GCC: how to compile library into one C file? - gcc

So we have for example ffmpeg that can be compiled into > than 4 libraries. Is it possible using GCC to compile\collect not into .a or .so libraries but into some wary large one C files?


Funny question, why would you want to do this? Anyway, you can try and see whether gcc -E does what you want - it only invokes the preprocessor. So you might have to adapt the source code to include other source code files.
MiLu#Dago: ~ > cat x.c
int x_num() {
return 7;
}
MiLu#Dago: ~ > cat moin.c
#include <stdio.h>
#include "x.c"
int main() {
printf("moin: %d\n", x_num());
return 0;
}
MiLu#Dago: ~ > gcc -E moin.c | wc -l
1150
I think I'd rather try the cat utility:
find -type f -name \*.c | xargs cat > all_my.c
Still, what are you trying to achieve?


You may use cpp with -M family of arguments to generate a make file describing the code dependencies, then add a rule for .c and .h files to append content to the same file.
If the application is not using #ifdef guards within .h files it may cause warnings about macros defined twice.

Related

grep "--exclude=" cancels "--include=" while file extensions filtering

I want to grep only header files in source code. And seems that it is well-known and already solved problem, but --include= doesn't work for me:
$ grep -rnI --include=\*.h 'static volatile'
net/ipv4/ipconfig.c:174:static volatile int ic_got_reply __initdata; /* Proto(s) that replied */
drivers/mtd/maps/sbc_gxx.c:83:static volatile int page_in_window = -1; // Current page in window.
drivers/parport/parport_mfc3.c:79:static volatile int dummy; /* for trigger readds */
^C
I expect that there will be only header files with .h extension.
By the way, --exclude= works fine:
$ grep -rnI --exclude=\*.c 'static volatile'
arch/mips/include/asm/mach-pmcs-msp71xx/msp_gpio_macros.h:63:static volatile u32 * const MSP_GPIO_DATA_REGISTER[] = {
arch/mips/include/asm/mach-pmcs-msp71xx/msp_gpio_macros.h:83:static volatile u32 * const MSP_GPIO_MODE_REGISTER[] = {
arch/arm/mach-w90x900/include/mach/uncompress.h:28:static volatile u32 * const uart_base = (u32 *)UART0_PA;
arch/arm/mach-gemini/include/mach/uncompress.h:19:static volatile unsigned long * const UART = (unsigned long *)GEMINI_UART_BASE;
^C
UPD:
$ grep --version
grep (GNU grep) 2.27
...
$ type grep
grep is aliased to `grep --color --exclude=tags --exclude-dir="build*"'
But it works well when alias is cancelled such way: $ \grep -rnI --include=\*.h 'static volatile'.
So why does --exclude= cancels --include=?

For alias overriding, see: how can I override alias set in .bash_aliases
I couldn't find documentation relating to how multiple include/exclude rules affect the final list of files chosen for matching. But I think it is safe to assume that in case of conflict, the rule defined latest in the command line wins.
--exclude=tags will ignore any file or directory whose name is exactly tags
--include='*.h' will allow any file ending with .h to be matched
In this case, there is no conflict between the two rules. The first one allows grep to choose any file not named tags. The second one tells grep to allow any file ending with .h
In case of a conflict like --exclude='*foo.txt' --include='*.txt' the include rule will override the exclude rule and a file ending with foo.txt will be considered for matching

If you just use find to find files and grep to g/re/p within files (there are subtle clues in the names of the tools as to their purpose!) then you won't have to deal with the confusing, muddy, complexity the GNU guys created when they gave grep options to find files.
find . -type f -name '*.h' -exec grep -nIH 'static volatile' {} +
is probably all you need.

How to place ELF sections using the GNU linker at addresses according to their name?

