Skip to contentSkip to conversation

DSPBRIDGE: Kill CAMelCaSiNg

RFC Nishanth Menon 3 participants 4 messages Sep 17, 2009 – Sep 21, 2009

lore
Share & tools
lore
#0 lore raw

Hi All,
Ref [1] w.r.t some patchset seen on linux driver project having a similar condition of ton of camelcasing and [2],

Can we consider running the following script for DSPBridge code? or if there are some more improvements folks can think of..

-----
#!/bin/bash
# Silly little script to replace CaMELCaSInG to lower_casing # Half a zillion thoughts later,

# why not use ctags which has c interpreter already to do
# function identification for us?
if [ ! -r 'tags'  ]; then
echo "Sorry.. no tags file!"
exit 1
fi
if [ ! -r "$1" ]; then
echo "I need a directory or a file for replacing.."
exit 2
fi
DIR_FILE=`echo "$1"|sed -e "s/\//\\\//g"`
# tags have the following notation in field 4 (tab seperated)
# e - enum value
# d - define
# f - function
# F - File itself
# g - enum type
# m - structure parameter
# p - prototype
# r - kconfig define(refered)
# s - structure type
# t - typedefined variable
# v - variable
# x - extern defined variable
# we can choose to have different rules:
# rename rule i am going to follow:
# Any define -> convert camelcase to lower_case_define->convert to caps
# Any function, enum,struct param,prototype,struct type, typedef var,extern def ->
# camel_case to lower_case_define
# All files will be made unix file only (no point in dos + unix mixtures..)

#**
#* @brief - changes camel casing to ones with smaller case
#*
camel_to_norm(){
echo $1|tr '\_' '+' |sed 's/\([A-Z]\)/_\l\1/g' | sed 's/^_\([a-z]\)/\1/g' |\
tr 'A-Z' 'a-z' | sed "s/\([a-z][a-z]\+\)/-\1-/g"|tr -d '_'|\
tr '-' '_'|sed -e "s/_$//g"|sed -e "s/^_//g" | tr '+' '_' |\
sed -e "s/__*/_/g"|sed -e "s/_\([0-9][0-9]*\)/\1/g"|\
sed -e "s/^[a-z]\_\([a-z][a-z]\+\)/\1/g"

}

#**
#* @brief - all_to_upper - Moves all to upper case #*
all_to_upper(){
camel_to_norm $1 | tr 'a-z' 'A-Z'
}

#**
#* @brief generate_list - will generate list of symbols from tags matching #*
generate_list() {
cut -d' ' -f1,4 $TMPFILE|grep "$1\$"|cut -d' ' -f1|sort|uniq
}

#**
#* @brief find_files - finds list of files from tag file matching usage #*
find_files(){
cut -d' ' -f1,2 $TMPFILE|grep "^$1"|cut -d' ' -f2|sort|uniq }
TMPFILE=/tmp/cleaner.$$.tmp
TMPFILE1=/tmp/cleaner.$$.1.tmp

# grab the output for only the ones we want..
grep  "$DIR_FILE" tags |grep -v "^\!"|sed -e "s/\/\^.*\"/REPLACE_1/g">$TMPFILE

replace_pattern(){
echo "=====>Search and replace for pattern $1 with $2"
# Convert the functions back to smaller case and replace them in the files for token in generate_list "[$1]"
do
if [ $2 -eq "1" ]; then
new=camel_to_norm $token
else
new=all_to_upper $token
fi
if [ "$new" != "$token" ]; then
echo "==>Replacing $token with $low"
for file in find_files $low
do
echo "->replacing $token to $new in file $file"
cp $file $TMPFILE1
dos2unix $TMPFILE1
sed -e "s/\(W*\)$token\(\W*\)/\1$new\2/g" $TMPFILE1>$file
done
else
echo "==>not replacing $token"
fi
done
}
echo "================= Replacing Defines ===============" replace_pattern d 2
echo "================= Replacing Functions ===============" replace_pattern efgmpstvx 1
rm -f $TMPFILE $TMPFILE1

-----

Signature · 5 lines
Regards,
Nishanth Menon
Ref:
[1] http://driverdev.linuxdriverproject.org/pipermail/devel/2009-September/002482.html
[2] http://nishanthmenon.blogspot.com/2009/08/replace-camelcasing-with-lowercasing.html
#1 lore raw
Quoted 6 lines · Nishanth Menon

Hi All,
Ref [1] w.r.t some patchset seen on linux driver project having a similar condition of ton of camelcasing and [2],

Can we consider running the following script for DSPBridge code? or if there are some more improvements folks can think of..

