jj log, extra space in some changes #9462

arialdomartini · 2026-05-10T07:48:27Z

arialdomartini
May 10, 2026

I use this template:

simple_log = """
pad_end(4,
    self.change_id().shortest() ++
    if(self.divergent() || self.hidden(),
        "/" ++ self.change_offset(),
        ""
    )
)
++ if(self.immutable(), " 🔒", "   ")
++ "  "
++ if(self.empty(),
    label("empty", "empty "),
    ""
)
++ self.description().first_line()
++ " " ++ self.bookmarks()
  """

which is extremely compact:

but I cannot understand why the graph adds an extra space to some changes (the x in the example), which by the way are not necessarily merge changes.

This seams to have nothing to do with the template itself. Even reducing the template to:


simple_log = """
"-"
"""

the result is:


  @  -
  ○  -
  ◆    -
  ├─╮
  │ │
  │ ~
  │
  ~  (elided revisions)
  │ ○  -
  ├─╯
  ◆  -
  │
  ~

notice the extra space in the 3rd change from the top.

I just wonder:

if this is an error, for which I should open an issue.
if this can be somehow removed / controlled, using a configuration.

arialdomartini · 2026-05-10T08:06:03Z

arialdomartini
May 10, 2026
Author

In the case of marge changes I could fix the issue like this:

simple_log = """
if(self.parents().len() > 1, raw_escape_sequence("\b\b"), "")
++ pad_end(4,....

but this does not work in case the problem affect changes which are not merges, but yet they get an extra space (I think, because the actual merge commit is elided in the visualization).

0 replies

arialdomartini · 2026-05-10T08:22:18Z

arialdomartini
May 10, 2026
Author

As an extra question: I think that pad_end doesn't account for the graph prefix width: it pads relative to the start of the template output, not the terminal margin, so columns drift whenever the graph column changes width (merges, side branches, etc.).

Is there a way to the graph space occupation into account?

1 reply

josephlou5 May 23, 2026
Collaborator

Correct, pad_(start|end) are not based on position in the terminal but simply operate on strings. There is no way to take the graph render into account because to do so would require figuring out what the entire graph looks like (with all merges, parents, children, other branches, etc.) before rendering each commit. For optimization reasons, jj renders one commit at a time, so this is not possible. But regardless, #9434 may be interesting to you.

arialdomartini · 2026-05-13T09:59:01Z

arialdomartini
May 13, 2026
Author

I think this is related to the discussion in #8407.

0 replies

josephlou5 · 2026-05-23T01:25:01Z

josephlou5
May 23, 2026
Collaborator

In your screenshot, x is indented in case the template takes multiple lines, in which case we would want the template to line up to the right of the second lane below x. So because x is rendered with multiple parents, its template will necessarily be indented, because jj cannot know whether the template will be one line or multiple. In your case it looks like your template is ALWAYS one line, but again, jj doesn't (can't) know that. So it's not really a bug, but just a consequence of how the graph is rendered.

I'm not exactly sure, but you said x is not a merge commit? Or did you mean that sometimes you see this indentation happening for non-merge commits? Either way, it sounds like #6884.

2 replies

arialdomartini May 25, 2026
Author

Yes, it's not a merge commit. Please, consider this example:

jj log -T builtin_log_oneline

  @  l arialdo.martini 2026-04-25 14:20:56 09 (no description set)
  ○  u arialdo.martini 2026-03-22 17:05:18 master* 2 Simulation Report doc
  ◆    x yousefyou1 2026-02-23 06:13:14 master@origin 5 the down slider is disabled
  ├─╮
  │ │
  │ ~
  │
  ~  (elided revisions)
  │ ○  s arialdo.martini 2026-05-10 10:17:41 e xxxxxxxxxxx
  ├─╯
  ◆  q arialdo.martini 2026-01-30 01:57:39 last 00 In Finance graph BCR and Payback are in table only
  │
  ~

