qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Markus Armbruster <armbru@redhat.com>
To: Eric Blake <eblake@redhat.com>
Cc: pbonzini@redhat.com, qemu-devel@nongnu.org, mdroth@linux.vnet.ibm.com
Subject: Re: [PATCH 7/6] Makefile: Make Makefile depend on generated qga files, too
Date: Wed, 04 Dec 2019 14:19:14 +0100	[thread overview]
Message-ID: <87blso9qjx.fsf@dusky.pond.sub.org> (raw)
In-Reply-To: <42dbda81-f1a0-bb15-cee6-3040f41f38ef@redhat.com> (Eric Blake's message of "Wed, 4 Dec 2019 06:58:16 -0600")

Eric Blake <eblake@redhat.com> writes:

> On 12/4/19 12:56 AM, Markus Armbruster wrote:
>
>>>> +++ b/Makefile
>>>> @@ -130,6 +130,15 @@ GENERATED_QAPI_FILES += qapi/qapi-doc.texi
>>>>      generated-files-y += $(GENERATED_QAPI_FILES)
>>>>    +GENERATED_QGA_FILES := qga-qapi-types.c qga-qapi-types.h
>>>> +GENERATED_QGA_FILES += qga-qapi-visit.c qga-qapi-visit.h
>>>> +GENERATED_QGA_FILES += qga-qapi-commands.h qga-qapi-commands.c
>>>> +GENERATED_QGA_FILES += qga-qapi-init-commands.h qga-qapi-init-commands.c
>>>> +GENERATED_QGA_FILES += qga-qapi-doc.texi
>>>> +GENERATED_QGA_FILES := $(addprefix qga/qapi-generated/, $(GENERATED_QGA_FILES))
>>>
>>> Would it be worth using two separate variable names (maybe
>>> GENERATED_QGA_BASEFILES for the first list) rather than exploiting the
>>> arcane knowledge that consecutive use of := causes GNU make to rewrite
>>> an existing variable with new contents?
>>
>> Our rules.mak relies on this already.  It's full of magic, which
>> admittedly diminishes its suitability to serve as a good example.
>>
>> Your worry might be rooted in old "=" burns.  "=" makes the variable
>> recursively expanded, and
>>
>>      GENERATED_QGA_FILES = $(addprefix qga/qapi-generated/, $(GENERATED_QGA_FILES))
>
> Indeed, but I have to refer to the manual to remind myself of whether
> = or := is what I want in a given situation.

Trust me, you're either sure you want "=", or you want ":=".

On a green field, I recommend a hard rule "no = without a comment
explaining why".

>> would be an infinite loop.  ":=" makes it simply expanded; there's not
>> even a loop, let alone an infinite one.  The GNU Make manual explains
>> this clearly at
>> https://www.gnu.org/software/make/manual/html_node/Flavors.html
>>
>> Aside: there's a reason one of the two flavors is called "simple".  It
>> could additionally be called "not as slow".  One of my pet makefile
>> peeves: unthinking use of recursively expanded variables, complicating
>> semantics and slowing down builds.
>>
>> Back to this patch.  I had started to write the thing in longhand, but
>> got tired of repeating qga/qapi-generated/, so I factored that out.
>> Would longhand be easier to understand?
>
> It's more verbose.  My suggestion was more:
>
> GENERATED_QGA_BASENAMES := qga-qapi-types.c qga-qapi-types.h
> GENERATED_QGA_BASENAMES += qga-qapi-visit.c qga-qapi-visit.h
> ...
> GENERATED_QGA_FILES := $(addprefix qga/qapi-generated/,
> $(GENERATED_QGA_BASENAMES))
>
> to avoid the reassignment-to-self issue altogether, while still
> remaining concise compared to longhand.

Either way, we use multiple assignments to build GENERATED_QGA_FILES.
The only difference is that the version using two variables would also
work with recursive expansion, due to the magic of +=.



  reply	other threads:[~2019-12-04 13:20 UTC|newest]

Thread overview: 23+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-11-20 18:25 [PATCH 0/6] qapi: Module fixes and cleanups Markus Armbruster
2019-11-20 18:25 ` [PATCH 1/6] qapi: Tweak "command returns a nice type" check for clarity Markus Armbruster
2019-11-20 19:08   ` Eric Blake
2019-11-20 18:25 ` [PATCH 2/6] tests/Makefile.include: Fix missing test-qapi-emit-events.[ch] Markus Armbruster
2019-11-20 19:16   ` Eric Blake
2019-11-20 18:25 ` [PATCH 3/6] qapi: Generate command registration stuff into separate files Markus Armbruster
2019-11-20 19:26   ` Eric Blake
2019-11-27 16:12     ` Markus Armbruster
2019-11-20 18:25 ` [PATCH 4/6] qapi: Proper intermediate representation for modules Markus Armbruster
2019-11-20 20:31   ` Eric Blake
2019-11-20 18:25 ` [PATCH 5/6] qapi: Fix code generation for empty modules Markus Armbruster
2019-11-20 20:35   ` Eric Blake
2019-11-20 18:25 ` [PATCH 6/6] qapi: Simplify QAPISchemaModularCVisitor Markus Armbruster
2019-11-20 20:54   ` Eric Blake
2019-11-20 19:40 ` [PATCH 0/6] qapi: Module fixes and cleanups Markus Armbruster
2019-11-20 23:58 ` no-reply
2019-11-21  6:27 ` Markus Armbruster
2019-11-29  9:59 ` [PATCH 7/6] Makefile: Make Makefile depend on generated qga files, too Markus Armbruster
2019-12-03 22:33   ` Eric Blake
2019-12-04  6:56     ` Markus Armbruster
2019-12-04 12:58       ` Eric Blake
2019-12-04 13:19         ` Markus Armbruster [this message]
2019-12-03  9:01 ` [PATCH 0/6] qapi: Module fixes and cleanups Markus Armbruster

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=87blso9qjx.fsf@dusky.pond.sub.org \
    --to=armbru@redhat.com \
    --cc=eblake@redhat.com \
    --cc=mdroth@linux.vnet.ibm.com \
    --cc=pbonzini@redhat.com \
    --cc=qemu-devel@nongnu.org \
    /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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).