Change all indentation from 3 spaces to 2 using findent #80

JamieJQuinn · 2024-03-26T12:00:10Z

Used pip-installed findent 4.3.1 to change all indentation from 3 spaces to 2 using the following fish command:

for f in (find . -name '*.f90'); findent -i2 -I2 < $f > $f.temp; mv $f.temp $f; end

semi-h · 2024-03-26T12:11:07Z

src/solver.f90

@@ -283,11 +283,11 @@ subroutine divergence_v2p(self, div_u, u, v, w)
      ! Interpolation for v field in x
      ! Interpolation for w field in x
      call self%backend%tds_solve(du_x, u, self%xdirps, &
-                                  self%xdirps%stagder_v2p)


Also here, I would expect the continuation from the parentheses, isn't this more common?

I believe we can play with this parameter

--align_paren[=<n>] align continuation lines not preceded by '&' with preceding unmatched left parenthesis. n=0, or flag omitted: do not align. n=1: do align (default).

though the default should be do leave it as is

semi-h · 2024-03-26T12:15:19Z

src/cuda/kernels/distributed.f90

-      last_r = ffr(1)
-
-      du(i, 1, b) = coeffs_s(1, 1)*u_s(i, 1, b) &
-                    + coeffs_s(2, 1)*u_s(i, 2, b) &


I think this would be very confusing, very surprised ~~fprettify~~ findent allows this to be honest.

Again not visible here on the comment tab, but + coeff... starts from all the way left with two spaces, however I would expect it to start from right below the previous variable after the =

yeah, I didn't see any way in the manual to indent continuation lines after a = nor a :: with findent

with :: its not the end of the world, one can argue its better to follow findent way, but with + I think things would get very confusing and hard to follow.

I am tempted to write a small script just to handle these tbh. This is pretty lacking indeed

semi-h · 2024-03-26T12:17:20Z

src/allocator.f90

@@ -1,37 +1,37 @@
-module m_allocator


This is also interesting, even the module declaration starts with two spacing. Is there a way to disable it? seems like a waste of two spaces to me. (Because its like +2 in every single line due to this)

This is the -I2 parameter, this can be removed

semi-h · 2024-03-26T12:19:43Z

src/allocator.f90

-     !! by calling the type bound procedure
-     !! [[m_allocator(module):release_block(subroutine)]].  The
-     !! released block is then pushed in front of the block list.
+  module m_allocator


Commented on the wrong side and its actually not visible due to that. I mean the 2 spaces before module m_allocator

see #80 (comment)

semi-h · 2024-03-26T12:27:46Z

src/solver.f90

-                                 u_y, v_y, w_y, du_y, dv_y, dw_y, &
-                                 u_z, w_z, dw_z
+        u_y, v_y, w_y, du_y, dv_y, dw_y, &
+        u_z, w_z, dw_z


This is actually everywhere but had to comment on a specific example. I would expect the continuation line to start aligned with the first variable like it used to be.

semi-h · 2024-03-26T12:28:50Z

src/cuda/kernels/distributed.f90

+    last_r = ffr(1)
+
+    du(i, 1, b) = coeffs_s(1, 1)*u_s(i, 1, b) &
+      + coeffs_s(2, 1)*u_s(i, 2, b) &


Commenting here just to show the actual change with spacing

semi-h · 2024-03-26T12:40:12Z

src/ordering.f90

-         case (RDR_Z2Y)
-            dir_in = DIR_Z
-            dir_out = DIR_Y
+       case (RDR_X2Y)


Another interesting point. case is shifted only one space, not two, and also the lines inside the case are shifted by one.

This one will be fixed by using --indent_select=4 --indent_case=2

semi-h · 2024-03-26T12:43:28Z

src/poisson_fft.f90

-         yky(i) = yky(ny-i+2)
-         eys(i) = eys(ny-i+2)
-         yk2(i) = yk2(ny-i+2)
+        yky(i) = yky(ny-i+2)


Looks like findent (at least with current settings) doesn't enforce anything between operators. (I think this particular line is not compatible with our current preference (spacing around - and + but no spacing around * and /) it would be nice if findent can identify and enforce a spacing based on our preference.

findent doesn't support modifying these things. It is really mainly doing indentation.

