10036
Comment:
|
11421
|
Deletions are marked like this. | Additions are marked like this. |
Line 53: | Line 53: |
|| unstable || instability || || || in use || || unsettled || ? || || || || || uprooted || ? || || || || || orphaned || orphan? || || || || |
|| unstable || instability || || too generic || in use || || unsettled || ? || || || no || || uprooted || ? || || bad timing || || || orphaned || orphan? || pretty good || || || |
Line 76: | Line 76: |
|| latecomer || ? || || || old abandoned name || | || latecomer || ? || || mouthful || old abandoned name || |
Line 78: | Line 78: |
|| trumped || || || || || | |
Line 116: | Line 117: |
|| Duplicated || duplication || || || || | || Duplicated || duplication || || content are different || || |
Line 118: | Line 119: |
|| forked || fork? || || || || | || forked || fork? || || too overloaded || || |
Line 124: | Line 125: |
|| competing || ? || || || || || scattered || ? || || || || || unmerged || ? || || || || || alternative || ? || || || || || twinned || ? || || || || || spread || ? || || || || || dispersed || ? || || || || || deviated || ? || || || || || unjoined || ? || || || || |
|
Line 134: | Line 144: |
|| Successor || supersede || || || in use || | || Successor || supersede || || long? obscure? || in use || |
Line 138: | Line 148: |
|| new || ? || || too generic || || || post-<image> || ? || || || || || obsoleting || ? || || || || |
|
Line 147: | Line 159: |
|| predecessor || precede || || || || | || predecessor || precede || proper complement to successor, has a different shape || || || |
Line 152: | Line 164: |
|| old || ? || || too generic || || || pre-<image> || ? || || || || || obsoleted || ? || || || || || replaced || ? || || || || |
|
Line 163: | Line 178: |
* etherpad with notes: https://public.etherpad-mozilla.org/p/evolve-naming_NO_SPAM |
Changeset Evolution - Vocabulary
This page is intended for developer
This pages gather data from discussion about the named use within the ChangesetEvolution concept.
Contents
Troubles
"troubled" changeset and troubles
Evolving history can introduce problems that need to be solved:
- They are intrinsic side effects of the free exchange and rewrite of draft changesets,
- They are an unhealthy state the user needs to take care of before moving forward,
- (Even if they can stay around unsolved for a while),
- They are not critical, we know how to detect and solve them,
- There are different kinds of them.
Changeset adjective |
Concept name |
pro |
con |
status |
troubled |
Troubles |
self-explanatory, not so weird |
uncommon |
in use, maybe? |
conflicted |
conflicts? |
|
confusion with merge conflict, doesn't work for unstable |
|
invalid |
? |
|
disabled people, doesn't suggest resolution |
|
unevolved |
unevolution? |
|
|
maybe? |
dirty |
dirtiness? |
|
terrible |
|
unhealthy |
unhealthiness |
|
confusion with code, suggests contamination |
|
unstable |
instability |
|
|
other use |
unsteady |
|
|
|
maybe? |
unsettled |
? |
|
|
maybe for divergence |
stalled |
? |
|
confusion with work |
|
stagnated |
stagnation |
|
|
|
problematic |
|
|
|
|
volatile |
|
cool word |
|
maybe? |
tainted |
taint |
|
|
|
"unstable" changeset and "instability"
Instability happens when a changeset with descendants is rewritten. The non-obsolete descendants of the now obsolete changeset are called "unstable changesets".
A Changeset is "unstable" because either:
- one of its parents is obsolete,
- one of its parents is unstable.
Automatic resolution of instability by rebasing the unstable changesets on the latest known successors of their obsolete parent. (A changeset that is unstable because its parent is unstable needs to wait for the unstable parent to be stabilized before we can solve it.)
Changeset adjective |
Concept name |
pro |
con |
status |
unstable |
instability |
|
too generic |
in use |
unsettled |
? |
|
|
no |
uprooted |
? |
|
bad timing |
|
orphaned |
orphan? |
pretty good |
|
|
precarious |
? |
|
|
|
dangling |
? |
|
|
|
stale |
? |
|
|
|
"Bumped" changeset and "bumping"
A changeset is called "Bumped" when it is the successor of a public changeset. The public changeset cannot be obsoleted or hidden anymore so both the old and new versions exist. This usually happens when someone is reworking a changeset while someone else is publishing it at the same time elsewhere. The two actions are eventually gathered somewhere and evolve detects there is an issue. To some extend "bumping" can be seen as "divergence with your past" as opposed to "divergence with another rewriting that happened in parallel).