Yes, I think we should do that, and make some sort of release, possibly squashing all the changes... my current branch counts 218 commits.

Quoted 56 lines

-----
#!/bin/bash
# Silly little script to replace CaMELCaSInG to lower_casing # Half a zillion thoughts later,

# why not use ctags which has c interpreter already to do
# function identification for us?
if [ ! -r 'tags'  ]; then
echo "Sorry.. no tags file!"
exit 1
fi
if [ ! -r "$1" ]; then
echo "I need a directory or a file for replacing.."
exit 2
fi
DIR_FILE=`echo "$1"|sed -e "s/\//\\\//g"`
# tags have the following notation in field 4 (tab seperated)
# e - enum value
# d - define
# f - function
# F - File itself
# g - enum type
# m - structure parameter
# p - prototype
# r - kconfig define(refered)
# s - structure type
# t - typedefined variable
# v - variable
# x - extern defined variable
# we can choose to have different rules:
# rename rule i am going to follow:
# Any define -> convert camelcase to lower_case_define->convert to caps
# Any function, enum,struct param,prototype,struct type, typedef var,extern
def ->
# camel_case to lower_case_define
# All files will be made unix file only (no point in dos + unix mixtures..)

#**
#* @brief - changes camel casing to ones with smaller case
#*
camel_to_norm(){
echo $1|tr '\_' '+' |sed 's/\([A-Z]\)/_\l\1/g' | sed 's/^_\([a-z]\)/\1/g' |\
 tr 'A-Z' 'a-z' | sed "s/\([a-z][a-z]\+\)/-\1-/g"|tr -d '_'|\
 tr '-' '_'|sed -e "s/_$//g"|sed -e "s/^_//g" | tr '+' '_' |\
 sed -e "s/__*/_/g"|sed -e "s/_\([0-9][0-9]*\)/\1/g"|\
 sed -e "s/^[a-z]\_\([a-z][a-z]\+\)/\1/g"

}

#**
#* @brief - all_to_upper - Moves all to upper case #*
all_to_upper(){
camel_to_norm $1 | tr 'a-z' 'A-Z'
}

Quoted 6 lines

#**
#* @brief generate_list - will generate list of symbols from tags matching #*
generate_list() {
cut -d' ' -f1,4 $TMPFILE|grep "$1\$"|cut -d' ' -f1|sort|uniq
}

It looks like my tag file was tabs instead of spaces, so that line makes everything explode... how about:
awk "\$4 ~ /[$1]/ { print \$1 }" $TMPFILE | sort | uniq

Quoted 6 lines

#**
#* @brief find_files - finds list of files from tag file matching usage #*
find_files(){
cut -d' ' -f1,2 $TMPFILE|grep "^$1"|cut -d' ' -f2|sort|uniq }

Maybe:
awk "\$1 ~ /^$1/ { print \$2 }" $TMPFILE | sort | uniq

Quoted 6 lines

TMPFILE=/tmp/cleaner.$$.tmp
TMPFILE1=/tmp/cleaner.$$.1.tmp

# grab the output for only the ones we want..
grep  "$DIR_FILE" tags |grep -v "^\!"|sed -e
"s/\/\^.*\"/REPLACE_1/g">$TMPFILE
Quoted 12 lines

replace_pattern(){
echo "=====>Search and replace for pattern $1 with $2"
# Convert the functions back to smaller case and replace them in the files for token in generate_list "[$1]"
do
 if [ $2 -eq "1" ]; then
 new=camel_to_norm $token
 else
 new=all_to_upper $token
 fi
 if [ "$new" != "$token" ]; then
 echo "==>Replacing $token with $low"

There's no $low, I guess you meant $new.

Quoted 1 line

 for file in find_files $low

Did you mean $token?

Quoted 4 lines

 do
  echo "->replacing $token to $new in file $file"
  cp $file $TMPFILE1
  dos2unix $TMPFILE1

This is simpler:
dos2unix -n $file $TMPFILE1

However, I think dos2unix should be run separately.

Quoted 1 line

  sed -e "s/\(W*\)$token\(\W*\)/\1$new\2/g" $TMPFILE1>$file

This is simpler:
sed -e "s/\<$token\>/$new/g" $TMPFILE1 > $file

Quoted 3 lines

 done
 else
 echo "==>not replacing $token"

No need to say that.

