Minor issue pulling from gitlab

Got this:

Code:

---

make chkpaw --no-print-directory
make -f run.mak chkpaw
git -C ./cloned/mak pull --verbose
POST git-upload-pack (155 bytes)
From https://gitlab.com/awsdert/mak
 = [up to date]      master    -> origin/master
run.mak:6: *** missing separator.  Stop.
make: *** [GNUmakefile:9: chkpaw] Error 2
Compilation failed.

---

From this:

Code:

---

include dir.mak

pull_mak:=git -C $(MAK_DIR) pull --verbose

$(info $(pull_mak))
$(shell $(pull_mak))

---

I'm not seeing the error, I know this is technically unrelated to C but I figured my fellow members here have had similar experience in this problem, any ideas?

For reference dir.mak looks like this:

Code:

---

TOP_DIR?=.
BIN_DIR?=$(TOP_DIR)/bin
LIB_DIR?=$(TOP_DIR)/lib
SRC_DIR?=$(TOP_DIR)/src
OBJ_DIR?=$(TOP_DIR)/obj
INC_DIR?=$(TOP_DIR)/include
CLONED_DIR?=$(TOP_DIR)/cloned
MAK_DIR?=$(CLONED_DIR)/mak

ALL_DIR:=$(BIN_DIR) $(LIB_DIR) $(SRC_DIR) $(OBJ_DIR) $(INC_DIR) $(CLONED_DIR)

directories: $(ALL_DIR:%=%/) $(MAK_DIR)/

%/:
        mkdir

$(MAK_DIR)/:
        git -C $(CLONED_DIR) clone https://gitlab.com/awsdert/mak.git

.PHONY: directories

---

If I remember correctly, a common reason for that error is the makefile being indented by spaces instead of tabs.

Quote:

---

Originally Posted by laserlight

If I remember correctly, a common reason for that error is the makefile being indented by spaces instead of tabs.

---

Nope, that's definitely not it, just went checking and did not find a single space where there should be a tab, didn't expect to find any anyways as I prefer tabs to spaces, any other ideas?

Shouldn't lines 5 and 6 go in a rule or whatever they call it?

EDIT:
Oh, looks like line 5 doesn't have to go in a rule, but it makes sense to keep it with line 6.

Never mind, finally thought "Ah! maybe that's..." and did this instead:

Code:

---

include dir.mak

pull_mak:=git -C $(MAK_DIR) pull

$(info $(pull_mak))
$(info $(shell $(pull_mak)))

---

shell was returning output and make was treating that output as part of the make script

On windows, I would guess spaces in the path and try changing

Code:

---

git -C $(CLONED_DIR) clone https://gitlab.com/awsdert/mak.git

---

to

Code:

---

git -C "$(CLONED_DIR)" clone https://gitlab.com/awsdert/mak.git

---

Tim S.

Quote:

---

Originally Posted by laserlight

Shouldn't lines 5 and 6 go in a rule or whatever they call it?

EDIT:
Oh, looks like line 5 doesn't have to go in a rule, but it makes sense to keep it with line 6.

---

Just seen your comment, I'm assuming you were talking about the shell command right?

Quote:

---

Originally Posted by stahta01

On windows, I would guess spaces in the path and try changing

Code:

---

git -C $(CLONED_DIR) clone https://gitlab.com/awsdert/mak.git

---

to

Code:

---

git -C "$(CLONED_DIR)" clone https://gitlab.com/awsdert/mak.git

---

Tim S.

---

I don't use spaces in directory names, not really sure how to go about putting in the quotes for automatic variables though, all I can think of is this:

Code:

---

$(^:%="%")

---

Either way it turned out that wasn't the problem

Quote:

---

Originally Posted by awsdert

shell was returning output and make was treating that output as part of the make script

---

info is normally used to print some info rather than to actually do something though. Wouldn't it be simpler to just write say:

Code:

---

all:
    $(pull_mak)

---

EDIT:
Although I guess naming the target "pull" rather than "all" would make more sense. But then if you really want to always pull, then maybe your approach is okay? It seems strange to me though, since presumably there may be things you want to do with make that doesn't involve pulling from a git repo.

Quote:

---

Originally Posted by laserlight

info is normally used to print some info rather than to actually do something though. Wouldn't it be simpler to just write say:

Code:

---

all:
    $(pull_mak)

---

EDIT:
Although I guess naming the target "pull" rather than "all" would make more sense. But then if you really want to always pull, then maybe your approach is okay? It seems strange to me though, since presumably there may be things you want to do with make that doesn't involve pulling from a git repo.

---

This particular project that I'm pulling is where I'm starting to store all the makefile functions and per system/compiler variables plus the auto mappes variables, so in this case yeah an always pull is needed before they get used to ensure any improvements or additional mappings are used. It's also the same method I prefer to apply to any 3rd party project, at least the code will be current then

Quote:

---

Originally Posted by awsdert

Do you expect us to be phsycic? How are we to know what the "..." means?

---

That comment looks like a copy-paste of (some parts of) The difference between forking and cloning a repository.

Quote:

---

Originally Posted by christop

That comment looks like a copy-paste of (some parts of) The difference between forking and cloning a repository.

---

Oh wow, one of us was psychic, thx btw :)