##// END OF EJS Templates
wireprotov2: implement commands as a generator of objects...
wireprotov2: implement commands as a generator of objects Previously, wire protocol version 2 inherited version 1's model of having separate types to represent the results of different wire protocol commands. As I implemented more powerful commands in future commits, I found I was using a common pattern of returning a special type to hold a generator. This meant the command function required a closure to do most of the work. That made logic flow more difficult to follow. I also noticed that many commands were effectively a sequence of objects to be CBOR encoded. I think it makes sense to define version 2 commands as generators. This way, commands can simply emit the data structures they wish to send to the client. This eliminates the need for a closure in command functions and removes encoding from the bodies of commands. As part of this commit, the handling of response objects has been moved into the serverreactor class. This puts the reactor in the driver's seat with regards to CBOR encoding and error handling. Having error handling in the function that emits frames is particularly important because exceptions in that function can lead to things getting in a bad state: I'm fairly certain that uncaught exceptions in the frame generator were causing deadlocks. I also introduced a dedicated error type for explicit error reporting in command handlers. This will be used in subsequent commits. There's still a bit of work to be done here, especially around formalizing the error handling "protocol." I've added yet another TODO to track this so we don't forget. Test output changed because we're using generators and no longer know we are at the end of the data until we hit the end of the generator. This means we can't emit the end-of-stream flag until we've exhausted the generator. Hence the introduction of 0-sized end-of-stream frames. Differential Revision: https://phab.mercurial-scm.org/D4472

File last commit:

r38749:d79f3afb default
r39595:07b58266 default
Show More
test-commit-interactive-curses.t
444 lines | 8.8 KiB | text/troff | Tads3Lexer
/ tests / test-commit-interactive-curses.t
#require tic
Set up a repo
$ cp $HGRCPATH $HGRCPATH.pretest
$ cat <<EOF >> $HGRCPATH
> [ui]
> interactive = true
> interface = curses
> [experimental]
> crecordtest = testModeCommands
> EOF
Record with noeol at eof (issue5268)
$ hg init noeol
$ cd noeol
$ printf '0' > a
$ printf '0\n' > b
$ hg ci -Aqm initial
$ printf '1\n0' > a
$ printf '1\n0\n' > b
$ cat <<EOF >testModeCommands
> c
> EOF
$ HGEDITOR="\"sh\" \"`pwd`/editor.sh\"" hg commit -i -m "add hunks" -d "0 0"
$ cd ..
Normal repo
$ hg init a
$ cd a
Committing some changes but stopping on the way
$ echo "a" > a
$ hg add a
$ cat <<EOF >testModeCommands
> TOGGLE
> X
> EOF
$ hg commit -i -m "a" -d "0 0"
no changes to record
[1]
$ hg tip
changeset: -1:000000000000
tag: tip
user:
date: Thu Jan 01 00:00:00 1970 +0000
Committing some changes
$ cat <<EOF >testModeCommands
> X
> EOF
$ hg commit -i -m "a" -d "0 0"
$ hg tip
changeset: 0:cb9a9f314b8b
tag: tip
user: test
date: Thu Jan 01 00:00:00 1970 +0000
summary: a
Check that commit -i works with no changes
$ hg commit -i
no changes to record
[1]
Committing only one file
$ echo "a" >> a
>>> open('b', 'wb').write(b"1\n2\n3\n4\n5\n6\n7\n8\n9\n10\n") and None
$ hg add b
