part of comment on b78a720cee:
The only reason it's getting into master is because it looks cool!
I hate it when something that looks cool doesn't work right :(
creating a repo on gitolite-admin push is *needed* in order to get
descriptions and export-ok files to work right
Summary: much as I did not want to use "excludes", I guess if we don't put the
code in "master" it's OK to at least *write* (and test) the code!
See the example config file for how to use it.
See "design choices" section in the "faq, tips, etc" document for how it
works.
the "create a new repo" code moves from compile to auth.
Only someone who has W access can create it, but he can do so even on a
"R" operation (like clone or ls-remote).
This is a pre-requisite for rebel's wildcard repos, where
autovivification is the only way you can create arbitrary repos matching
a pattern.
The only reason it's getting into master is because it looks cool!
----
OK that's a lie; the real reason is to keep the two branches as similar
as possible, though they;ve diverged quite a bit since the "only
one-line difference" days where "rebel" just meant "deny/exclude"
rules!)
writing the export_ok files and the gitweb project list are now
unconditional. They're idempotent anyway, and I doubt anyone cared
about all the fancy logic to detect and report *just* the new ones on
each compile.
This paves the way for gitweb ownership to be added later; that code was
becoming too complex otherwise...
- refuse to install to root
- when a pubkey is being used that was not freshly created by
ourselves, warn the user that this key can not be used to get shell
access to the server. Prevents some corner cases of people being
locked out...
Also, change the final message to be even more clear that this is all on
the workstation, not the server
I don't have a use for "@all" at all (pun not intended!) other than the
"testing" repo, but <teemu dot matilainen at iki dot fi> sent in a patch
to mark those repos with "R" and "W" in the permissions list, and I
started thinking about it.
This could actually be useful if we *differentiated* such access from
normal (explicit username) access. From the "corporate environment"
angle, it would be nice if a project manager could quickly check if any
of his projects have erroneously been made accessible by @all.
So what we do now is print "@" in the corresponding column if "@all" has
the corresponding access.
Also, when someone has access both as himself *and* via @all, we print
the "@"; printing the "R" or "W" would hide the "@", and wouldn't
correctly satisfy the use case described above.
We detect an upgrade situation by the presence of
$GL_ADMINDIR/conf/gitolite.conf -- if it exists, we reason, this is not
a fresh install. And if so we skip setting up PTA, and the initial
clone.
Well, turns out this is not always true. I've had a few cases where the
first install didn't go right, but left enough stuff in to make the
subsequent attempt think this is an upgrade.
[This mostly happened to me when I was testing the "oldgits" branch, and
also when I was making it work from msysgit I think... regardless of
why, it'd be good to fix]
So this changes the flow somewhat. Now the *only* difference between a
fresh install and an ugrade is the "initial_conf_key" function call (you
don't want to overwrite an existing conf file or keydir!)
I was trying to determine how close gitolite can come to the ACL model
of a proprietary product called codebeamer, and one of the items was how
to make a "role" (like QA_Lead) have different "members" in different
projects.
I then realised delegation already does that! Which is great, but as I
thought about it more, I realised... well, we'll let the in-code
comments speak for themselves :-)
Anyway, all it needed was a 1-line fix, luckily... <phew> And it would
have only affected people who use delegation.
- all $HOME/blah becomes "$HOME/blah" (bl**dy "Documents and Settings" crap)
- replace bash regex with perl, and in one case replace the check with
something else
- rsync changed to appropriate scp
- since we no longer insist on running from a specific directory, create
tmpgli dir *after* you cd to the right place
You can now add your own hooks into src/hooks/ and they get propagated
along with the update hook that is present there now. Please read the
new section in the admin document, and make sure you understand the
security implications of accidentally fiddling with the "update" script.
This also prompted a major rename spree of all the files to be
consistent, etc. Plus people said that the .sh and .pl suffixes should
be avoided (and I was feeling the same way). I've also been
inconsistent with that "gl-" prefix, so I cleaned that up, and the 00-
and 99- were also funny animals.
Time to get all this cleaned up before we get 1.0 :)
So these are the changes, in case you're looking at just the commit
message and not the diffstat:
src/pta-hook.sh -> src/ga-post-update-hook
src/conf-convert.pl -> src/gl-conf-convert
src/00-easy-install.sh -> src/gl-easy-install
src/99-emergency-addkey.sh -> src/gl-emergency-addkey
src/install.pl -> src/gl-install
src/update-hook.pl -> src/hooks/update
**upgrades no longer touch the config or the keydir**
When you first install gitolite, the easy install script has to do two
*distinct* things:
* install the software
* create and seed the gitolite-admin repo with a minimum config file
and the newly created pubkey
That's fine for an install, because nothing exists yet anyway.
Subsequent invocations of the script should only do the first task (so
that gitolite itself can be upgraded), and not attempt to fiddle with
the config file and pubkeys.
Unfortunately, until now I had not been separating these two activities
cleanly enough. For instance, the commit message for 8e47e01 said:
IMPORTANT: we assume that $admin_name remains the same in an upgrade
-- that's how we detect it is an upgrade! Change that name or his
pubkey, and you're toast!
Ouch!