I'm trying to put together a GNU linker script that places two types of sections:
the normal sections at fixed and predefined addresses. .text at 0x100000, .data at 0x200000, .bss at 0x300000 and so on
special sections according to their name. for example __at_0x13370000 should be placed at address 0x13370000
the first type is trivial and is explained everywhere. no problem there.
the issue that I have is with the second type. I can't figure out how to process a section's name to turn it into the address at which it is placed. the __at_ prefix used in the example above could be chosen freely.
I was previously using the ARM RVCT linker (armlink) where any section named in the schema of .ARM.__at_xxxxxx (where xxxxxx is a decimal or hexadecimal value) was placed at the corresponding address.
http://infocenter.arm.com/help/topic/com.arm.doc.dui0803e/pge1362066000571.html
I would like to switch to the GCC toolchain and the GNU linker, but in order for that to be possible, I need to somehow replicate this behaviour.
Thanks

You can extract section addresses from object files before linking and generate a small addend for linker script:
readelf -t *.o \
| grep -o '^ *\[ *[0-9]\+\] \.[a-zA-Z_0-9]\+_at_[x0-9]\+' \
| sed 's!.*\(\..*\)_at_\([x0-9]\+\).*!\1_at_\2 \2!' \
| sort -u \
| while read name addr; do
cat <<EOF
$name $addr:
{
KEEP(*($name)) ;
}
EOF
done > fixed.lds
fixed.lds can now be INCLUDEd into main linker script.

Print out all necessary for compilation files

I have bunch of Make files though directory hierarchy, some of them define C (CPP) and H files to use in compilations, some of them not, that are not adds to CPFLAFS -I../../SomeDir1 -I../SomeDir2/SomeDir2/Inc. etc.
Some of the Make files are called via make -C PATH Target .
The thing is hierarchy of sources grew for years and covers different projects and target platforms.
Now there are no people left in the company who remember "why that" and there is not enough documentation left.
The question: is there way to know particular sources and header files used to build the final target "all" ?
Thank you

There are ways, I know of two. Both require that you run make all, and before that remove all generated files so every file in question would actually be rebuilt (usually make clean).
Using strace
If strace is available, you can find the files by tracing and analyzing the system calls.
You can do it like this:
$ strace -f -o trace make all
$ grep 'open.*\.[ch]"' trace | grep -v 'open("/\(tmp\|usr\)' | sed -e 's/.*open("//' -e 's/".*//' | sort -u
Using Dependency Files (doesn't work with all compilers)
If your compiler can generate dependency files (i.e. gcc, armcc), you can analyze the dependency files. Make sure that dependency files are generated by putting the corresponding option in your CFLAGS or CPPFLAGS. For gcc, this would be -MMD. If Makefiles are written nicely, you can set CPPFLAGS on the command line, I will use this example here.
$ make all CPPFLAGS+=-MMD
$ find -name "*.d" -exec cat {} \; | sed -e 's/[:\\]//' -e 's/ /\n/g' | grep '\.[ch]$' | sort -u

Sublime Text build file for C++11

I'm currently trying to set up my build system in Sublime Text to compile C++11 Code and additionally use libpqxx library for some Postgres database stuff.
Right now, without C++11, my g++ command looks like this:
"cmd": ["g++", "${file}", "-o", "${file_path}/${file_base_name}", "-I/usr/include/pqxx", "-lpqxx", "-lpq"]
Which works fine.
Now I add "-std=c++0x" to the line, and it doesn't work. However, I wrote a simple testprogram that has #include <regex> in it, startet out with a default build file and added the C++11 flag to look like this:
"cmd": ["g++", "-std=c++0x", "${file}", "-o", "${file_path}/${file_base_name}"]
It works fine.
I also tried running the build from terminal, using this line:
g++ Connect.cpp -o Connect -std=c++0x -I/usr/include/pqxx -lpqxx, -lpq
But with no luck either.
Further I tried different places for the -std=c++0x. Right after g++, after the input and output file names, after the include, after the libary linking. And I trief to alter the std flag to -std+gnu++11, -std=c++11, -std=gnu++0x. Nothing helped.
What am I doing wrong?
(by the way my GCC/G++ has version 4.9, so there shouldn't be any problems with C++11 itself. The testfile.cpp without lipqxx include and linking works as mentioned before)
The solution: -I/usr/include/pqxx needed to be taken out.

Indeed, the solution is to remove the -I/usr/include/pqxx. But why?
The file he wants to include is /usr/include/pqxx/pqxx.
He is including it with #include <pqxx/pqxx> (He didnt provide that information on stackoverflow). With this include statement he doesn't need the additional include path since /usr/include is in the standard path.
But why is the build failing?
% ~ cpp -Wall -std=c++11 -M test.cpp > test2
% ~ cpp -Wall -std=c++11 -I/usr/include/pqxx -M test.cpp > test
% ~ cat test2 | sed "s/ \\\\//g" | sed "s/ /\n/g" | sort > test2
% ~ cat test | sed "s/ \\\\//g" | sed "s/ /\n/g" | sort > test
% ~ diff test test2
235d234
<
270a270
> /usr/include/c++/4.9.0/array
404a405
> /usr/include/c++/4.9.0/tuple
496d496
< /usr/include/pqxx/tuple
The -M option outputs a makefile.compatible list of headerfiles. The rest is just reformating to have a nicer diff.
Result: with -I/usr/include/pqxx the pqxx/tuple is included instead of the standard c++ tuple. That causes the compilation errors he didnt provide here.