$ cat <<EOF >testModeCommands
> TOGGLE
> KEY_DOWN
> X
> EOF
$ hg commit -i -m "one file" -d "0 0"
$ hg tip
changeset: 1:fb2705a663ea
tag: tip
user: test
date: Thu Jan 01 00:00:00 1970 +0000
summary: one file
$ hg cat -r tip a
a
$ cat a
a
a
Committing only one hunk while aborting edition of hunk
- Untoggle all the hunks, go down to the second file
- unfold it
- go down to second hunk (1 for the first hunk, 1 for the first hunkline, 1 for the second hunk, 1 for the second hunklike)
- toggle the second hunk
- toggle on and off the amend mode (to check that it toggles off)
- edit the hunk and quit the editor immediately with non-zero status
- commit
$ printf "printf 'editor ran\n'; exit 1" > editor.sh
$ echo "x" > c
$ cat b >> c
$ echo "y" >> c
$ mv c b
$ cat <<EOF >testModeCommands
> A
> KEY_DOWN
> f
> KEY_DOWN
> KEY_DOWN
> KEY_DOWN
> KEY_DOWN
> TOGGLE
> a
> a
> e
> X
> EOF
$ HGEDITOR="\"sh\" \"`pwd`/editor.sh\"" hg commit -i -m "one hunk" -d "0 0"
editor ran
$ rm editor.sh
$ hg tip
changeset: 2:7d10dfe755a8
tag: tip
user: test
date: Thu Jan 01 00:00:00 1970 +0000
summary: one hunk
$ hg cat -r tip b
1
2
3
4
5
6
7
8
9
10
y
$ cat b
x
1
2
3
4
5
6
7
8
9
10
y
$ hg commit -m "other hunks"
$ hg tip
changeset: 3:a6735021574d
tag: tip
user: test
date: Thu Jan 01 00:00:00 1970 +0000
summary: other hunks
$ hg cat -r tip b
x
1
2
3
4
5
6
7
8
9
10
y
Newly added files can be selected with the curses interface
$ hg update -C .
0 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ echo "hello" > x
$ hg add x
$ cat <<EOF >testModeCommands
> TOGGLE
> TOGGLE
> X
> EOF
$ hg st
A x
? testModeCommands
$ hg commit -i -m "newly added file" -d "0 0"
$ hg st
? testModeCommands
Amend option works
$ echo "hello world" > x
$ hg diff -c .
diff -r a6735021574d -r 2b0e9be4d336 x
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/x Thu Jan 01 00:00:00 1970 +0000
@@ -0,0 +1,1 @@
+hello
$ cat <<EOF >testModeCommands
> a
> X
> EOF
$ hg commit -i -m "newly added file" -d "0 0"
saved backup bundle to $TESTTMP/a/.hg/strip-backup/2b0e9be4d336-3cf0bc8c-amend.hg
$ hg diff -c .
diff -r a6735021574d -r c1d239d165ae x
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/x Thu Jan 01 00:00:00 1970 +0000
@@ -0,0 +1,1 @@
+hello world
Make file empty
$ printf "" > x
$ cat <<EOF >testModeCommands
> X
> EOF
$ hg ci -i -m emptify -d "0 0"
$ hg update -C '.^' -q
Editing a hunk puts you back on that hunk when done editing (issue5041)
To do that, we change two lines in a file, pretend to edit the second line,
exit, toggle the line selected at the end of the edit and commit.
The first line should be recorded if we were put on the second line at the end
of the edit.
$ hg update -C .
0 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ echo "foo" > x
$ echo "hello world" >> x
$ echo "bar" >> x
$ cat <<EOF >testModeCommands
> f
> KEY_DOWN
> KEY_DOWN
> KEY_DOWN
> KEY_DOWN
> e
> TOGGLE
> X
> EOF
$ printf "printf 'editor ran\n'; exit 0" > editor.sh
$ HGEDITOR="\"sh\" \"`pwd`/editor.sh\"" hg commit -i -m "edit hunk" -d "0 0" -q
editor ran
$ hg cat -r . x
foo
hello world
Testing the review option. The entire final filtered patch should show
up in the editor and be editable. We will unselect the second file and
the first hunk of the third file. During review, we will decide that
"lower" sounds better than "bottom", and the final commit should
reflect this edition.
$ hg update -C .
