* [PATCH] [gdb/build] Fix build breaker with -std=c++11
@ 2023-08-06 23:59 Tom de Vries
2023-08-07 14:35 ` Simon Marchi
0 siblings, 1 reply; 10+ messages in thread
From: Tom de Vries @ 2023-08-06 23:59 UTC (permalink / raw)
To: gdb-patches; +Cc: Tom Tromey
When building with -std=c++11 I run into:
...
gdb/dwarf2/cooked-index.c: In member function \
‘void cooked_index::start_writing_index(dwarf2_per_bfd*)’:
gdb/dwarf2/cooked-index.c:469:10: error: lambda capture initializers only \
available with -std=c++14 or -std=gnu++14 [-Werror]
ctx = std::move (ctx)] ()
^~~
...
Fix this by capturing a copy instead:
...
= gdb::thread_pool::g_thread_pool->post_task ([this, per_bfd, ctx] ()
...
Tested on x86_64-linux.
Reported-By: Tom Tromey <tom@tromey.com>
---
gdb/dwarf2/cooked-index.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/gdb/dwarf2/cooked-index.c b/gdb/dwarf2/cooked-index.c
index df82e86b690..87aac5aaa8b 100644
--- a/gdb/dwarf2/cooked-index.c
+++ b/gdb/dwarf2/cooked-index.c
@@ -465,8 +465,7 @@ cooked_index::start_writing_index (dwarf2_per_bfd *per_bfd)
/* This must be set after all the finalization tasks have been
started, because it may call 'wait'. */
m_write_future
- = gdb::thread_pool::g_thread_pool->post_task ([this, per_bfd,
- ctx = std::move (ctx)] ()
+ = gdb::thread_pool::g_thread_pool->post_task ([this, per_bfd, ctx] ()
{
maybe_write_index (per_bfd, ctx);
});
base-commit: 5e66f55c62e306afbcc93856bf06e542ddd00997
--
2.35.3
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH] [gdb/build] Fix build breaker with -std=c++11
2023-08-06 23:59 [PATCH] [gdb/build] Fix build breaker with -std=c++11 Tom de Vries
@ 2023-08-07 14:35 ` Simon Marchi
2023-08-07 17:15 ` Tom de Vries
2023-08-07 22:16 ` Tom Tromey
0 siblings, 2 replies; 10+ messages in thread
From: Simon Marchi @ 2023-08-07 14:35 UTC (permalink / raw)
To: Tom de Vries, Tom de Vries via Gdb-patches, gdb-patches; +Cc: Tom Tromey
On August 6, 2023 7:59:15 p.m. EDT, Tom de Vries via Gdb-patches <gdb-patches@sourceware.org> wrote:
>When building with -std=c++11 I run into:
>...
>gdb/dwarf2/cooked-index.c: In member function \
> ‘void cooked_index::start_writing_index(dwarf2_per_bfd*)’:
>gdb/dwarf2/cooked-index.c:469:10: error: lambda capture initializers only \
> available with -std=c++14 or -std=gnu++14 [-Werror]
> ctx = std::move (ctx)] ()
> ^~~
>...
>
>Fix this by capturing a copy instead:
>...
> = gdb::thread_pool::g_thread_pool->post_task ([this, per_bfd, ctx] ()
>...
>
>Tested on x86_64-linux.
>
>Reported-By: Tom Tromey <tom@tromey.com>
When we encounter things like this, should we add some greppable TODO comment to record an actionable item for when we switch to a more recent C++ version?
Simon
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH] [gdb/build] Fix build breaker with -std=c++11
2023-08-07 14:35 ` Simon Marchi
@ 2023-08-07 17:15 ` Tom de Vries
2023-08-07 18:01 ` John Baldwin
2023-08-07 22:16 ` Tom Tromey
1 sibling, 1 reply; 10+ messages in thread
From: Tom de Vries @ 2023-08-07 17:15 UTC (permalink / raw)
To: Simon Marchi, Tom de Vries via Gdb-patches; +Cc: Tom Tromey
[-- Attachment #1: Type: text/plain, Size: 966 bytes --]
On 8/7/23 16:35, Simon Marchi wrote:
>
>
> On August 6, 2023 7:59:15 p.m. EDT, Tom de Vries via Gdb-patches <gdb-patches@sourceware.org> wrote:
>> When building with -std=c++11 I run into:
>> ...
>> gdb/dwarf2/cooked-index.c: In member function \
>> ‘void cooked_index::start_writing_index(dwarf2_per_bfd*)’:
>> gdb/dwarf2/cooked-index.c:469:10: error: lambda capture initializers only \
>> available with -std=c++14 or -std=gnu++14 [-Werror]
>> ctx = std::move (ctx)] ()
>> ^~~
>> ...
