All of lore.kernel.org
 help / color / mirror / Atom feed
From: Junio C Hamano <gitster@pobox.com>
To: Dave Borowitz <dborowitz@google.com>
Cc: git@vger.kernel.org, kusmabite@gmail.com
Subject: Re: [PATCH v3 1/2] Makefile: use curl-config to determine curl flags
Date: Mon, 28 Apr 2014 12:56:54 -0700	[thread overview]
Message-ID: <xmqqtx9dp6rd.fsf@gitster.dls.corp.google.com> (raw)
In-Reply-To: <1398713704-15428-1-git-send-email-dborowitz@google.com> (Dave Borowitz's message of "Mon, 28 Apr 2014 12:35:03 -0700")

Dave Borowitz <dborowitz@google.com> writes:

> Use this only when CURLDIR is not explicitly specified, to continue
> supporting older builds. Moreover, if CURL_CONFIG is unset or running
> it returns no results (e.g. because it is missing), default to the old
> behavior of blindly setting -lcurl.
>  	ifdef CURLDIR
> +		CURL_LIBCURL=
>  	else
> +		CURL_CONFIG ?= curl-config
> +		ifeq "$(CURL_CONFIG)" ""
> +			CURL_LIBCURL =
> +		else
> +			CURL_LIBCURL := $(shell $(CURL_CONFIG) --libs)
>  		endif

This "ifeq" is redundant and will never set CURL_LIBCURL to empty
without running the "else" part, I think.  In a Makefile, a variable
explicitly set to empty and a variable that is unset are treated the
same.

	$ cat >Makefile <<EOF
	CURL_CONFIG ?= curl-config
	ifeq "$(CURL_CONFIG)" ""
		X=Empty
	else
		X=NotEmpty
	endif

	ifdef "$(CURL_CONFIG)"
		Z=Defined
	else
		Z=Undefined
	endif

	all::
		@echo "$(X) $(Z)"
	EOF
	$ make -f Makefile CURL_CONFIG=""
	Empty Undefined

That does not mean the patch will give us a broken behaviour,
though.  It just means the ifeq/else part will be redundant.

>  	endif
> +
> +	ifeq "$(CURL_LIBCURL)" ""

This will catch the "$(shell $(CURL_CONFIG) --libs) assigned an
empty string to CURL_LIBCURL" case, so the result is good.

I haven't checked what it would look like if we turn this into an
incremental patch to be applied on top of 'master' (which would give
us a place to document better why we do not rely on the presense of
curl-config), but if we can do so, that would be more preferable
than having to revert the merge of the previous one and then
applying these two patches anew.

Thanks.

> +		ifdef CURLDIR
> +			# Try "-Wl,-rpath=$(CURLDIR)/$(lib)" in such a case.
> +			BASIC_CFLAGS += -I$(CURLDIR)/include
> +			CURL_LIBCURL = -L$(CURLDIR)/$(lib) $(CC_LD_DYNPATH)$(CURLDIR)/$(lib) -lcurl
> +		else
> +			CURL_LIBCURL = -lcurl
> +		endif
> +		ifdef NEEDS_SSL_WITH_CURL
> +			CURL_LIBCURL += -lssl
> +			ifdef NEEDS_CRYPTO_WITH_SSL
> +				CURL_LIBCURL += -lcrypto
> +			endif
> +		endif
> +		ifdef NEEDS_IDN_WITH_CURL
> +			CURL_LIBCURL += -lidn
> +		endif
> +	else
> +		BASIC_CFLAGS += $(shell $(CURL_CONFIG) --cflags)
>  	endif
>  
>  	REMOTE_CURL_PRIMARY = git-remote-http$X

  parent reply	other threads:[~2014-04-28 19:57 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-04-28 19:35 [PATCH v3 1/2] Makefile: use curl-config to determine curl flags Dave Borowitz
2014-04-28 19:35 ` [PATCH v3 2/2] Makefile: allow static linking against libcurl Dave Borowitz
2014-04-28 19:44 ` [PATCH v3 1/2] Makefile: use curl-config to determine curl flags Jonathan Nieder
2014-04-28 19:51   ` Dave Borowitz
2014-04-28 19:56 ` Junio C Hamano [this message]
2014-04-28 20:00   ` Junio C Hamano
2014-04-28 21:13   ` Junio C Hamano

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=xmqqtx9dp6rd.fsf@gitster.dls.corp.google.com \
    --to=gitster@pobox.com \
    --cc=dborowitz@google.com \
    --cc=git@vger.kernel.org \
    --cc=kusmabite@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.