How do you get the list of targets in a makefile?

I've used rake a bit (a Ruby make program), and it has an option to get a list of all the available targets, eg
> rake --tasks
rake db:charset # retrieve the charset for your data...
rake db:collation # retrieve the collation for your da...
rake db:create # Creates the databases defined in y...
rake db:drop # Drops the database for your curren...
...
but there seems to be no option to do this in GNU make.
Apparently the code is almost there for it, as of 2007 - http://www.mail-archive.com/help-make#gnu.org/msg06434.html.
Anyway, I made little hack to extract the targets from a makefile, which you can include in a makefile.
list:
#grep '^[^#[:space:]].*:' Makefile
It will give you a list of the defined targets. It's just a start - it doesn't filter out the dependencies, for instance.
> make list
list:
copy:
run:
plot:
turnin:

Under Bash (at least), this can be done automatically with tab completion:
make spacetabtab

Note: This answer has been updated to still work as of GNU make v4.3 - let us know if you come across something that breaks.
This is an attempt to improve on Brent Bradburn's great approach as follows:
uses a more robust command to extract the target names, which hopefully prevents any false positives (and also does away with the unnecessary sh -c)
does not invariably target the makefile in the current directory; respects makefiles explicitly specified with -f <file>
excludes hidden targets - by convention, these are targets whose name starts neither with a letter nor a digit
makes do with a single phony target
prefixes the command with # to prevent it from being echoed before execution
Curiously, GNU make has no feature for listing just the names of targets defined in a makefile. While the -p option produces output that includes all targets, it buries them in a lot of other information and also executes the default target (which could be suppressed with -f/dev/null).
Place the following rule in a makefile for GNU make to implement a target named list that simply lists all target names in alphabetical order - i.e.: invoke as make list:
.PHONY: list
list:
#LC_ALL=C $(MAKE) -pRrq -f $(lastword $(MAKEFILE_LIST)) : 2>/dev/null | awk -v RS= -F: '/(^|\n)# Files(\n|$$)/,/(^|\n)# Finished Make data base/ {if ($$1 !~ "^[#.]") {print $$1}}' | sort | egrep -v -e '^[^[:alnum:]]' -e '^$#$$'
Important: On pasting this, make sure that the last line is indented by exactly 1 actual tab char. (spaces do not work).
Note that sorting the resulting list of targets is the best option, since not sorting doesn't produce a helpful ordering in that the order in which the targets appear in the makefile is not preserved.
Also, the sub-targets of a rule comprising multiple targets are invariably output separately and will therefore, due to sorting, usually not appear next to one another; e.g., a rule starting with a z: will not have targets a and z listed next to each other in the output, if there are additional targets.
Explanation of the rule:
.PHONY: list
declares target list a phony target, i.e., one not referring to a file, which should therefore have its recipe invoked unconditionally
LC_ALL=C makes sure that make's output in in English, as parsing of the output relies on that.Tip of the hat to Bastian Bittorf
$(MAKE) -pRrq -f $(lastword $(MAKEFILE_LIST)) : 2>/dev/null
Invokes make again in order to print and parse the database derived from the makefile:
-p prints the database
-Rr suppresses inclusion of built-in rules and variables
-q only tests the up-to-date-status of a target (without remaking anything), but that by itself doesn't prevent execution of recipe commands in all cases; hence:
-f $(lastword $(MAKEFILE_LIST)) ensures that the same makefile is targeted as in the original invocation, regardless of whether it was targeted implicitly or explicitly with -f ....
Caveat: This will break if your makefile contains include directives; to address this, define variable THIS_FILE := $(lastword $(MAKEFILE_LIST)) before any include directives and use -f $(THIS_FILE) instead.
: is a deliberately invalid target that is meant to ensure that no commands are executed; 2>/dev/null suppresses the resulting error message. Note: This relies on -p printing the database nonetheless, which is the case as of GNU make 3.82. Sadly, GNU make offers no direct option to just print the database, without also executing the default (or given) task; if you don't need to target a specific Makefile, you may use make -p -f/dev/null, as recommended in the man page.
-v RS=
This is an awk idiom that breaks the input into blocks of contiguous non-empty lines.
/(^|\n)# Files(\n|$$)/,/(^|\n)# Finished Make data base/
Matches the range of lines in the output that contains all targets, across paragraphs - by limiting parsing to this range, there is no need to deal with false positives from other output sections.
Note: Between make versions 3.x and 4.3, paragraph structuring in make's output changed, so (^|\n) / (\n|$$) ensures that the lines that identify the start and the end of the cross-paragraph range of lines of interest are detected irrespective of whether they occur at the start or inside / at the end of a paragraph.
if ($$1 !~ "^[#.]")
Selectively ignores blocks:
# ... ignores non-targets, whose blocks start with # Not a target:
. ... ignores special targets
All other blocks should each start with a line containing only the name of an explicitly defined target followed by :
egrep -v -e '^[^[:alnum:]]' -e '^$#$$' removes unwanted targets from the output:
'^[^[:alnum:]]' ... excludes hidden targets, which - by convention - are targets that start neither with a letter nor a digit.
'^$#$$' ... excludes the list target itself
Running make list then prints all targets, each on its own line; you can pipe to xargs to create a space-separated list instead.

