From: Jason Merrill <jason@redhat.com>
To: Bernhard Reutner-Fischer <rep.dot.nop@gmail.com>
Cc: gcc-patches@gcc.gnu.org,
Bernhard Reutner-Fischer <aldot@gcc.gnu.org>,
Nathan Sidwell <nathan@acm.org>
Subject: Re: [PATCH 2/5] c++: Set the locus of the function result decl
Date: Fri, 18 Nov 2022 11:06:29 -0500 [thread overview]
Message-ID: <3e8757c8-cb16-9446-6953-5d6de9deeb5b@redhat.com> (raw)
In-Reply-To: <20221118114938.48a39cbd@nbbrfq>
On 11/18/22 05:49, Bernhard Reutner-Fischer wrote:
> On Thu, 17 Nov 2022 18:52:36 -0500
> Jason Merrill <jason@redhat.com> wrote:
>
>> On 11/17/22 14:02, Bernhard Reutner-Fischer wrote:
>>> On Thu, 17 Nov 2022 09:53:32 -0500
>>> Jason Merrill <jason@redhat.com> wrote:
>
>>>> Instead, you want to copy the location for instantiations, i.e. check
>>>> DECL_TEMPLATE_INSTANTIATION instead of !DECL_USE_TEMPLATE.
>>>
>>> No, that makes no difference.
>>
>> Hmm, when I stop there when processing the instantiation the template's
>> DECL_RESULT has the right location information, e.g. for
>>
>> template <class T> int f() { return 42; }
>>
>> int main()
>> {
>> f<int>();
>> }
>>
>> #1 0x0000000000f950e8 in instantiate_body (pattern=<template_decl
>> 0x7ffff7ff5080 f>, args=<tree_vec 0x7fffe9712ae0>, d=<function_decl
>> 0x7fffe971e600 f>, nested_p=false) at /home/jason/gt/gcc/cp/pt.cc:26470
>> #0 start_preparsed_function (decl1=<function_decl 0x7fffe971e600 f>,
>> attrs=<tree 0x0>, flags=1) at /home/jason/gt/gcc/cp/decl.cc:17252
>> (gdb) p expand_location (input_location)
>> $13 = {file = 0x4962370 "wa.C", line = 1, column = 24, data = 0x0, sysp
>> = false}
>> (gdb) p expand_location (DECL_SOURCE_LOCATION (DECL_RESULT
>> (DECL_TEMPLATE_RESULT (DECL_TI_TEMPLATE (decl1)))))
>> $14 = {file = 0x4962370 "wa.C", line = 1, column = 20, data = 0x0, sysp
>> = false}
>
> Yes, that works. Sorry if i was not clear: The thing in the cover
> letter in this series does not work, the mini_vector reduced testcase
> from the libstdc++-v3/include/ext/bitmap_allocator.h.
> class template member function return type location, would that be it?
>
> AFAIR the problem was that that these member functions get their result
> decl late. When they get them, there are no
> declspecs->locations[ds_type_spec] around anywhere to tuck that on the
> resdecl. While the result decl is clear, there is no obvious way where
> to store the ds_type_spec (somewhere in the template, as you told me).
>
> Back then I tried moving the resdecl building from
> start_preparsed_function to grokfndecl but that did not work out easily
> IIRC and i ultimately gave up to move stuff around rather blindly.
> I also tried to find a spot where i could store the ds_type_spec locus
> somewhere in grokmethod, but i think the problem was the same, i had
> just the type where i cannot store a locus and did not find a place
> where i could smuggle the locus along.
Ah, so the problem is deferred parsing of methods, rather than
templates. Building the DECL_RESULT sooner does seem like the right
approach to handling that, whether that's in grokfndecl or grokmethod.
> So, to make that clear. Your template function (?) works:
>
> $ XXX=1 ./xg++ -B. -S -o /dev/null ../tmp4/return-narrow-2j.cc
> ../tmp4/return-narrow-2j.cc: In function ‘int f()’:
> ../tmp4/return-narrow-2j.cc:1:20: warning: result decl locus sample
> 1 | template <class T> int f() { return 42; }
> | ^~~
> | the return type
> ../tmp4/return-narrow-2j.cc: In function ‘int main()’:
> ../tmp4/return-narrow-2j.cc:3:1: warning: result decl locus sample
> 3 | int main()
> | ^~~
> | the return type
> ../tmp4/return-narrow-2j.cc: In instantiation of ‘int f() [with T = int]’:
> ../tmp4/return-narrow-2j.cc:5:10: required from here
> ../tmp4/return-narrow-2j.cc:1:20: warning: result decl locus sample
> 1 | template <class T> int f() { return 42; }
> | ^~~
> | the return type
>
>
> The class member fn not so much (IMHO, see attached):
>
> $ XXX=1 ./xg++ -B. -S -o /dev/null ../tmp4/return-narrow-2.cc
> ../tmp4/return-narrow-2.cc: In member function ‘const long unsigned int __mini_vector< <template-parameter-1-1> >::_M_space_left()’:
> ../tmp4/return-narrow-2.cc:9:3: warning: result decl locus sample
> 9 | { return _M_finish != 0; }
> | ^
> | the return type
> ../tmp4/return-narrow-2.cc: In instantiation of ‘const long unsigned int __mini_vector< <template-parameter-1-1> >::_M_space_left() [with <template-parameter-1-1> = std::pair<long int, long int>]’:
> ../tmp4/return-narrow-2.cc:11:17: required from here
> ../tmp4/return-narrow-2.cc:9:3: warning: result decl locus sample
> 9 | { return _M_finish != 0; }
> | ^
> | the return type
> ../tmp4/return-narrow-2.cc: In instantiation of ‘const long unsigned int __mini_vector< <template-parameter-1-1> >::_M_space_left() [with <template-parameter-1-1> = int]’:
> ../tmp4/return-narrow-2.cc:12:17: required from here
> ../tmp4/return-narrow-2.cc:9:3: warning: result decl locus sample
> 9 | { return _M_finish != 0; }
> | ^
> | the return type
>
>
>>
>>> But really I'm not interested in the template case, i only mentioned
>>> them because they don't work and in case somebody wanted to have correct
>>> locations.
