Skip to contentSkip to conversation

Lindent: Handle missing indent gracefully

PATCH In review Jean Delvare 3 participants 7 messages Jul 10, 2015 – Jul 15, 2015

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 Lindent: Handle missing indent gracefully

If indent is not found, bail out immediately instead of spitting random shell script error messages.

---
 scripts/Lindent |    3 +++
 1 file changed, 3 insertions(+)
scripts/Lindent +3 -0
@@ -1,6 +1,9 @@
 #!/bin/sh
 PARAM="-npro -kr -i8 -ts8 -sob -l80 -ss -ncs -cp1"
 RES=`indent --version`
+if [ "$RES" = "" ]; then
+	exit 1
+fi
 V1=`echo $RES | cut -d' ' -f3 | cut -d'.' -f1`
 V2=`echo $RES | cut -d' ' -f3 | cut -d'.' -f2`
 V3=`echo $RES | cut -d' ' -f3 | cut -d'.' -f3`
Signature · 2 lines
Jean Delvare
SUSE L3 Support
#1 lore raw
Quoted 2 lines · Jean Delvare

If indent is not found, bail out immediately instead of spitting random shell script error messages.

OK, but can't we just delete Lindent instead?

Quoted 13 lines
scripts/Lindent +3 -0
@@ -1,6 +1,9 @@
 #!/bin/sh
 PARAM="-npro -kr -i8 -ts8 -sob -l80 -ss -ncs -cp1"
 RES=`indent --version`
+if [ "$RES" = "" ]; then
+	exit 1
+fi
 V1=`echo $RES | cut -d' ' -f3 | cut -d'.' -f1`
 V2=`echo $RES | cut -d' ' -f3 | cut -d'.' -f2`
 V3=`echo $RES | cut -d' ' -f3 | cut -d'.' -f3`
#2 lore raw

Hi Joe,

Quoted 5 lines · Joe Perches
Quoted 2 lines · Jean Delvare

If indent is not found, bail out immediately instead of spitting random shell script error messages.

OK, but can't we just delete Lindent instead?

Because...?

Signature · 2 lines
Jean Delvare
SUSE L3 Support
#3 lore raw
Quoted 1 line · Jean Delvare

Hi Joe,

howdy Jean.

Quoted 8 lines · Joe Perches
Quoted 5 lines · Jean Delvare
Quoted 2 lines

If indent is not found, bail out immediately instead of spitting random shell script error messages.

OK, but can't we just delete Lindent instead?

Because...?

It's just not very useful in today's development space.

indent is quite bad at handling long lines.

It wraps code at arbitrary points to fit a column boundary rather than for readability or sensibility.

Code that may have deep indentation using a few spaces per block level can look horrible post Lindent.

For instance, reiserfs commit bd4c625c061c ("reiserfs: run scripts/Lindent on reiserfs code")

There are other tools that reflow whitespace styles.

#4 lore raw
Quoted 10 lines · On Fri, 10 Jul 2015 10:04:07 -0700 Joe Perches <[email protected]> wrote:
Quoted 8 lines

Le Friday 10 July 2015 __ 04:51 -0700, Joe Perches a __crit :

Quoted 5 lines · Jean Delvare
Quoted 2 lines

If indent is not found, bail out immediately instead of spitting random shell script error messages.

OK, but can't we just delete Lindent instead?

Because...?

It's just not very useful in today's development space.

I've very occasionally used Lindent. It's useful if the input is an utter mess. You feed it through Lindent as a first pass then get in and do the remainder by hand.

It can be less work than doing the whole conversion by hand.

#5 lore raw
Quoted 11 lines · Andrew Morton
Quoted 9 lines · On Fri, 10 Jul 2015 10:04:07 -0700 Joe Perches <[email protected]> wrote:
Quoted 8 lines

Le Friday 10 July 2015 __ 04:51 -0700, Joe Perches a __crit :

Quoted 5 lines · Jean Delvare
Quoted 2 lines

If indent is not found, bail out immediately instead of spitting random shell script error messages.

OK, but can't we just delete Lindent instead?

Because...?

Quoted 7 lines
Quoted 1 line

It's just not very useful in today's development space.

I've very occasionally used Lindent. It's useful if the input is an utter mess. You feed it through Lindent as a first pass then get in and do the remainder by hand.

It can be less work than doing the whole conversion by hand.

That's true, it can be, but I think Lindent mostly doesn't work particularly well for reviewing and it can require a lot more rework.

My biggest complaint about Lindent is that it can produce _awful_ looking code when it has to wrap longish lines.

I think that generally, checkpatch --fix-inplace works better and it can work in discrete steps.

I submitted a little script a while back that does most of what Lindent does.

https://lkml.org/lkml/2014/7/11/794

uncrustify also kinda works without the line wrapping nuttiness. It's not very good about using Linux's pointer location style.

http://uncrustify.sourceforge.net/

clang-format works reasonably well.
It can respect existing line wrapping.

http://clang.llvm.org/docs/ClangFormatStyleOptions.html

#6 lore raw
Quoted 18 lines · Joe Perches
Quoted 17 lines · Andrew Morton
Quoted 10 lines · On Fri, 10 Jul 2015 10:04:07 -0700 Joe Perches <[email protected]> wrote:
Quoted 8 lines

Le Friday 10 July 2015 __ 04:51 -0700, Joe Perches a __crit :

Quoted 5 lines · Jean Delvare
Quoted 2 lines

If indent is not found, bail out immediately instead of spitting random shell script error messages.

OK, but can't we just delete Lindent instead?

Because...?

It's just not very useful in today's development space.

I've very occasionally used Lindent. It's useful if the input is an utter mess. You feed it through Lindent as a first pass then get in and do the remainder by hand.

Quoted 5 lines
Quoted 1 line

It can be less work than doing the whole conversion by hand.

That's true, it can be, but I think Lindent mostly doesn't work particularly well for reviewing and it can require a lot more rework.

Well, this is a tool for submitters, not reviewers, so no surprise here.

Quoted 6 lines

My biggest complaint about Lindent is that it can produce _awful_ looking code when it has to wrap longish lines.

I think that generally, checkpatch --fix-inplace works better and it can work in discrete steps.

Thanks for the hint, this is indeed useful. Using the same tool to fix-up formatting and to check for incoming patches makes a lot of sense. However it seems that checkpatch.pl does not fix curly brace placement, nor does it delete blank lines at end of files. So indent can still do. So Lindent (or just indent actually, see below) can still be useful, at least until the missing features are added to checkpatch.

Quoted 9 lines

I submitted a little script a while back that does most of what Lindent does.

https://lkml.org/lkml/2014/7/11/794

uncrustify also kinda works without the line wrapping nuttiness. It's not very good about using Linux's pointer location style.

Quoted 6 lines

http://uncrustify.sourceforge.net/

clang-format works reasonably well.
It can respect existing line wrapping.

http://clang.llvm.org/docs/ClangFormatStyleOptions.html

Good to know, but unless these tools come with a preset for the Linux kernel style, or the kernel itself provides it (as Lindent does for for indent) I'm afraid it won't be too useful in practice.

BTW I noticed while looking for other options that recent versions of indent (as of November 2007) have a "--linux-style" option which serves the same purpose as Lindent. The result is almost the same in my case, but --linux-style translates to a much longer list of options. Also some options seem better than what Lindent uses (e.g. -il1 instead of -il0.)

In the light of this I would tend to agree with Joe that Lindent could go away and references to it be replaced with "indent --linux-style".

Signature · 2 lines
Jean Delvare
SUSE L3 Support