch) return true end function get_netsync_read_permitted(branch, keyid) return true end Just make it: function get_netsync_read_permitted(branch, keyid) return true end If you don't, you probably have: function get_netsync_anonymous_read_permitted(branch) return false end function get_netsync_read_permitted(branch, keyid) conditionals here return false end Just make this: function get_netsync_read_permitted(keyid) conditionals here return false end Or, to be extra certain: function get_netsync_read_permitted(branch, keyid) if keyid == nil then return false end conditionals here return false end (Of course, if you're upgrading from 0.20, you also need to delete the branch argument from you get_netsync_write_permitted hook or you will get mysterious permission denied messages.) Enjoy, -- Nathaniel ___ Monotone-devel mailing list Monotone-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/monotone-devel -- Conrad Steenberg [EMAIL PROTECTED] California Institute of technology smime.p7s Description: S/MIME cryptographic signature ___ Monotone-devel mailing list Monotone-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/monotone-devel