Quoted 8 lines

 fi
done
}
echo "================= Replacing Defines ===============" replace_pattern d 2
echo "================= Replacing Functions ===============" replace_pattern efgmpstvx 1
rm -f $TMPFILE $TMPFILE1

I did a few more cleanups (yeah, I'm too picky) and here's the result. Unfortunately I still think it needs some work in the camel conversion (E_FOO, __FOOBAR__, etc.)

----
#!/bin/bash

# Silly little script to replace CaMELCaSInG to lower_casing # Half a zillion thoughts later,

# why not use ctags which has c interpreter already to do
# function identification for us?

if [ ! -r 'tags' ]; then
echo "Sorry.. no tags file!"
exit 1
fi

if [ ! -r "$1" ]; then
echo "I need a directory or a file for replacing.." exit 2
fi

DIR_FILE=echo "$1" | sed -e "s/\//\\\//g"

# tags have the following notation in field 4 (tab seperated)
# e - enum value
# d - define
# f - function
# F - File itself
# g - enum type
# m - structure parameter
# p - prototype
# r - kconfig define(refered)
# s - structure type
# t - typedefined variable
# v - variable
# x - extern defined variable
# we can choose to have different rules:
# rename rule i am going to follow:
# Any define -> convert camelcase to lower_case_define->convert to caps
# Any function, enum,struct param,prototype,struct type, typedef
var,extern def ->
# camel_case to lower_case_define
# All files will be made unix file only (no point in dos + unix mixtures..)

#**
#* @brief - changes camel casing to ones with smaller case #*

camel_to_norm() {
	echo $1 | tr '\_' '+' | sed 's/\([A-Z]\)/_\l\1/g' | sed
's/^_\([a-z]\)/\1/g' | \
	tr 'A-Z' 'a-z' | sed "s/\([a-z][a-z]\+\)/-\1-/g" | tr -d '_'| \
	tr '-' '_' | sed -e "s/_$//g" | sed -e "s/^_//g" | tr '+' '_' | \
	sed -e "s/__*/_/g" | sed -e "s/_\([0-9][0-9]*\)/\1/g" | \
	sed -e "s/^[a-z]\_\([a-z][a-z]\+\)/\1/g"
}

#**
#* @brief - all_to_upper - Moves all to upper case #*

all_to_upper() {
	camel_to_norm $1 | tr 'a-z' 'A-Z'
}

#**
#* @brief generate_list - will generate list of symbols from tags matching #*

generate_list() {
	awk "\$4 ~ /[$1]/ { print \$1 }" $TMPFILE | sort | uniq
}

#**
#* @brief find_files - finds list of files from tag file matching usage #*

find_files() {
	awk "\$1 ~ /^$1/ { print \$2 }" $TMPFILE | sort | uniq
}

TMPFILE=/tmp/cleaner.$$.tmp
TMPFILE1=/tmp/cleaner.$$.1.tmp

# grab the output for only the ones we want..
grep "$DIR_FILE" tags | grep -v "^\!" | sed -e
"s/\/\^.*\"/REPLACE_1/g" > $TMPFILE
replace_pattern() {
	echo "=====> Search and replace for pattern $1 with $2"
	# Convert the functions back to smaller case and replace them in the files
	for token in `generate_list $1`
	do
		if [ $2 -eq "1" ]; then
			new=`camel_to_norm $token`
		else
			new=`all_to_upper $token`
		fi
		if [ "$new" != "$token" ]; then
			# echo "==> Replacing '$token' with '$new'"
			for file in `find_files $token`
			do
				echo "==> Replacing $token to $new in file $file"
				dos2unix -n $file $TMPFILE1
				sed -e "s/\<$token\>/$new/g" $TMPFILE1 > $file
			done
		fi
	done
}

echo "================= Replacing Defines ===============" replace_pattern d 2
echo "================= Replacing Functions ===============" replace_pattern efgmpstvx 1

rm -f $TMPFILE $TMPFILE1
----

Cheers.

Signature · 5 lines
Felipe Contreras
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#2 lore raw
Quoted 3 lines · Felipe Contreras

Yes, I think we should do that, and make some sort of release, possibly squashing all the changes... my current branch counts 218 commits.

I think it should be somehow announced before. Because, f.e., my patches is so big to rewrite them
(and most of them are accepted already, although I can't see commits in public repository)

Quoted 1 line

I did a few more cleanups (yeah, I'm too picky) and here's the result.

