qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: John Snow <jsnow@redhat.com>
To: Markus Armbruster <armbru@redhat.com>
Cc: qemu-devel@nongnu.org, Eduardo Habkost <ehabkost@redhat.com>,
	Cleber Rosa <crosa@redhat.com>
Subject: Re: [PATCH v5 11/36] qapi/common.py: Add indent manager
Date: Wed, 7 Oct 2020 14:08:33 -0400	[thread overview]
Message-ID: <0a507019-1e59-2967-9c7a-fa3712a74ae0@redhat.com> (raw)
In-Reply-To: <87d01uifmn.fsf@dusky.pond.sub.org>

On 10/7/20 4:50 AM, Markus Armbruster wrote:
> John Snow <jsnow@redhat.com> writes:
> 
>> Code style tools really dislike the use of global keywords, because it
>> generally involves re-binding the name at runtime which can have strange
>> effects depending on when and how that global name is referenced in
>> other modules.
>>
>> Make a little indent level manager instead.
>>
>> Signed-off-by: John Snow <jsnow@redhat.com>
>> Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
>> Reviewed-by: Cleber Rosa <crosa@redhat.com>
> 
> Intentation is a job for QAPIGen (and its subtypes).  But if this patch
> is easier to achieve this series' goal, I don't mind.
> 

I agree, but refactoring it properly is beyond my capacity right now.

This was the dumbest thing I could do to get pylint/mypy passing, which 
required the elimination (or suppression) of the global keyword.

Creating a stateful object was the fastest way from A to B.

>> ---
>>   scripts/qapi/common.py | 49 ++++++++++++++++++++++++++++--------------
>>   scripts/qapi/visit.py  |  7 +++---
>>   2 files changed, 36 insertions(+), 20 deletions(-)
>>
>> diff --git a/scripts/qapi/common.py b/scripts/qapi/common.py
>> index cee63eb95c7..b35318b72cf 100644
>> --- a/scripts/qapi/common.py
>> +++ b/scripts/qapi/common.py
>> @@ -93,33 +93,50 @@ def c_name(name, protect=True):
>>   pointer_suffix = ' *' + eatspace
>>   
>>   
>> -def genindent(count):
>> -    ret = ''
>> -    for _ in range(count):
>> -        ret += ' '
>> -    return ret
>> +class Indentation:
>> +    """
>> +    Indentation level management.
>>   
>> +    :param initial: Initial number of spaces, default 0.
>> +    """
>> +    def __init__(self, initial: int = 0) -> None:
>> +        self._level = initial
>>   
>> -indent_level = 0
>> +    def __int__(self) -> int:
>> +        return self._level
>>   
>> +    def __repr__(self) -> str:
>> +        return "{}({:d})".format(type(self).__name__, self._level)
>>   
>> -def push_indent(indent_amount=4):
>> -    global indent_level
>> -    indent_level += indent_amount
>> +    def __str__(self) -> str:
>> +        """Return the current indentation as a string of spaces."""
>> +        return ' ' * self._level
>>   
>> +    def __bool__(self) -> bool:
>> +        """True when there is a non-zero indentation."""
>> +        return bool(self._level)
>>   
>> -def pop_indent(indent_amount=4):
>> -    global indent_level
>> -    indent_level -= indent_amount
>> +    def increase(self, amount: int = 4) -> None:
>> +        """Increase the indentation level by ``amount``, default 4."""
>> +        self._level += amount
>> +
>> +    def decrease(self, amount: int = 4) -> None:
>> +        """Decrease the indentation level by ``amount``, default 4."""
>> +        if self._level < amount:
>> +            raise ArithmeticError(
>> +                f"Can't remove {amount:d} spaces from {self!r}")
> 
> Raise a fancy error when there's an actual need for it.  You're not
> coding a framework thousands of people you never heard of will put to
> uses you cannot imagine.
> 

It's not fancy, it's just a normal built-in exception, like 
AssertionError or any other:

Traceback (most recent call last):
   File "<stdin>", line 1, in <module>
ArithmeticError: Can't remove 4 spaces from Indent(0)

vs

Traceback (most recent call last):
   File "<stdin>", line 1, in <module>
AssertionError: Can't remove 4 spaces from Indent(0)


I feel like it's kind of a lateral move? I realize you feel this is an 
overfancy class doing what we hope is a temporary job, and that I have 
overengineered the hell out of a tiny do-nothing class... but I suppose 
that's also why I feel weird changing it around so much to accomplish so 
little.