So now I decided to clean things up. The "Usage" message tells you
clearly what to do for an upgrade.
Should have been like this from the beginning, but hey we got there
eventually :)
----
Code-wise, this is a major refactor of the easy install script. It uses
an old forgotten trick to get forward refs for bash functions ;-) and in
the process cleans up the flow quite a bit.
- "it's an upgrade" is decided by presence of gitolite.conf (not a pubkey)
- admin_name optional (and will be ignored if given) for upgrades
plus a lot of comments and some minor text changes
Packaging gitolite for debian requires the rc file to be in /etc/gitolite.
But non-root installs must still be supported, and they need it in $HOME.
This means the rc file is no longer in a fixed place, which needs code to find
the rc file first. See comments inside new file 'gitolite.pm' for details.
The rest of the changes are in the other programs, to replace the hard-coded
rc filename with a call to this new code.
- example config file is now all comments (should have been that way anyway)
- we detect if it is an upgrade and act accordingly (see below)
IMPORTANT: we assume that $admin_name remains the same in an upgrade -- that's
how we detect it is an upgrade! Change that name or his pubkey, and you're
toast!
- added comments to easy install to help do it manually
- README: some stuff moved to tips doc, brief summary of extras
(over gitosis) added
- INSTALL: major revamp, easy install and manual install,
much shorter and much more readable!
plus other docs changed as needed, and updated the tips doc to roll in
some details from "update.mkd" in the "ml" branch
- add better comments on the 2 main hashes
- work around an inefficiency caused by the exclude prep code needing
a list instead of a hash at a certain place
[Note: this is a fairly involved commit, compared to most of the others.
See doc/5-delegation.mkd for a user-level feature description.]
parse delegated config fragments (found as conf/fragments/*.conf). Any
repos being referenced within a fragment config *must* belong to the
"@group" with the same name as the fragment.
That is, a fragment called conf/fragments/abc.conf can only refer to repos
that are members of the "@abc" repo group. It cannot specify access
control for any other repos. If it does, those settings are ignored, and
a warning message is produced.
since the delegated config must have the flexibility of (re-)defining
group names for internal convenience, and since all such definitions go
into the same "groups" hash, it is quite easy for conf/fragments/abc.conf
to write in its own (re-)definition of "@abc"! That would be a neat
little security hole :)
The way to close it is to consider only members of the "@abc" groupset
defined in the main ("master") config file for this purpose.
collect the delegated config fragments from correspondingly named branches of
the gitolite-admin repo, and put them all in conf/fragments/
also deprecate changes to conf and keydir locations from now on
Again, prep for delegation, when we'll be reading fragments of config rules
from various files and tacking them onto the %repos hash.
note: this patch best viewed with "git diff -w", clicking "Ignore space
change" in gitk, or eqvt :-)
- warn about files in keydir/ that dont end with ".pub"
- warn about pubkey files for which the user is not mentioned in config
- warn more sternly about the opposite (user in config, no pubkey!)
update hook: add reponame to message on deny
auth: minor typo
Summary:
DONT forget to run src/gl-compile-conf as the last step in the upgrade
Details:
The compiled file format has changed quite a bit, to make it easier for the
rebel edition coming up :-)
compile:
- we don't split RW/RW+ into individual perms anymore
- we store the info required for the first level check separately now:
(repo, R/W, user)
- the order for second level check is now:
repo, user, [{ref=>perms}...] (list of hashes)
update hook logic: the first refex that:
- matches the incoming ref, AND
- contains the perm you're trying to use,
causes the match loop to exit with success. Fallthrough is failure
- detect/warn git version < 1.6.2
- create documentation with details on client-side workaround
- change the "git init --bare" to (older) "git --bare init", since the old
syntax still works anyway
why should just usernames have all the fun :) The "expand_userlist" function
is now "expand_list" and serves generically. The example conf has also been
updated correspondingly
(thanks to SethX for feedback)
- install: a little more verbosity in the mkdir
- install and example conf: some of the help text made more clear
- auth: error message on bad $cmd is now clearer, plus no perl-warnings to
confuse people
- logs go into $GL_ADMINDIR/logs by default, named by year-month
- logfile name template (including dir prefix) now in $GL_LOGT
- two new env vars passed down: GL_TS and GL_LOG (timestamp, logfilename)
- log messages timestamps more compact, fields tab-delimited
- old and new SHAs cut to 14 characters
system("...") run from perl on sol does not seem to like "~" (regardless of
what $SHELL is set to), so use $ENV{HOME} instead
thanks again to evocallaghan
- install.sh is now install.pl (had to happen sooner or later!)
- now handles updates more gracefully, doesn't overwrite important stuff :)
- makes the install sequence much easier to understand
(just run it and follow the prompts!)
- made ~/.gitolite.rc much clearer to edit
In the "create new repos" loop, we need an absolute value for REPO_BASE, in
order to be able to chdir back and forth. But (taking the "normal user with
no privileges" assumption too far!) we assumed REPO_BASE would be within
$HOME, and relative to it. So it fails when someone wants the repo_base
elsewhere.
Now we don't prefix $HOME if REPO_BASE is already absolute (begins with a "/")
bug reported by evocallaghan