>>
>> Fix this by capturing a copy instead:
>> ...
>> = gdb::thread_pool::g_thread_pool->post_task ([this, per_bfd, ctx] ()
>> ...
>>
>> Tested on x86_64-linux.
>>
>> Reported-By: Tom Tromey <tom@tromey.com>
>
> When we encounter things like this, should we add some greppable TODO comment to record an actionable item for when we switch to a more recent C++ version?
Hi Simon,
thanks for the review.
Fine by me, added comment.
Thanks,
- Tom
[-- Attachment #2: 0001-gdb-build-Fix-build-breaker-with-std-c-11.patch --]
[-- Type: text/x-patch, Size: 1845 bytes --]
From ef0b55aaf4eaa9a1efece7656df1a8a997f97412 Mon Sep 17 00:00:00 2001
From: Tom de Vries <tdevries@suse.de>
Date: Mon, 7 Aug 2023 00:31:04 +0200
Subject: [PATCH] [gdb/build] Fix build breaker with -std=c++11
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
When building with -std=c++11 I run into:
...
gdb/dwarf2/cooked-index.c: In member function \
‘void cooked_index::start_writing_index(dwarf2_per_bfd*)’:
gdb/dwarf2/cooked-index.c:469:10: error: lambda capture initializers only \
available with -std=c++14 or -std=gnu++14 [-Werror]
ctx = std::move (ctx)] ()
^~~
...
Fix this by capturing a copy instead:
...
= gdb::thread_pool::g_thread_pool->post_task ([this, per_bfd, ctx] ()
...
Tested on x86_64-linux.
Reported-By: Tom Tromey <tom@tromey.com>
---
gdb/dwarf2/cooked-index.c | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/gdb/dwarf2/cooked-index.c b/gdb/dwarf2/cooked-index.c
index df82e86b690..775b1383ce5 100644
--- a/gdb/dwarf2/cooked-index.c
+++ b/gdb/dwarf2/cooked-index.c
@@ -463,10 +463,11 @@ cooked_index::start_writing_index (dwarf2_per_bfd *per_bfd)
index_cache_store_context ctx (global_index_cache, per_bfd);
/* This must be set after all the finalization tasks have been
- started, because it may call 'wait'. */
+ started, because it may call 'wait'.
+ TODO (c++14): The capture of ctx here can be rewritten to
+ "ctx = std::move (ctx)" once the standard is updated to c++14. */
m_write_future
- = gdb::thread_pool::g_thread_pool->post_task ([this, per_bfd,
- ctx = std::move (ctx)] ()
+ = gdb::thread_pool::g_thread_pool->post_task ([this, per_bfd, ctx] ()
{
maybe_write_index (per_bfd, ctx);
});
base-commit: 8d27b09d0877982dc2252c01a3f1d9c631b2d6eb
--
2.35.3
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH] [gdb/build] Fix build breaker with -std=c++11
2023-08-07 17:15 ` Tom de Vries
@ 2023-08-07 18:01 ` John Baldwin
2023-08-07 23:11 ` Tom de Vries
0 siblings, 1 reply; 10+ messages in thread
From: John Baldwin @ 2023-08-07 18:01 UTC (permalink / raw)
To: Tom de Vries, Simon Marchi, Tom de Vries via Gdb-patches; +Cc: Tom Tromey
On 8/7/23 10:15 AM, Tom de Vries via Gdb-patches wrote:
> On 8/7/23 16:35, Simon Marchi wrote:
>>
>>
>> On August 6, 2023 7:59:15 p.m. EDT, Tom de Vries via Gdb-patches <gdb-patches@sourceware.org> wrote:
>>> When building with -std=c++11 I run into:
>>> ...
>>> gdb/dwarf2/cooked-index.c: In member function \
>>> ‘void cooked_index::start_writing_index(dwarf2_per_bfd*)’:
>>> gdb/dwarf2/cooked-index.c:469:10: error: lambda capture initializers only \
>>> available with -std=c++14 or -std=gnu++14 [-Werror]
>>> ctx = std::move (ctx)] ()
>>> ^~~
>>> ...
>>>
>>> Fix this by capturing a copy instead:
>>> ...
>>> = gdb::thread_pool::g_thread_pool->post_task ([this, per_bfd, ctx] ()
>>> ...
>>>
>>> Tested on x86_64-linux.
>>>
>>> Reported-By: Tom Tromey <tom@tromey.com>
>>
>> When we encounter things like this, should we add some greppable TODO comment to record an actionable item for when we switch to a more recent C++ version?
>
> Hi Simon,
>
> thanks for the review.
>
> Fine by me, added comment.