This obviously won't work in many cases, but if your Makefile was created by CMake you might be able to run make help.
$ make help
The following are some of the valid targets for this Makefile:
... all (the default if no target is provided)
... clean
... depend
... install
etc

I combined these two answers: https://stackoverflow.com/a/9524878/86967 and https://stackoverflow.com/a/7390874/86967
and did some escaping so that this could be used from inside a makefile.
.PHONY: no_targets__ list
no_targets__:
list:
sh -c "$(MAKE) -p no_targets__ | awk -F':' '/^[a-zA-Z0-9][^\$$#\/\\t=]*:([^=]|$$)/ {split(\$$1,A,/ /);for(i in A)print A[i]}' | grep -v '__\$$' | sort"
.
$ make -s list
build
clean
default
distclean
doc
fresh
install
list
makefile ## this is kind of extraneous, but whatever...
run

As mklement0 points out, a feature for listing all Makefile targets is missing from GNU-make, and his answer and others provides ways to do this.
However, the original post also mentions rake, whose tasks switch does something slightly different than just listing all tasks in the rakefile. Rake will only give you a list of tasks that have associated descriptions. Tasks without descriptions will not be listed. This gives the author the ability to both provide customized help descriptions and also omit help for certain targets.
If you want to emulate rake's behavior, where you provide descriptions for each target, there is a simple technique for doing this: embed descriptions in comments for each target you want listed.
You can either put the description next to the target or, as I often do, next to a PHONY specification above the target, like this:
.PHONY: target1 # Target 1 help text
target1: deps
[... target 1 build commands]
.PHONY: target2 # Target 2 help text
target2:
[... target 2 build commands]
...
.PHONY: help # Generate list of targets with descriptions
help:
#grep '^.PHONY: .* #' Makefile | sed 's/\.PHONY: \(.*\) # \(.*\)/\1 \2/' | expand -t20
Which will yield
$ make help
target1 Target 1 help text
target2 Target 2 help text
...
help Generate list of targets with descriptions
You can also find a short code example in this gist and here too.
Again, this does not solve the problem of listing all the targets in a Makefile. For example, if you have a big Makefile that was maybe generated or that someone else wrote, and you want a quick way to list its targets without digging through it, this won't help.
However, if you are writing a Makefile, and you want a way to generate help text in a consistent, self-documenting way, this technique may be useful.

My favorite answer to this was posted by Chris Down at Unix & Linux Stack Exchange. I'll quote.
This is how the bash completion module for make gets its list:
make -qp | awk -F':' '/^[a-zA-Z0-9][^$#\/\t=]*:([^=]|$)/ {split($1,A,/ /);for(i in A)print A[i]}'
It prints out a newline-delimited list of targets, without paging.
User Brainstone suggests piping to sort -u to remove duplicate entries:
make -qp | awk -F':' '/^[a-zA-Z0-9][^$#\/\t=]*:([^=]|$)/ {split($1,A,/ /);for(i in A)print A[i]}' | sort -u
Source: How to list all targets in make? (Unix&Linux SE)