Differences in style, I suppose.

Feel free to change it around to suit your tastes, I don't think it's 
worth spending a lot of ping-pong time on this paintsink in particular.

--js

>> +        self._level -= amount
>> +
>> +
>> +indent = Indentation()
>>   
>>   
>>   # Generate @code with @kwds interpolated.
>> -# Obey indent_level, and strip eatspace.
>> +# Obey indent, and strip eatspace.
>>   def cgen(code, **kwds):
>>       raw = code % kwds
>> -    if indent_level:
>> -        indent = genindent(indent_level)
>> -        raw = re.sub(r'^(?!(#|$))', indent, raw, flags=re.MULTILINE)
>> +    if indent:
>> +        raw = re.sub(r'^(?!(#|$))', str(indent), raw, flags=re.MULTILINE)
>>       return re.sub(re.escape(eatspace) + r' *', '', raw)
>>   
>>   
>> diff --git a/scripts/qapi/visit.py b/scripts/qapi/visit.py
>> index 808410d6f1b..14f30c228b7 100644
>> --- a/scripts/qapi/visit.py
>> +++ b/scripts/qapi/visit.py
>> @@ -18,9 +18,8 @@
>>       c_name,
>>       gen_endif,
>>       gen_if,
>> +    indent,
>>       mcgen,
>> -    pop_indent,
>> -    push_indent,
>>   )
>>   from .gen import QAPISchemaModularCVisitor, ifcontext
>>   from .schema import QAPISchemaObjectType
>> @@ -69,7 +68,7 @@ def gen_visit_object_members(name, base, members, variants):
>>       if (visit_optional(v, "%(name)s", &obj->has_%(c_name)s)) {
>>   ''',
>>                            name=memb.name, c_name=c_name(memb.name))
>> -            push_indent()
>> +            indent.increase()
>>           ret += mcgen('''
>>       if (!visit_type_%(c_type)s(v, "%(name)s", &obj->%(c_name)s, errp)) {
>>           return false;
>> @@ -78,7 +77,7 @@ def gen_visit_object_members(name, base, members, variants):
>>                        c_type=memb.type.c_name(), name=memb.name,
>>                        c_name=c_name(memb.name))
>>           if memb.optional:
>> -            pop_indent()
>> +            indent.decrease()
>>               ret += mcgen('''
>>       }
>>   ''')



  reply	other threads:[~2020-10-07 18:09 UTC|newest]

