Durham Goode <durham@fb.com> [Tue, 20 Sep 2016 12:24:01 -0700] rev 29963
manifest: move treeinmem onto manifestlog
A previous patched moved all the serialization related options onto
manifestrevlog (since it is responsible for serialization). Let's move the
treeinmem option on manifestlog, since it is responsible for materialization
decisions. This reduces the number of dependencies manifestlog has on the old
manifest type as well, so we can eventually make them completely independent of
each other.
Augie Fackler <augie@google.com> [Mon, 19 Sep 2016 17:14:43 -0400] rev 29962
copy: document current behavior of 'hg cp --after'
I'm about to propose an output change here, but the existing behavior
was untested!
Nathan Goldbaum <ngoldbau@illinois.edu> [Tue, 20 Sep 2016 10:03:50 -0500] rev 29961
crecord: add an event that scrolls the selected line to the top of the screen
Using ctrl-l for this purpose seems to be a fairly widely used practice,
presumably following emacs. This doesn't scroll the selected line all
the way to the top of the window, instead it leaves a 3 line buffer for
context. Use curses.unctrl() to resolve keypressed to '^L' to avoid
hard-coding hexadecimal key codes.
Yuya Nishihara <yuya@tcha.org> [Tue, 03 May 2016 14:24:00 +0900] rev 29960
log: drop hack to fix order of revset (issue5100)
Specify ordered=revset.followorder instead.
This patch effectively backs out c407583cf5f6. revs.sort(reverse=True)
is replaced by revs.reverse() because the matcher should no longer reorder
revisions.
Yuya Nishihara <yuya@tcha.org> [Tue, 03 May 2016 14:18:28 +0900] rev 29959
revset: add option to make matcher takes the ordering of the input set
This allows us to evaluate match(subset) as if 'subset & expr', which will
be the complete fix for the issue5100.
Martin von Zweigbergk <martinvonz@google.com> [Mon, 19 Sep 2016 09:14:35 -0700] rev 29958
strip: don't use "full" and "partial" to describe bundles
The partial bundle is not a subset of the full bundle, and the full
bundle is not full in any way that i see. The most obvious
interpretation of "full" I can think of is that it has all commits
back to the null revision, but that is not what the "full" bundle
is. The "full" bundle is simply a backup of what the user asked us to
strip (unless --no-backup). The "partial" bundle contains the
revisions we temporarily stripped because they had higher revision
numbers that some commit that the user asked us to strip.
The "full" bundle is already called "backup" in the code, so let's use
that in user-facing messages too. Let's call the "partial" bundle
"temporary" in the code.
Martin von Zweigbergk <martinvonz@google.com> [Mon, 19 Sep 2016 09:14:32 -0700] rev 29957
strip: clarify that user action is required to recover temp bundle
If strip fails when applying the temporary bundle, the commits in the
temporary bundle have not yet been applied, so the user will almost
definitely want to apply the bundle. We should be more clear to the
user about that than our current "partial bundle stored in...".
Note that we will probably not be able to recover it automatically,
since whatever made it fail (e.g. a hook) will most likely make it
fail again. We need to give control back to the user to fix the
problem before trying again.
Martin von Zweigbergk <martinvonz@google.com> [Thu, 15 Sep 2016 09:45:29 -0700] rev 29956
strip: report both bundle files in case of exception (issue5368)
If strip fails while recovering the temporary bundle (e.g. because a
hook fails), we tell the user only about the backup bundle, not about
the temporary bundle. Since the user did not ask to strip the commits
in the temporary bundle, that's the more important bundle to mention,
so let's do that (and also mention the backup bundle as usual).
Martin von Zweigbergk <martinvonz@google.com> [Thu, 15 Sep 2016 10:18:56 -0700] rev 29955
strip: simplify some repeated conditions
We check "if saveheads or savebases" in several places to see if we
should or have created a bundle of the changesets to apply after
truncating the revlogs. One of the conditions is actually just "if
saveheads", but since there can't be savebases without saveheads, that
is effectively the same condition. It seems simpler to check only once
and from then on see if we created the file.
Mathias De Maré <mathias.demare@gmail.com> [Mon, 29 Aug 2016 07:07:15 +0200] rev 29954
config: add template support
V2:
- Limit escaping to plain formatting only
- Use the formatter consistently (no more ui.debug)
- Always include 'name' and 'value'
V3:
- Always convert 'value' to string (this also makes sure we handle functions)
- Keep real debug message as ui.debug for now
- Add additional tests.
Note: I'm not quite sure about the best approach to handling
the 'print the full config' case.
For me, it printed the 'ui.promptecho' key at the end.
I went with globs there as that at least tests the json display reliably.