If you have bash completion for make installed, the completion script will define a function _make_target_extract_script. This function is meant to create a sed script which can be used to obtain the targets as a list.
Use it like this:
# Make sure bash completion is enabled
source /etc/bash_completion
# List targets from Makefile
sed -nrf <(_make_target_extract_script --) Makefile

Focusing on an easy syntax for describing a make target, and having a clean output, I chose this approach:
help:
#grep -B1 -E "^[a-zA-Z0-9_-]+\:([^\=]|$$)" Makefile \
| grep -v -- -- \
| sed 'N;s/\n/###/' \
| sed -n 's/^#: \(.*\)###\(.*\):.*/\2###\1/p' \
| column -t -s '###'
#: Starts the container stack
up: a b
command
#: Pulls in new container images
pull: c d
another command
make-target-not-shown:
# this does not count as a description, so leaving
# your implementation comments alone, e.g TODOs
also-not-shown:
So treating the above as a Makefile and running it gives you something like
> make help
up Starts the container stack
pull Pulls in new container images
Explanation for the chain of commands:
First, grep all targets and their preceeding line, see https://unix.stackexchange.com/a/320709/223029.
Then, get rid of the group separator, see https://stackoverflow.com/a/2168139/1242922.
Then, we collapse each pair of lines to parse it later, see https://stackoverflow.com/a/9605559/1242922.
Then, we parse for valid lines and remove those which do not match, see https://stackoverflow.com/a/8255627/1242922, and also give the output our desired order: command, then description.
Lastly, we arrange the output like a table.

Add this target to your Makefile:
help:
#echo "\nTARGETS:\n"
#make -qpRr | egrep -e '^[a-z].*:$$' | sed -e 's~:~~g' | sort
#echo ""
make -qpRr = make --question --print-data-base --no-builtin-variables --no-builtin-rules
egrep -e '^[a-z].*:$$': searches for lines which start with lowercase and ends with ":"
sed -e 's~:~~g': deletes the ":"
Then just run:
make help
This works for me 😉
PD: more info at...
make --help

I took a few answers mentioned above and compiled this one, which can also generate a nice description for each target and it works for targets with variables too.
Example Makefile:
APPS?=app1 app2
bin: $(APPS:%=%.bin)
## Help: A composite target that relies only on other targets
$(APPS:%=%.bin): %.bin:
## Help: A target with variable name, value = $*
test:
## Help: A normal target without variables
# A target without any help description
clean:
# A hidden target
.hidden:
help:
#printf "%-20s %s\n" "Target" "Description"
#printf "%-20s %s\n" "------" "-----------"
#make -pqR : 2>/dev/null \
| awk -v RS= -F: '/^# File/,/^# Finished Make data base/ {if ($$1 !~ "^[#.]") {print $$1}}' \
| sort \
| egrep -v -e '^[^[:alnum:]]' -e '^$#$$' \
| xargs -I _ sh -c 'printf "%-20s " _; make _ -nB | (grep -i "^# Help:" || echo "") | tail -1 | sed "s/^# Help: //g"'
Example output:
$ make help
Target Description
------ -----------
app1.bin A target with variable name, value = app1
app2.bin A target with variable name, value = app2
bin A composite target that relies only on other targets
clean
test A normal target without variables
How does it work:
The top part of the make help target works exactly as posted by mklement0 here - How do you get the list of targets in a makefile?.
After getting the list of targets, it runs make <target> -nB as a dry run for each target and parses the last line that starts with ## Help: for the description of the target. And that or an empty string is printed in a nicely formatted table.
As you can see, the variables are even expanded within the description as well, which is a huge bonus in my book :).

#nobar's answer helpfully shows how to use tab completion to list a makefile's targets.
This works great for platforms that provide this functionality by default (e.g., Debian, Fedora).
On other platforms (e.g., Ubuntu) you must explicitly load this functionality, as implied by #hek2mgl's answer:
. /etc/bash_completion installs several tab-completion functions, including the one for make
Alternatively, to install only tab completion for make:
. /usr/share/bash-completion/completions/make
For platforms that don't offer this functionality at all, such as OSX, you can source the following commands (adapated from here) to implement it:
_complete_make() { COMPREPLY=($(compgen -W "$(make -pRrq : 2>/dev/null | awk -v RS= -F: '/^# File/,/^# Finished Make data base/ {if ($1 !~ "^[#.]") {print $1}}' | egrep -v '^[^[:alnum:]]' | sort | xargs)" -- "${COMP_WORDS[$COMP_CWORD]}")); }
complete -F _complete_make make
Note: This is not as sophisticated as the tab-completion functionality that comes with Linux distributions: most notably, it invariably targets the makefile in the current directory, even if the command line targets a different makefile with -f <file>.