Its actually good. Because fprettify is too restrictive with spacing around operators, and I don't really like it. For example, when passing a keyword argument to a function, I would rather not using spaces around =, but of course prefer spaces around assingment = normally. And fprettify can't tell them apart and want you to pick one and use it everywhere. Also with if and allocate. You may want space after if, but no space after allocate, but fprettify won't let you do this because it considers these two in the same group and the rule is defined for the whole group...

semi-h · 2024-03-26T12:59:18Z

src/common.f90

-                               RDR_Y2X, 0, RDR_Y2Z, RDR_Y2C, &
-                               RDR_Z2X, RDR_Z2Y, 0, RDR_Z2C, &
-                               RDR_C2X, RDR_C2Y, RDR_C2Z, 0], shape=[4, 4])
+      RDR_Y2X, 0, RDR_Y2Z, RDR_Y2C, &


This line is a good one to test findent. With the settings @Nanoseb recommended I would expect this line to start after the [ above. Curious to see what will happen

To my surprise, it does work:

integer, protected :: & rdr_map(4, 4) = reshape([0, RDR_X2Y, RDR_X2Z, RDR_X2C, & RDR_Y2X, 0, RDR_Y2Z, RDR_Y2C, & RDR_Z2X, RDR_Z2Y, 0, RDR_Z2C, & RDR_C2X, RDR_C2Y, RDR_C2Z, 0], shape=[4, 4])

With findent --align_paren=1 < common.f90

Nanoseb · 2024-03-26T13:06:59Z

Used pip-installed findent 4.3.1 to change all indentation from 3 spaces to 2 using the following fish command:
for f in (find . -name '*.f90'); findent -i2 -I2 < $f > $f.temp; mv $f.temp $f; end

I would add

--align_paren=1 to align multilines parenthesis)
--refactor_end to add subroutine names after end subroutine
--indent_select=4 --indent_case=2 to for case/select statements

semi-h · 2024-03-26T13:25:49Z

Is there any setting for alignment for multi-line assignments and for multi-line variable declarations?

Nanoseb · 2024-03-26T13:34:38Z

Is there any setting for alignment for multi-line assignments and for multi-line variable declarations?

no, unfortunately I haven't seen anything about that in the help page

to use in combination with findent

Nanoseb · 2024-03-26T14:50:29Z

Because the main limitation of findent is line continuation after =, :: and :. I wrote a small script to handle only these. It is to be used on findent output.

findent < in.f90 | script/indent_continuation.py > out.f90

Nanoseb · 2024-03-29T15:46:13Z

using fprettify instead see #81

Apply indentation change

ab37d9e

semi-h reviewed Mar 26, 2024

View reviewed changes

Add script to handle indent continuation

706d8fb

to use in combination with findent

Nanoseb changed the title ~~Change all indentation from 3 spaces to 2~~ Change all indentation from 3 spaces to 2 using findent Mar 27, 2024

Nanoseb closed this Mar 29, 2024

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Change all indentation from 3 spaces to 2 using findent #80

Change all indentation from 3 spaces to 2 using findent #80

JamieJQuinn commented Mar 26, 2024

semi-h Mar 26, 2024

Nanoseb Mar 26, 2024

semi-h Mar 26, 2024 •

edited

semi-h Mar 26, 2024

Nanoseb Mar 26, 2024

semi-h Mar 26, 2024

Nanoseb Mar 26, 2024

semi-h Mar 26, 2024 •

edited

Nanoseb Mar 26, 2024

semi-h Mar 26, 2024

Nanoseb Mar 26, 2024

semi-h Mar 26, 2024

semi-h Mar 26, 2024

semi-h Mar 26, 2024

Nanoseb Mar 26, 2024 •

edited

semi-h Mar 26, 2024

Nanoseb Mar 26, 2024

semi-h Mar 26, 2024

semi-h Mar 26, 2024

Nanoseb Mar 26, 2024 •

edited

semi-h Mar 26, 2024

Nanoseb commented Mar 26, 2024 •

edited

semi-h commented Mar 26, 2024

Nanoseb commented Mar 26, 2024

Nanoseb commented Mar 26, 2024

Nanoseb commented Mar 29, 2024

Change all indentation from 3 spaces to 2 using findent #80

Change all indentation from 3 spaces to 2 using findent #80

Conversation

JamieJQuinn commented Mar 26, 2024

Choose a reason for hiding this comment

Choose a reason for hiding this comment

semi-h Mar 26, 2024 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

semi-h Mar 26, 2024 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Nanoseb Mar 26, 2024 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Nanoseb Mar 26, 2024 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Nanoseb commented Mar 26, 2024 • edited

semi-h commented Mar 26, 2024

Nanoseb commented Mar 26, 2024

Nanoseb commented Mar 26, 2024

Nanoseb commented Mar 29, 2024

semi-h Mar 26, 2024 •

edited

semi-h Mar 26, 2024 •

edited

Nanoseb Mar 26, 2024 •

edited

Nanoseb Mar 26, 2024 •

edited

Nanoseb commented Mar 26, 2024 •

edited