1 files updated, 0 files merged, 0 files removed, 0 files unresolved
$ echo "top" > c
$ cat x >> c
$ echo "bottom" >> c
$ mv c x
$ echo "third a" >> a
$ echo "we will unselect this" >> b
$ cat > editor.sh <<EOF
> cat "\$1"
> cat "\$1" | sed s/bottom/lower/ > tmp
> mv tmp "\$1"
> EOF
$ cat > testModeCommands <<EOF
> KEY_DOWN
> TOGGLE
> KEY_DOWN
> f
> KEY_DOWN
> TOGGLE
> R
> EOF
$ HGEDITOR="\"sh\" \"`pwd`/editor.sh\"" hg commit -i -m "review hunks" -d "0 0"
# To remove '-' lines, make them ' ' lines (context).
# To remove '+' lines, delete them.
# Lines starting with # will be removed from the patch.
#
# If the patch applies cleanly, the edited patch will immediately
# be finalised. If it does not apply cleanly, rejects files will be
# generated. You can use those when you try again.
diff --git a/a b/a
--- a/a
+++ b/a
@@ -1,2 +1,3 @@
a
a
+third a
diff --git a/x b/x
--- a/x
+++ b/x
@@ -1,2 +1,3 @@
foo
hello world
+bottom
$ hg cat -r . a
a
a
third a
$ hg cat -r . b
x
1
2
3
4
5
6
7
8
9
10
y
$ hg cat -r . x
foo
hello world
lower
Check spacemovesdown
$ cat <<EOF >> $HGRCPATH
> [experimental]
> spacemovesdown = true
> EOF
$ cat <<EOF >testModeCommands
> TOGGLE
> TOGGLE
> X
> EOF
$ hg status -q
M b
M x
$ hg commit -i -m "nothing to commit?" -d "0 0"
no changes to record
[1]
Check ui.interface logic for the chunkselector
The default interface is text
$ cp $HGRCPATH.pretest $HGRCPATH
$ chunkselectorinterface() {
> $PYTHON <<EOF
> from mercurial import hg, ui;\
> repo = hg.repository(ui.ui.load(), ".");\
> print(repo.ui.interface("chunkselector"))
> EOF
> }
$ chunkselectorinterface
text
If only the default is set, we'll use that for the feature, too
$ cp $HGRCPATH.pretest $HGRCPATH
$ cat <<EOF >> $HGRCPATH
> [ui]
> interface = curses
> EOF
$ chunkselectorinterface
curses
If TERM=dumb, we use text, even if the config says curses
$ chunkselectorinterface
curses
$ TERM=dumb chunkselectorinterface
text
(Something is keeping TERM=dumb in the environment unless I do this, it's not
scoped to just that previous command like in many shells)
$ TERM=xterm chunkselectorinterface
curses
It is possible to override the default interface with a feature specific
interface
$ cp $HGRCPATH.pretest $HGRCPATH
$ cat <<EOF >> $HGRCPATH
> [ui]
> interface = text
> interface.chunkselector = curses
> EOF
$ chunkselectorinterface
curses
$ cp $HGRCPATH.pretest $HGRCPATH
$ cat <<EOF >> $HGRCPATH
> [ui]
> interface = curses
> interface.chunkselector = text
> EOF
$ chunkselectorinterface
text
If a bad interface name is given, we use the default value (with a nice
error message to suggest that the configuration needs to be fixed)
$ cp $HGRCPATH.pretest $HGRCPATH
$ cat <<EOF >> $HGRCPATH
> [ui]
> interface = blah
> EOF
$ chunkselectorinterface
invalid value for ui.interface: blah (using text)
text
$ cp $HGRCPATH.pretest $HGRCPATH
$ cat <<EOF >> $HGRCPATH
> [ui]
> interface = curses
> interface.chunkselector = blah
> EOF
$ chunkselectorinterface
invalid value for ui.interface.chunkselector: blah (using curses)
curses
$ cp $HGRCPATH.pretest $HGRCPATH
$ cat <<EOF >> $HGRCPATH
> [ui]
> interface = blah
> interface.chunkselector = curses
> EOF
$ chunkselectorinterface
invalid value for ui.interface: blah
curses
$ cp $HGRCPATH.pretest $HGRCPATH
$ cat <<EOF >> $HGRCPATH
> [ui]
> interface = blah
> interface.chunkselector = blah
> EOF
$ chunkselectorinterface
invalid value for ui.interface: blah
invalid value for ui.interface.chunkselector: blah (using text)
text