Thread overview: 125+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-10-05 19:51 [PATCH v5 00/36] qapi: static typing conversion, pt1 John Snow
2020-10-05 19:51 ` [PATCH v5 01/36] docs: repair broken references John Snow
2020-10-05 19:51 ` [PATCH v5 02/36] qapi: modify docstrings to be sphinx-compatible John Snow
2020-10-06 11:21   ` Markus Armbruster
2020-10-06 15:23     ` John Snow
2020-10-07  7:24       ` Markus Armbruster
2020-10-07 17:00         ` John Snow
2020-10-05 19:51 ` [PATCH v5 03/36] qapi-gen: Separate arg-parsing from generation John Snow
2020-10-06 11:51   ` Markus Armbruster
2020-10-06 15:59     ` John Snow
2020-10-07  7:54       ` Markus Armbruster
2020-10-07 14:52         ` John Snow
2020-10-08  5:56           ` Markus Armbruster
2020-10-08 17:33             ` John Snow
2020-10-06 16:46     ` John Snow
2020-10-07  8:07   ` Markus Armbruster
2020-10-07 14:36     ` John Snow
2020-10-08  6:51       ` Markus Armbruster
2020-10-08 16:37         ` John Snow
2020-10-08 16:50         ` John Snow
2020-10-09  7:12           ` Markus Armbruster
2020-10-07  8:12   ` Markus Armbruster
2020-10-07 14:41     ` John Snow
2020-10-08  7:15       ` Markus Armbruster
2020-10-08 17:14         ` John Snow
2020-10-09  7:19           ` Markus Armbruster
2020-10-05 19:51 ` [PATCH v5 04/36] qapi: move generator entrypoint into module John Snow
2020-10-05 19:51 ` [PATCH v5 05/36] qapi: Prefer explicit relative imports John Snow
2020-10-06 11:33   ` Philippe Mathieu-Daudé
2020-10-05 19:51 ` [PATCH v5 06/36] qapi: Remove wildcard includes John Snow
2020-10-06 11:34   ` Philippe Mathieu-Daudé
2020-10-05 19:51 ` [PATCH v5 07/36] qapi: enforce import order/styling with isort John Snow
2020-10-07  8:15   ` Markus Armbruster
2020-10-05 19:51 ` [PATCH v5 08/36] qapi: delint using flake8 John Snow
2020-10-07  8:19   ` Markus Armbruster
2020-10-07 14:54     ` John Snow
2020-10-05 19:51 ` [PATCH v5 09/36] qapi: add pylintrc John Snow
2020-10-05 19:51 ` [PATCH v5 10/36] qapi/common.py: Remove python compatibility workaround John Snow
2020-10-05 19:51 ` [PATCH v5 11/36] qapi/common.py: Add indent manager John Snow
2020-10-07  8:50   ` Markus Armbruster
2020-10-07 18:08     ` John Snow [this message]
2020-10-07 18:18       ` Eduardo Habkost
2020-10-08  7:23         ` Markus Armbruster
2020-10-08 17:45           ` John Snow
2020-10-05 19:51 ` [PATCH v5 12/36] qapi/common.py: delint with pylint John Snow
2020-10-05 19:51 ` [PATCH v5 13/36] qapi/common.py: Replace one-letter 'c' variable John Snow
2020-10-06 11:35   ` Philippe Mathieu-Daudé
2020-10-05 19:51 ` [PATCH v5 14/36] qapi/common.py: check with pylint John Snow
2020-10-05 19:51 ` [PATCH v5 15/36] qapi/common.py: add type hint annotations John Snow
2020-10-07  9:03   ` Markus Armbruster
2020-10-07 15:01     ` John Snow
2020-10-05 19:51 ` [PATCH v5 16/36] qapi/common.py: Convert comments into docstrings, and elaborate John Snow
2020-10-07  9:14   ` Markus Armbruster
2020-10-07 15:23     ` John Snow
2020-10-08  7:20       ` Markus Armbruster
2020-10-05 19:51 ` [PATCH v5 17/36] qapi/common.py: move build_params into gen.py John Snow
2020-10-07  9:21   ` Markus Armbruster
2020-10-07 15:26     ` John Snow
2020-10-07 18:10     ` Eduardo Habkost
2020-10-05 19:51 ` [PATCH v5 18/36] qapi: establish mypy type-checking baseline John Snow
2020-10-07  9:25   ` Markus Armbruster
2020-10-07 15:33     ` John Snow
2020-10-08  7:29       ` Markus Armbruster
2020-10-05 19:51 ` [PATCH v5 19/36] qapi/events.py: add type hint annotations John Snow