This help target will only print targets which have ## followed by a description. This allows for documenting both public and private targets. Using the .DEFAULT_GOAL makes the help more discoverable.
Only sed, xargs and printf used which are pretty common.
Using the < $(MAKEFILE_LIST) allows for the makefile to be called something other than Makefile for instance Makefile.github
You can customize the output to suit your preference in the printf. This example is set up to match the OP's request for rake style output
When cutting and pasting the below make file, don't forget to change the 4 spaces indentation to tabs.
# vim:ft=make
# Makefile
.DEFAULT_GOAL := help
.PHONY: test help
help: ## these help instructions
#sed -rn 's/^([a-zA-Z_-]+):.*?## (.*)$$/"\1" "\2"/p' < $(MAKEFILE_LIST) | xargs printf "make %-20s# %s\n"
lint: ## style, bug and quality checker
pylint src test
private: # for internal usage only
#true
test: private ## run pytest with coverage
pytest --cov test
Here is the output from the Makefile above. Notice the private target doesn't get output because it only has a single # for it's comment.
$ make
make help # these help instructions
make lint # style, bug and quality checker
make test # run pytest with coverage

This is far from clean, but did the job, for me.
make -p 2&>/dev/null | grep -A 100000 "# Files" | grep -v "^$" | grep -v "^\(\s\|#\|\.\)" | grep -v "Makefile:" | cut -d ":" -f 1
I use make -p that dumps the internal database, ditch stderr, use a quick and dirty grep -A 100000 to keep the bottom of the output. Then I clean the output with a couple of grep -v, and finally use cut to get what's before the colon, namely, the targets.
This is enough for my helper scripts on most of my Makefiles.
EDIT: added grep -v Makefile that is an internal rule

For a Bash Script
Here's a very simple way to do this in bash -- based on the comment by #cibercitizen1 above:
grep : Makefile | awk -F: '/^[^.]/ {print $1;}'
See also the more authoritative answer by #Marc.2377, too, which says how the Bash completion module for make does it.

To expand on the answer given by #jsp, you can even evaluate variables in your help text with the $(eval) function.
The proposed version below has these enhanced properties:
Will scan any makefiles (even included)
Will expand live variables referenced in the help comment
Adds documentation anchor for real targets (prefixed with # TARGETDOC:)
Adds column headers
So to document, use this form:
RANDOM_VARIABLE := this will be expanded in help text
.PHONY: target1 # Target 1 help with $(RANDOM_VARIABLE)
target1: deps
[... target 1 build commands]
# TARGETDOC: $(BUILDDIR)/real-file.txt # real-file.txt help text
$(BUILDDIR)/real-file.txt:
[... $(BUILDDIR)/real-file.txt build commands]
Then, somewhere in your makefile:
.PHONY: help # Generate list of targets with descriptions
help:
## find all help in targets and .PHONY and evaluate the embedded variables
$(eval doc_expanded := $(shell grep -E -h '^(.PHONY:|# TARGETDOC:) .* #' $(MAKEFILE_LIST) | sed -E -n 's/(\.PHONY|# TARGETDOC): (.*) # (.*)/\2 \3\\n/'p | expand -t40))
#echo
#echo ' TARGET HELP' | expand -t40
#echo ' ------ ----' | expand -t40
#echo -e ' $(doc_expanded)'