Example output:
[
{
"name": "ui.username",
"source": "/home/mathias/.hgrc:2",
"value": "Mathias De Maré <mathias.demare@gmail.com>"
}
]
Mathias De Maré <mathias.demare@gmail.com> [Mon, 29 Aug 2016 17:19:09 +0200] rev 29953
formatter: introduce isplain() to replace (the inverse of) __nonzero__() (API)
V2: also remove and replace __nonzero__
Augie Fackler <augie@google.com> [Tue, 30 Aug 2016 15:55:07 -0400] rev 29952
diffopts: notice a negated boolean flag in diffopts
This means that if you have git-diffs enabled by default (pretty
common) and you hit the rare (but real) case where a git-diff breaks
patch(1) or some other tool, you can easily disable it by just
specifying --no-git on the command line.
I feel a little bad about the isinstance() check, but some values in
diffopts are not booleans and so we need to preserve false iff the
flag is a boolean flag: failing to do this means we end up with empty
string defaults for flags clobbering meaningful values from the [diff]
section in hgrc.
Augie Fackler <augie@google.com> [Tue, 13 Sep 2016 22:57:57 -0400] rev 29951
flags: allow specifying --no-boolean-flag on the command line (BC)
This makes it much easier to enable some anti-foot-shooting features
(like update --check) by default, because now all boolean flags can be
explicitly disabled on the command line without having to use HGPLAIN
or similar. Flags which don't deserve this treatment can be removed
from consideration by adding them to the nevernegate set in fancyopts.
This doesn't make it any easier to identify when a flag is set: opts
still always gets filled in, either with the user-specified flag value
or with the default from the flags list in the command
table. Improving that would probably clean things up a bit, but for
now if you want a boolean flag and care if it was explicitly false or
default false (or true, but nobody uses that functionality because
before now it was nonsense) you need to use None as your default
rather than True or False.
This doesn't (yet) update help output, because I'm not quite sure how
to do that cleanly.
Yuya Nishihara <yuya@tcha.org> [Tue, 03 May 2016 13:36:12 +0900] rev 29950
revset: make sort() noop depending on ordering requirement (BC)
See the previous patch for why.
Yuya Nishihara <yuya@tcha.org> [Tue, 03 May 2016 13:36:12 +0900] rev 29949
revset: make reverse() noop depending on ordering requirement (BC)
Because smartset.reverse() may modify the underlying subset, it should be
called only if the set can define the ordering.
In the following example, 'a' and 'c' is the same object, so 'b.reverse()'
would reverse 'a' unexpectedly.
# '0:2 & reverse(all())'
<filteredset
<spanset- 0:2>, # a
<filteredset # b
<spanset- 0:2>, # c
<spanset+ 0:9>>>
Yuya Nishihara <yuya@tcha.org> [Tue, 03 May 2016 12:52:50 +0900] rev 29948
revset: fix order of nested 'range' expression (BC)
Enforce range order only if necessary as the comment says "carrying the
sorting over would be more efficient."
Yuya Nishihara <yuya@tcha.org> [Wed, 01 Jun 2016 20:54:04 +0900] rev 29947
revset: forward ordering requirement to argument of present()
present() is special in that it returns the argument set with no
modification, so the ordering requirement should be forwarded.
We could make present() fix the order like orset(), but that would be silly
because we know the extra filtering cost is unnecessary.
Nathan Goldbaum <ngoldbau@illinois.edu> [Wed, 14 Sep 2016 11:39:47 -0500] rev 29946
crecord: delete commented line
Durham Goode <durham@fb.com> [Tue, 13 Sep 2016 16:00:41 -0700] rev 29945
manifest: move dirlog up to manifestrevlog
This removes dirlog and its associated cache from manifest and puts it in
manifestrevlog. The notion of there being sub-logs is specific to the revlog
implementation, and therefore belongs on the revlog class.
This patch will enable future patches to move the serialization logic for
manifests onto manifestrevlog, which will allow us to move manifest.add onto
manifestlog in a way that it just calls out to manifestrevlog for the
serialization.
Durham Goode <durham@fb.com> [Tue, 13 Sep 2016 16:00:41 -0700] rev 29944
manifest: move revlog specific options from manifest to manifestrevlog
The manifestv2 and treeondisk options are specific to how we serialize the
manifest into revlogs, so let's move them onto the manifestrevlog class. This
will allow us to add a manifestlog.add() function in a future diff that will
rely on manifestrevlog to make decisions about how to serialize the given
manifest to disk.
We have to move a little bit of extra logic about the 'dir' as well, since it is
used in conjunction with the treeondisk option to decide the revlog file name.
It's probably good to move this down to the manifestrevlog class anyway, since
it's specific to the revlog.