>>> I remember just frustration when i looked at those a year ago.
>>
>> I'd like to get the template case right while we're looking at it. I
>> guess I can add that myself if you're done trying.
>>
>>> Is the hunk for normal functions OK for trunk?
>>
>> You also need a testcase for the desired behavior, with e.g.
>> { dg-error "23:" }
>
> I'd have to think about how to test that with trunk, yes.
> There are no existing warnings that want to point to the return type,
> are there?
Good point. Do any of your later patches add such a warning?
> Maybe a g++.dg/plugin/result_decl_plugin.c then.
>
> set plugin_test_list [list
> hmz. That strikes me as not all that flexible.
> We could glob *_plugin.[cC][c]*, and have foo_plugin.lst contain it's
> files. Whatever.
>
> thanks,
next prev parent reply other threads:[~2022-11-18 16:06 UTC|newest]
Thread overview: 34+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-11-12 23:45 [PATCH 0/5] function result decl location; type demotion Bernhard Reutner-Fischer
2022-11-12 23:45 ` [PATCH 1/5] c: Set the locus of the function result decl Bernhard Reutner-Fischer
2022-11-14 21:25 ` Joseph Myers
2022-11-12 23:45 ` [PATCH 2/5] c++: " Bernhard Reutner-Fischer
2022-11-15 23:52 ` Jason Merrill
2022-11-17 8:56 ` Bernhard Reutner-Fischer
2022-11-17 14:53 ` Jason Merrill
2022-11-17 19:02 ` Bernhard Reutner-Fischer
2022-11-17 23:52 ` Jason Merrill
2022-11-18 10:49 ` Bernhard Reutner-Fischer
2022-11-18 16:06 ` Jason Merrill [this message]
2022-11-18 18:26 ` Bernhard Reutner-Fischer
2022-11-19 9:56 ` Bernhard Reutner-Fischer
2022-11-20 17:06 ` Bernhard Reutner-Fischer
2022-11-22 20:25 ` Jason Merrill
2022-11-23 15:28 ` Jason Merrill
2022-12-02 19:30 ` Jason Merrill
2022-12-02 19:55 ` Bernhard Reutner-Fischer
2022-11-12 23:45 ` [PATCH 3/5] Fortran: Narrow return types [PR78798] Bernhard Reutner-Fischer
2022-11-13 10:13 ` Janne Blomqvist
2022-11-13 10:39 ` Bernhard Reutner-Fischer
2022-11-13 20:29 ` Harald Anlauf
2022-11-13 20:29 ` Harald Anlauf
2022-11-13 21:50 ` Bernhard Reutner-Fischer
2023-05-10 16:47 ` [PATCH v2] " Bernhard Reutner-Fischer
2023-05-14 12:27 ` Mikael Morin
2023-05-14 13:04 ` Thomas Koenig
2023-05-14 15:24 ` Bernhard Reutner-Fischer
2023-05-14 18:06 ` Mikael Morin
2023-05-18 19:52 ` Bernhard Reutner-Fischer
2022-11-12 23:45 ` [PATCH 4/5] value-range: Add as_string diagnostics helper Bernhard Reutner-Fischer
2022-11-12 23:55 ` Andrew Pinski
2022-11-17 3:30 ` Jeff Law
2022-11-12 23:45 ` [PATCH 5/5] gimple: Add pass to note possible type demotions; IPA pro/demotion; DO NOT MERGE Bernhard Reutner-Fischer
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=3e8757c8-cb16-9446-6953-5d6de9deeb5b@redhat.com \
--to=jason@redhat.com \
--cc=aldot@gcc.gnu.org \
--cc=gcc-patches@gcc.gnu.org \
--cc=nathan@acm.org \
--cc=rep.dot.nop@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 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 read-only IMAP folder(s) and NNTP newsgroup(s).