Actually even this could be simplified :-)

Quoted 6 lines

#**
#* @brief - changes camel casing to ones with smaller case #*
camel_to_norm() {
       echo $1 | tr '\_' '+' | sed 's/\([A-Z]\)/_\l\1/g' | sed 's/^_\([a-z]\)/\1/g' | \

I suppose there is could be used code like: sed -e '<expr1>; <expr2>' instead of two or more sed calls.

Quoted 1 line

       tr 'A-Z' 'a-z' | sed "s/\([a-z][a-z]\+\)/-\1-/g" | tr -d '_'| \

tr -d '_' is the 's/_//g' expression for sed

Quoted 3 lines

       tr '-' '_' | sed -e "s/_$//g" | sed -e "s/^_//g" | tr '+' '_' | \
       sed -e "s/__*/_/g" | sed -e "s/_\([0-9][0-9]*\)/\1/g" | \
       sed -e "s/^[a-z]\_\([a-z][a-z]\+\)/\1/g"

The similar. It could be optimized to a few sed calls (perhaps even to one call with all expressions in a sequence).

Quoted 3 lines
# grab the output for only the ones we want..
grep "$DIR_FILE" tags | grep -v "^\!" | sed -e
"s/\/\^.*\"/REPLACE_1/g" > $TMPFILE

sed could match and skip lines like grep -v 'smth.': sed -e '/^!/d; ...'

Quoted 2 lines

                               dos2unix -n $file $TMPFILE1
                               sed -e "s/\<$token\>/$new/g" $TMPFILE1 > $file

dos2unix actually one expression for sed, AFAIK.
And new sed versions (starting from 3.8?) have -i which means 'in place'.

P.S. And personally I prefer to use perl :-)

Signature · 6 lines
With Best Regards,
Andy Shevchenko
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#3 lore raw
Quoted 9 lines · Andy Shevchenko
Quoted 3 lines · Felipe Contreras

Yes, I think we should do that, and make some sort of release, possibly squashing all the changes... my current branch counts 218 commits.

I think it should be somehow announced before. Because, f.e., my patches is so big to rewrite them
(and most of them are accepted already, although I can't see commits in public repository)

Indeed. I'm planning on making my own "release" and pick your patches, but for a more "official" release we (or Omar) should make an official announcement so people send their patches before drawing the line.

Quoted 12 lines
Quoted 1 line

I did a few more cleanups (yeah, I'm too picky) and here's the result.

Actually even this could be simplified :-)

Quoted 6 lines

#**
#* @brief - changes camel casing to ones with smaller case #*
camel_to_norm() {
       echo $1 | tr '\_' '+' | sed 's/\([A-Z]\)/_\l\1/g' | sed 's/^_\([a-z]\)/\1/g' | \

I suppose there is could be used code like: sed -e '<expr1>; <expr2>' instead of two or more sed calls.

Quoted 8 lines
Quoted 1 line

       tr 'A-Z' 'a-z' | sed "s/\([a-z][a-z]\+\)/-\1-/g" | tr -d '_'| \

tr -d '_' is the 's/_//g' expression for sed

Quoted 3 lines

       tr '-' '_' | sed -e "s/_$//g" | sed -e "s/^_//g" | tr '+' '_' | \
       sed -e "s/__*/_/g" | sed -e "s/_\([0-9][0-9]*\)/\1/g" | \
       sed -e "s/^[a-z]\_\([a-z][a-z]\+\)/\1/g"

The similar. It could be optimized to a few sed calls (perhaps even to one call with all expressions in a sequence).

Yeah, I think this code should be re-factored, but for that I guess we'll need a list of simple checks to make sure that we get what we want.

Quoted 8 lines
Quoted 3 lines
# grab the output for only the ones we want..
grep "$DIR_FILE" tags | grep -v "^\!" | sed -e
"s/\/\^.*\"/REPLACE_1/g" > $TMPFILE

sed could match and skip lines like grep -v 'smth.': sed -e '/^!/d; ...'

Quoted 2 lines

                               dos2unix -n $file $TMPFILE1
                               sed -e "s/\<$token\>/$new/g" $TMPFILE1 > $file

dos2unix actually one expression for sed, AFAIK.

Yes, but dos2unix (or equivalent) will generate a *huge* diff, that's why I think it should be done in a different step.

Quoted 1 line

And new sed versions (starting from 3.8?) have -i which means 'in place'.

That would be better IMHO.

Quoted 1 line

P.S. And personally I prefer to use perl :-)

For this kind of usage I would prefer it as well (if the code doesn't become obfuscated), but I'd prefer ruby myself :)

Cheers.

Signature · 5 lines
Felipe Contreras
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html