make doesn't support this by default and other answers have shown how to extract the list of possible targets automatically.
However, in case you want to have more control with the listing without any side-effects (such as using the .PHONY target to mark the documentation which prevents the logic of using the target names as actual files which Make uses to decide which targets needs to be rebuilt), you can invent your own syntax just for the documentation. I prefer to use ### like this:
CPUS ?= $(shell nproc)
MAKEFLAGS += -j $(CPUS) -l $(CPUS) -s
# Basic paths
PREFIX ?= usr
BINDIR ?= $(PREFIX)/bin
ETCDIR ?= etc
MANDIR ?= $(PREFIX)/share/man
# ...
### help: Show help message (default target)
# use "help" as the default target (first target in the Makefile)
.PHONY: help
help:
#printf "%s\n\n" "make: List of possible targets:"
#grep '^### .*:' $(lastword $(MAKEFILE_LIST)) | sed 's/^### \([^:]*\): \(.*\)/\1:\t\2/' | column -ts "$$(printf '\t')"
### install: Install all files in $PREFIX (used by debian binary package build scripts)
install:
install -D -o root -g root -m 755 ...
...
### release: Increase package version number
release:
debchange --release
(as usual, the indented files must start with exactly one tabulator but stackoverflow cannot reproduce that detail correctly.)
Output will look like this:
$ make
make: List of possible targets:
help: Show help message (default target)
install: Install all files in $PREFIX (used by debian binary package build scripts)
release: Increase package version number
This works because only lines starting with ### and having a : character are considered as the documentation to output. Note that this intentionally does not extract the actual target name but fully trusts the documentation lines only. This allows always emitting correct output for very complex Makefile tricks, too. Also note that this avoids needing to put the documentation line on any specific position relative to actual rule. I also intentionally avoid sorting the output because the order of output can be fully controlled from the Makefile itself simply by listing the documentation lines in preferred order.
You could obviously invent any other syntax you like and even do something like
### en: install: Install all files in $PREFIX
### fi: asennus: asenna kaikki tiedostot hakemistoon $PREFIX
and only print lines that match the current locale to support multiple languages and having aliases to localize the target names, too:
.PHONY: asennus
asennus: install
The most important question is why do you want to list the targets? Do you want actual documentation or some kind of debugging information?

This is a modification to jsp's very helpful answer (https://stackoverflow.com/a/45843594/814145). I like the idea of getting not only a list of targets but also their descriptions. jsp's Makefile puts the description as the comment, which I found often will be repeated in the target's description echo command. So instead, I extract the description from the echo command for each target.
Example Makefile:
.PHONY: all
all: build
: "same as 'make build'"
.PHONY: build
build:
#echo "Build the project"
.PHONY: clean
clean:
#echo "Clean the project"
.PHONY: help
help:
#echo -n "Common make targets"
#echo ":"
#cat Makefile | sed -n '/^\.PHONY: / h; /\(^\t#*echo\|^\t:\)/ {H; x; /PHONY/ s/.PHONY: \(.*\)\n.*"\(.*\)"/ make \1\t\2/p; d; x}'| sort -k2,2 |expand -t 20
Output of make help:
$ make help
Common make targets:
make all same as 'make build'
make build Build the project
make clean Clean the project
make help Common make targets
Notes:
Same as jsp's answer, only PHONY targets may be listed, which may or may not work for your case
In addition, it only lists those PHONY targets that have a echo or : command as the first command of the recipe. : means "do nothing". I use it here for those targets that no echo is needed, such as all target above.
There is an additional trick for the help target to add the ":" in the make help output.

Plenty of workable solutions here, but as I like saying, "if it's worth doing once, it's worth doing again."
I did upvote the sugestion to use (tab)(tab), but as some have noted, you may not have completion support, or, if you have many include files, you may want an easier way to know where a target is defined.
I have not tested the below with sub-makes...I think it wouldn't work. As we know, recursive makes considered harmful.
.PHONY: list ls
ls list :
## search all include files for targets.
## ... excluding special targets, and output dynamic rule definitions unresolved.
#for inc in $(MAKEFILE_LIST); do \
echo ' =' $$inc '= '; \
grep -Eo '^[^\.#[:blank:]]+.*:.*' $$inc | grep -v ':=' | \
cut -f 1 | sort | sed 's/.*/ &/' | sed -n 's/:.*$$//p' | \
tr $$ \\\ | tr $(open_paren) % | tr $(close_paren) % \
; done
# to get around escaping limitations:
open_paren := \(
close_paren := \)
Which I like because:
list targets by include file.
output raw dynamic target definitions (replaces variable delimiters with modulo)
output each target on a new line
seems clearer (subjective opinion)
Explanation:
foreach file in the MAKEFILE_LIST
output the name of the file
grep lines containing a colon, that are not indented, not comments, and don't start with a period
exclude immediate assignment expressions (:=)
cut, sort, indent, and chop rule-dependencies (after colon)
munge variable delimiters to prevent expansion
Sample Output:
= Makefile =
includes
ls list
= util/kiss/snapshots.mk =
rotate-db-snapshots
rotate-file-snapshots
snap-db
snap-files
snapshot
= util/kiss/main.mk =
dirs
install
%MK_DIR_PREFIX%env-config.php
%MK_DIR_PREFIX%../srdb