Durham Goode <durham@fb.com> [Tue, 13 Sep 2016 16:26:30 -0700] rev 29943
manifest: adds manifestctx.readfast
This adds a copy of manifest.readfast to manifestctx.readfast and adds a
consumer of it. It currently looks like duplicate code, but a future patch
causes these functions to diverge as tree concepts are added to the tree
version.
Durham Goode <durham@fb.com> [Tue, 13 Sep 2016 16:25:21 -0700] rev 29942
manifest: add manifestctx.readdelta()
This adds an implementation of readdelta to the new manifestctx class and adds a
couple consumers of it. This currently appears to have some duplicate code, but
future patches cause this function to diverge when things like "shallow" are
introduced.
Pierre-Yves David <pierre-yves.david@ens-lyon.org> [Wed, 14 Sep 2016 17:12:39 +0200] rev 29941
merge with stable
Martin von Zweigbergk <martinvonz@google.com> [Tue, 13 Sep 2016 13:49:42 -0700] rev 29940
rebase: make debug logging more consistent
We emit some lines that mix revision numbers with nodeids, which makes
little sense to me.
Yuya Nishihara <yuya@tcha.org> [Sun, 26 Jun 2016 18:41:28 +0900] rev 29939
revset: fix order of nested '_(|int|hex)list' expression (BC)
This fixes the order of 'x & (y + z)' where 'y' and 'z' are trivial, and the
other uses of _list()-family functions. The original functions are renamed to
'_ordered(|int|hex)list' to say clearly that they do not follow the subset
ordering.
Yuya Nishihara <yuya@tcha.org> [Sun, 26 Jun 2016 18:17:12 +0900] rev 29938
revset: fix order of nested 'or' expression (BC)
This fixes the order of 'x & (y + z)' where 'y' and 'z' are not trivial.
The follow-order 'or' operation is slower than the ordered operation if
an input set is large:
#0 #1 #2 #3
0) 0.002968 0.002980 0.002982 0.073042
1) 0.004513 0.004485 0.012029 0.075261
#0: 0:4000 & (0:1099 + 1000:2099 + 2000:3099)
#1: 4000:0 & (0:1099 + 1000:2099 + 2000:3099)
#2: 10000:0 & (0:1099 + 1000:2099 + 2000:3099)
#3: file("path:hg") & (0:1099 + 1000:2099 + 2000:3099)
I've tried another implementation, but which appeared to be slower than
this version.
ss = [getset(repo, fullreposet(repo), x) for x in xs]
return subset.filter(lambda r: any(r in s for s in ss), cache=False)
Yuya Nishihara <yuya@tcha.org> [Sun, 07 Aug 2016 17:58:50 +0900] rev 29937
revset: add 'takeorder' attribute to mark functions that need ordering flag
Since most functions shouldn't need 'order' flag, it is passed only when
explicitly required. This avoids large API breakage.
Yuya Nishihara <yuya@tcha.org> [Sun, 07 Aug 2016 17:46:12 +0900] rev 29936
revset: pass around ordering flags to operations
Some operations and functions will need them to fix ordering bugs.
Yuya Nishihara <yuya@tcha.org> [Sun, 07 Aug 2016 17:48:52 +0900] rev 29935
revset: add stub to handle parentpost operation
All operations will take 'order' flag, but p1() function won't.
Yuya Nishihara <yuya@tcha.org> [Tue, 16 Feb 2016 22:02:16 +0900] rev 29934
revset: infer ordering flag to teach if operation should define/follow order
New flag 'order' is the hint to determine if a function or operation can
enforce its ordering requirement or take the ordering already defined. It
will be used to fix a couple of ordering bugs, such as:
a) 'x & (y | z)' disregards the order of 'x' (issue5100)
b) 'x & y:z' is listed from 'y' to 'z'
c) 'x & y' can be rewritten as 'y & x' if weight(x) > weight(y)
(a) and (b) are bugs of the revset core. Before this, there was no way to
tell if 'orset()' and 'rangeset()' can enforce its ordering. These bugs
could be addressed by overriding __and__() of the initial set to take the
ordering of the other set:
class fullreposet:
def __and__(self, other):
# allow other to enforce its ordering
return other
but it would expose (c), which is a hidden bug of optimize(). So, in either
ways, optimize() have to know the current ordering requirement. Otherwise,
it couldn't rewrite expressions by weights with no output change, nor tell
how a revset function or operation should order the entries.
'order' is tri-state. It starts with 'define', and shifts to 'follow' by
'x & y'. It changes back to 'define' on function call 'f(x)' or function-like
operation 'x (f) y' because 'f' may have its own ordering requirement for 'x'
and 'y'. The state 'any' will allow us to avoid extra cost that would be
necessary to constrain ordering where it isn't important, 'not x'.