So in summary bumped changesets are:
* superseding something public, * trying to obsolete something that cannot be obsolete, * trying to bring "a change" but they are too late to do so.
The automated solution for this is to create a new changeset with the diff between the public changeset and the successors (The diff introduced by the amend).
Changeset adjective |
Concept name |
pro |
con |
status |
latecomer |
? |
|
mouthful |
old abandoned name |
bumped |
bumping |
|
|
in use |
trumped |
|
|
|
|
invalidated |
invalidation? |
|
|
|
behind |
? |
|
|
|
superseded |
? |
|
|
|
lagging |
lagginess? |
|
|
|
obviated |
? |
|
|
|
unlucky |
? |
|
|
|
tardy |
tardiness |
|
|
|
rewritten |
? |
|
|
|
covering |
? |
|
|
|
shadowed |
shadowing |
|
|
|
replacer |
? |
|
|
|
belated |
? |
|
|
|
mutated |
mutant |
|
|
|
mislead |
? |
|
|
|
naive |
? |
|
|
|
unaware |
? |
|
|
|
mindless |
? |
|
|
|
disenchanting |
? |
|
|
|
dangling replacement |
? |
|
|
|
undermined |
? |
|
|
|
inhibited |
? |
|
|
|
deferred |
? |
|
|
|
obviated |
? |
|
|
|
forestalled |
? |
|
|
|
"divergent" changeset and "divergence"
Divergence happens when two changesets claim to have superseded the same changeset. The canonical way to do this is to have two people in two different repos rewriting the same changeset. The divergence is detected when one pulls from the other. (There is currently no other easy way to get divergence locally).
* Divergence denotes that two (or more) different and independent "edits" happened on the same changeset, * No version can be called "better/newer" than the other one, so both are "alive" at the same time, duplicating all the common changes, * There is at least one latest common precursor that both divergent changesets claim to rewrite. They are said to be divergent from that latest common precursor. * A divergent changeset is divergent with the other non-obsolete changeset.
The automated way to solve this, is to merge the two divergent changesets using the precursor they are divergent from as a base.
Changeset adjective |
Concept name |
pro |
con |
status |
divergent |
divergence |
|
|
in use |
Duplicated |
duplication |
|
content are different |
|
concurrent |
? |
|
|
|
forked |
fork? |
|
too overloaded |
|
conflicting |
conflict? |
|
|
|
conflicting replacement |
? |
|
|
|
contested |
? |
|
|
|
disputed |
? |
|
|
|
torn |
? |
|
|
|
competing |
? |
|
|
|
scattered |
? |
|
|
|
unmerged |
? |
|
|
|
alternative |
? |
|
|
|
twinned |
? |
|
|
|
spread |
? |
|
|
|
dispersed |
? |
|
|
|
deviated |
? |
|
|
|
unjoined |
? |
|
|
|
Obsolescence graph
Obsolescence markers create an orthogonal graph between the changesets, that track what changeset gets rewritten into which other. For this we use obsolescence-marker. A small data structure that contains various information including the fact that "Changesets [Y] are replacing changeset X". We are interested in names used to describe the X ← Y relation. In both directions and from the point of view of each side of it.
"Successor"
"successor" are the "new" part of that relation.
Noun |
verb |
pro |
con |
status |
Successor |
supersede |
|
long? obscure? |
in use |
Successor |
suceed |
|
|
in use |
replacement |
replace |
|
|
|
next |
? |
|
|
|
new |
? |
|
too generic |
|
post-<image> |
? |
|
|
|
obsoleting |
? |
|
|
|
"Precursor"
"precusor" are the "old" part of that relation. The precursors will become obsolete and hidden (except in multiple cases).
Noun |
verb |
pro |
con |
status |
precursor |
precede? |
|
|
in use |
original |
? |
|
|
|
predecessor |
precede |
proper complement to successor, has a different shape |
|
|
progenitor |
? |
|
|
|
previous |
? |
|
|
|
prior |
? |
|
|
|
antecedent |
? |
|
|
|
old |
? |
|
too generic |
|
pre-<image> |
? |
|
|
|
obsoleted |
? |
|
|
|
replaced |
? |
|
|
|
== See also ==
mailing list discussion: [RFC] naming of obsolescence troubles (2012-September)
etherpad with notes: https://public.etherpad-mozilla.org/p/evolve-naming_NO_SPAM