This one was helpful to me because I wanted to see the build targets required (and their dependencies) by the make target. I know that make targets cannot begin with a "." character. I don't know what languages are supported, so I went with egrep's bracket expressions.
cat Makefile | egrep "^[[:alnum:][:punct:]]{0,}:[[:space:]]{0,}[[:alnum:][:punct:][:space:]]{0,}$"

Yet another additional answer to above.
tested on MacOSX using only cat and awk on terminal
cat Makefile | awk '!/SHELL/ && /^[A-z]/ {print $1}' | awk '{print substr($0, 1, length($0)-1)}'
will output of the make file like below:
target1
target2
target3
in the Makefile, it should be the same statement, ensure that you escape the variables using $$variable rather than $variable.
Explanation
cat - spits out the contents
| - pipe parses output to next awk
awk - runs regex excluding "shell" and accepting only "A-z" lines then prints out the $1 first column
awk - yet again removes the last character ":" from the list
this is a rough output and you can do more funky stuff with just AWK. Try to avoid sed as its not as consistent in BSDs variants i.e. some works on *nix but fails on BSDs like MacOSX.
More
You should be able add this (with modifications) to a file for make, to the default bash-completion folder /usr/local/etc/bash-completion.d/
meaning when you "make tab tab" .. it will complete the targets based on the one liner script.

For AWK haters, and for simplicity, this contraption works for me:
help:
make -qpRr $(lastword $(MAKEFILE_LIST)) | egrep -v '(^(\.|:|#|\s|$)|=)' | cut -d: -f1
(for use outside a Makefile, just remove $(lastword ...) or replace it with the Makefile path).
This solution will not work if you have "interesting" rule names but will work well for most simple setups. The main downside of a make -qp based solution is (as in other answers here) that if the Makefile defines variable values using functions - they will still be executed regardless of -q, and if using $(shell ...) then the shell command will still be called and its side effects will happen. In my setup often the side effects of running shell functions is unwanted output to standard error, so I add 2>/dev/null after the make command.

I usually do:
grep install_targets Makefile
It would come back with something like:
install_targets = install-xxx1 install-xxx2 ... etc
I hope this helps

tl;dr I personally copy-paste the same help target for every Makefile I build.
.SILENT:
.PHONY: help
## This help screen
help:
printf "Available targets\n\n"
awk '/^[a-zA-Z\-\_0-9]+:/ { \
helpMessage = match(lastLine, /^## (.*)/); \
if (helpMessage) { \
helpCommand = substr($$1, 0, index($$1, ":")-1); \
helpMessage = substr(lastLine, RSTART + 3, RLENGTH); \
printf "%-30s %s\n", helpCommand, helpMessage; \
} \
} \
{ lastLine = $$0 }' $(MAKEFILE_LIST)
I also maintain a copy of it in this Github gist:
https://gist.github.com/Olshansk/689fc2dee28a44397c6e31a0776ede30

Very simple AWK solution:
all:
#awk -F'[ :]' '!/^all:/ && /^([A-z_-]+):/ {print "make " $$1}' Makefile
(Note: This doesn't cover all the corner-cases as the accepted answer, as explained here.)

Try this one:
make -qp | awk -F':' '/^[^ \t.%][-A-Za-z0-9_]*:/ {split($0,A,/ /);for(i in A)if(match(A[i],/^[^.%][-A-Za-z0-9_]*/))print substr(A[i],1,RLENGTH)}' | sort -u

This is a very simplified version of what the bash-completion script does.
make -npq : 2> /dev/null | \
awk -v RS= -F: '$1 ~ /^[^#%]+$/ { print $1 }'
Explanation:
make -npq: Print the database without executing anything
-v RS=: Separate records by whole paragraphs
-F:: Separate fields by : (so the rule name is $1)
$1 ~ /^[^#%]+$/: Match rules that don't contain # or % (comments or pattern rules)
{ print $1 }: Print the rule name
This is much simpler than mklement0's approach (which I fixed myself), and works better.

not sure why the previous answer was so complicated:
list:
cat Makefile | grep "^[A-z]" | awk '{print $$1}' | sed "s/://g"

Resources