2020-10-07 11:32   ` Markus Armbruster
2020-10-07 11:49     ` Markus Armbruster
2020-10-07 15:46       ` John Snow
2020-10-08  9:16         ` Markus Armbruster
2020-10-08 16:19           ` John Snow
2020-10-09  7:21             ` Markus Armbruster
2020-10-07 15:39     ` John Snow
2020-10-08  7:41       ` Markus Armbruster
2020-10-08 15:35         ` John Snow
2020-10-05 19:51 ` [PATCH v5 20/36] qapi/events.py: Move comments into docstrings John Snow
2020-10-05 19:51 ` [PATCH v5 21/36] qapi/commands.py: Don't re-bind to variable of different type John Snow
2020-10-07 11:34   ` Markus Armbruster
2020-10-05 19:51 ` [PATCH v5 22/36] qapi/commands.py: add type hint annotations John Snow
2020-10-05 19:51 ` [PATCH v5 23/36] qapi/commands.py: enable checking with mypy John Snow
2020-10-07 11:37   ` Markus Armbruster
2020-10-07 15:49     ` John Snow
2020-10-08  7:52       ` Markus Armbruster
2020-10-05 19:51 ` [PATCH v5 24/36] qapi/source.py: add type hint annotations John Snow
2020-10-07 11:55   ` Markus Armbruster
2020-10-07 16:04     ` John Snow
2020-10-08  8:42       ` Markus Armbruster
2020-10-09 14:30         ` John Snow
2020-10-09 14:37           ` John Snow
2020-10-05 19:51 ` [PATCH v5 25/36] qapi/source.py: delint with pylint John Snow
2020-10-05 19:51 ` [PATCH v5 26/36] qapi/gen.py: Fix edge-case of _is_user_module John Snow
2020-10-06 11:44   ` Philippe Mathieu-Daudé
2020-10-07 12:02   ` Markus Armbruster
2020-10-07 16:09     ` John Snow
2020-10-05 19:51 ` [PATCH v5 27/36] qapi/gen.py: add type hint annotations John Snow
2020-10-07 12:21   ` Markus Armbruster
2020-10-07 16:21     ` John Snow
2020-10-07 13:20   ` Markus Armbruster
2020-10-07 16:50     ` John Snow
2020-10-08  8:44       ` Markus Armbruster
2020-10-05 19:51 ` [PATCH v5 28/36] qapi/gen.py: Enable checking with mypy John Snow
2020-10-05 19:51 ` [PATCH v5 29/36] qapi/gen.py: Remove unused parameter John Snow
2020-10-07 12:22   ` Markus Armbruster
2020-10-07 16:23     ` John Snow
2020-10-05 19:51 ` [PATCH v5 30/36] qapi/gen.py: update write() to be more idiomatic John Snow
2020-10-07 12:32   ` Markus Armbruster
2020-10-07 16:25     ` John Snow
2020-10-05 19:51 ` [PATCH v5 31/36] qapi/gen.py: delint with pylint John Snow
2020-10-05 19:51 ` [PATCH v5 32/36] qapi/types.py: add type hint annotations John Snow
2020-10-05 19:51 ` [PATCH v5 33/36] qapi/types.py: remove one-letter variables John Snow
2020-10-07 12:42   ` Markus Armbruster
2020-10-07 16:31     ` John Snow
2020-10-05 19:51 ` [PATCH v5 34/36] qapi/visit.py: assert tag_member contains a QAPISchemaEnumType John Snow
2020-10-07 12:43   ` Markus Armbruster
2020-10-07 16:40     ` John Snow
2020-10-08  9:06       ` Markus Armbruster
2020-10-08 15:49         ` John Snow
2020-10-09  7:24           ` Markus Armbruster
2020-10-05 19:51 ` [PATCH v5 35/36] qapi/visit.py: remove unused parameters from gen_visit_object John Snow
2020-10-06 11:43   ` Philippe Mathieu-Daudé
2020-10-05 19:51 ` [PATCH v5 36/36] qapi/visit.py: add type hint annotations John Snow
2020-10-07 13:00   ` Markus Armbruster
2020-10-07 16:43     ` John Snow
2020-10-05 23:05 ` [PATCH v5 00/36] qapi: static typing conversion, pt1 Cleber Rosa
2020-10-05 23:57   ` John Snow
2020-10-06 17:51     ` Cleber Rosa
2020-10-07 13:05 ` 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=0a507019-1e59-2967-9c7a-fa3712a74ae0@redhat.com \
    --to=jsnow@redhat.com \
    --cc=armbru@redhat.com \
    --cc=crosa@redhat.com \
    --cc=ehabkost@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).