largefiles: fix explicit commit of normal/largefile switch Commit of corresponding normal/largefiles pairs would only commit the standin. That is usually fine, except if either the normal file or the standin is a remove while the other is an add. In that case it would either give duplicate colliding entries or lose the file. Instead, commit both filenames if one of them is a remove.

largefiles: test coverage of explicit commit of normal/largefile switch This reveals that a switch from normal to largefile violates the normal largefile invariant and gives a manifest with both a normal and standin file in the repository, while a switch from largefile to normal gives nothing. Not good.

Added signature for changeset b66e3ca0b90c

Added tag 3.6-rc for changeset b66e3ca0b90c

merge default into stable for code freeze

convert: fix Python syntax in 'splice in' message Instead of reporting spliced in ['82544090e14fe18091e04f1fb0f0d7991cbe6e7e'] as parents of 369fd983d9e13330e9f12d9fce820deae84ea223 report spliced in 82544090e14fe18091e04f1fb0f0d7991cbe6e7e as parents of 369fd983d9e13330e9f12d9fce820deae84ea223

rebase: fix wrong 'no changes to commit' when using --collapse --collapse will do that rebase doesn't commit until the final commit. The lack of a new commit would make it look like the rebase didn't contribute any changes. Instead, only warn about no commits when not using --collapse.

pull: all pass along extra opargs Wihtout this patch, commands.pull silently drops opargs defeating the whole purpose of the argument.

push: all pass along opargs Wihtout this patch, commands.push silently drops opargs defeating the whole purpose of the argument.

check-code: allow argument passing py2.6ism this backs out 131f7fe06e9e, because Python2.5 support was dropped

mail: drop python 2.5 self.sock.read workaround

url: drop support for python2.5

win32: drop reference to python2.5

editmerge: dequote other use of $ED We want to support editors with parameters, eg EDITOR="vim -O" or whatever. So remove the quotes from around $ED and assume that the editor variable is properly escaped already.

bundlerepo: properly extract compressed changegroup from bundle2 Before this bundle repository were unable to work with compressed bundle2. We use the same approach as with bundle1, we extract the changegroup in uncompressed form into a temporary file.

bundle2: make unbundle.compressed return True when compressed We were returning 'False' in all cases, even when the bundle2 was actually compressed.

bundlerepo: uncompress changegroup in bundle1 case only Uncompressing bundle2 needs to be handled differently.

bundlerepo: move temp-bundle writing logic into a closure We will reuse this logic for bundle2

shelve: properly process bundle2 bundle

histedit: properly apply bundle2 backups If the histedit fails, we restore a backup. We make sure this backup bundle can be in bundle2 format as general delta will require.

strip: pass source and url to bundle2 processing Restoring from a 'bundle2' was missing this data.

unbundle: use 'url' argument with applybundle This will let the hook get access to the 'url' value.

applybundle: take url as argument We allow specifying the url to carry it to hooks. This gets us closer to 'bundle1.apply(...)' and will allow us to remove regressions in multiple place where we forget to pass the url to hooks.

unbundle: use 'source' argument with applybundle This will let the hook get access to the 'source' value.

applybundle: take source as argument We allow specifying the source to carry it to hooks. This gets us closer to 'bundle1.apply(...)' and will allow us to remove regressions in multiple places where we forget to pass the source to hooks.

applybundle: set 'bundle2=1' env for all transaction This should be set for all bundle2 application, we enforce that at a low level. This is for courtesy with hooks.

unbundle: use 'bundle2.applybundle' This is one such place.

bundle2: introduce an "applybundle" function There is a case where the intent is clear and the transaction is not optional. We want to be able to alter that transaction in a wide and easy way. We cannot get a unified '.apply(repo)' method for bundle1 and bundle2 yet because the api are still a bit too far apart. But this is a good step forward to get the rc out.

commands.resolve: conclude merge driver if no unresolved files are left This can happen when either 'hg resolve --all' is called or a driver-resolved file is explicitly requested. This is done as part of 'hg resolve --all' so that users still have a chance to test their changes before committing them. The exact semantics here are still to be decided. This does not impact any non-experimental features. Thanks to Pierre-Yves David for some advice about this behavior in particular, and merge drivers in general.

commands.resolve: call driverpreprocess if we haven't run it yet We need to be careful about allowing --mark and --unmark to keep working -- we don't want the user to be stuck in a weird state. The exact behavior here is still to be decided, though.