Notice the extra space for x.
Compare the same with:

 jj log -T builtin_log_oneline -r ::@
  @  l arialdo.martini 2026-04-25 14:20:56 09 (no description set)
  ○  u arialdo.martini 2026-03-22 17:05:18 master* 2 Simulation Report doc
  ◆  x yousefyou1 2026-02-23 06:13:14 master@origin 5 the down slider is disabled
  ◆    ylm arialdo.martini 2026-02-11 19:21:35 29b7 (empty) Merge pull request #221 from Umwelt-Design-LLP/Visual-bug-Fix
  ├─╮
  │ ◆  ksv yousefyou1 2026-02-10 01:30:02 Visual-bug-Fix@origin ffc when you click on an analysis it take to the mcda results
  │ ◆  utzw yousefyou1 2026-02-03 18:57:09 1b7f the unit appear now beside the subtitle and the the health is not shown anymor  e in the mcda results when welfare optimization is chosen
  ├─╯
  ◆    zvz arialdo.martini 2026-01-30 02:11:36 43e (empty) Merge #220 last changes before the demo
  ├─╮
  │ ◆  q arialdo.martini 2026-01-30 01:57:39 last 00 In Finance graph BCR and Payback are in table only
  │ ◆  twk arialdo.martini 2026-01-30 01:56:24 9a40 newFinRadarData -> newFinanceBarData
  │ ◆  lkl arialdo.martini 2026-01-30 01:54:58 bf5 financeRadarData -> financeBarData
  │ ◆  los arialdo.martini 2026-01-30 01:54:12 4e7 In Environment graph, Net GHG is in the table only
  │ ◆  ktr arialdo.martini 2026-01-30 01:51:51 fe1 Fix #205: no hardcoded values for downloading the PDF report
  │ ◆  sup arialdo.martini 2026-01-30 01:51:05 597 Health graphs is properly sized
  │ ◆  qrk arialdo.martini 2026-01-30 01:49:41 8b1 Fix #136: Health is a bar chart instead of a Radar chart
  ├─╯

In the first case, an extra space is added to x. Notice that x is not a merge.
In the second case, there is no extra space for x but there is for its parent, ylm.

josephlou5 May 25, 2026
Collaborator

Yes, that sounds exactly like #6884. Something about x having a merge ancestor that isn't currently rendered or something?

As I said, the log for ylm will be intended just in case it has multiple lines of output. It's more obvious with your second log output because you can see that ylm lines up with ksv and utzw below. But in the first output with just x, it admittedly looks like x is unnecessarily intended, but it's just lining up out of the way of the graph symbols.

joyously · 2026-05-25T14:20:15Z

joyously
May 25, 2026

In the example, it looks like the extra space is because the next line has a corner symbol. The q line does not, but the zvz line does.

0 replies

Navigation Menu

Repository navigation

jj log, extra space in some changes #9462

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{editor}}'s edit

{{editor}}'s edit

Uh oh!

Replies: 5 comments 3 replies

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Select a reply

Uh oh!

jj log, extra space in some changes #9462

Uh oh!

Uh oh!

arialdomartini May 10, 2026

Replies: 5 comments · 3 replies

Uh oh!

arialdomartini May 10, 2026 Author

Uh oh!

arialdomartini May 10, 2026 Author

Uh oh!

josephlou5 May 23, 2026 Collaborator

Uh oh!

arialdomartini May 13, 2026 Author

Uh oh!

josephlou5 May 23, 2026 Collaborator

Uh oh!

arialdomartini May 25, 2026 Author

Uh oh!

josephlou5 May 25, 2026 Collaborator

Uh oh!

joyously May 25, 2026

arialdomartini
May 10, 2026

Replies: 5 comments 3 replies

arialdomartini
May 10, 2026
Author

arialdomartini
May 10, 2026
Author

josephlou5 May 23, 2026
Collaborator

arialdomartini
May 13, 2026
Author

josephlou5
May 23, 2026
Collaborator

arialdomartini May 25, 2026
Author

josephlou5 May 25, 2026
Collaborator

joyously
May 25, 2026