An alternative that is slightly more work but might be nicer long term would be
to #if on the C++ version, so for example:
#if __cplusplus >= 201402L
= gdb::thread_pool::g_thread_pool->post_task ([this, per_bfd,
ctx = std::move (ctx)] ()
#else
= gdb::thread_pool::g_thread_pool->post_task ([this, per_bfd, ctx]) ()
#endif
--
John Baldwin
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH] [gdb/build] Fix build breaker with -std=c++11
2023-08-07 14:35 ` Simon Marchi
2023-08-07 17:15 ` Tom de Vries
@ 2023-08-07 22:16 ` Tom Tromey
2023-08-08 0:30 ` Tom de Vries
2023-08-08 15:21 ` Simon Marchi
1 sibling, 2 replies; 10+ messages in thread
From: Tom Tromey @ 2023-08-07 22:16 UTC (permalink / raw)
To: Simon Marchi; +Cc: Tom de Vries, Tom de Vries via Gdb-patches, Tom Tromey
Simon> When we encounter things like this, should we add some greppable
Simon> TODO comment to record an actionable item for when we switch to a
Simon> more recent C++ version?
FWIW I know of one other instance of this, in gdb_demangle:
std::string copy = name;
run_on_main_thread ([=] ()
{
report_failed_demangle (copy.c_str (), core_dump_allowed,
Here, the name is copied twice, but with move captures there could be
just a single copy.
I wonder when we can switch to C++14.
Tom
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH] [gdb/build] Fix build breaker with -std=c++11
2023-08-07 18:01 ` John Baldwin
@ 2023-08-07 23:11 ` Tom de Vries
2023-08-08 5:37 ` John Baldwin
0 siblings, 1 reply; 10+ messages in thread
From: Tom de Vries @ 2023-08-07 23:11 UTC (permalink / raw)
To: John Baldwin, Simon Marchi, Tom de Vries via Gdb-patches; +Cc: Tom Tromey
[-- Attachment #1: Type: text/plain, Size: 1690 bytes --]
On 8/7/23 20:01, John Baldwin wrote:
> On 8/7/23 10:15 AM, Tom de Vries via Gdb-patches wrote:
>> On 8/7/23 16:35, Simon Marchi wrote:
>>>
>>>
>>> On August 6, 2023 7:59:15 p.m. EDT, Tom de Vries via Gdb-patches
>>> <gdb-patches@sourceware.org> wrote:
>>>> When building with -std=c++11 I run into:
>>>> ...
>>>> gdb/dwarf2/cooked-index.c: In member function \
>>>> ‘void cooked_index::start_writing_index(dwarf2_per_bfd*)’:
>>>> gdb/dwarf2/cooked-index.c:469:10: error: lambda capture initializers
>>>> only \
>>>> available with -std=c++14 or -std=gnu++14 [-Werror]
>>>> ctx = std::move (ctx)] ()
>>>> ^~~
>>>> ...
>>>>
>>>> Fix this by capturing a copy instead:
>>>> ...
>>>> = gdb::thread_pool::g_thread_pool->post_task ([this, per_bfd,
>>>> ctx] ()
>>>> ...
>>>>
>>>> Tested on x86_64-linux.
>>>>
>>>> Reported-By: Tom Tromey <tom@tromey.com>
>>>
>>> When we encounter things like this, should we add some greppable TODO
>>> comment to record an actionable item for when we switch to a more
>>> recent C++ version?
>>
>> Hi Simon,
>>
>> thanks for the review.
>>
>> Fine by me, added comment.
>
> An alternative that is slightly more work but might be nicer long term
> would be
> to #if on the C++ version, so for example:
>
> #if __cplusplus >= 201402L
> = gdb::thread_pool::g_thread_pool->post_task ([this, per_bfd,
> ctx = std::move (ctx)] ()
> #else
> = gdb::thread_pool::g_thread_pool->post_task ([this, per_bfd, ctx]) ()
> #endif
>
Thanks for the suggestion. Updated patch attached.
Thanks,
- Tom
[-- Attachment #2: 0001-gdb-build-Fix-build-breaker-with-std-c-11.patch --]
[-- Type: text/x-patch, Size: 1622 bytes --]
From ddbe8c3c8a8665e77f9ebb108c5627a6ebecb81f Mon Sep 17 00:00:00 2001
From: Tom de Vries <tdevries@suse.de>
Date: Mon, 7 Aug 2023 00:31:04 +0200
Subject: [PATCH] [gdb/build] Fix build breaker with -std=c++11
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
When building with -std=c++11 I run into:
...
gdb/dwarf2/cooked-index.c: In member function \
‘void cooked_index::start_writing_index(dwarf2_per_bfd*)’:
gdb/dwarf2/cooked-index.c:469:10: error: lambda capture initializers only \
available with -std=c++14 or -std=gnu++14 [-Werror]
ctx = std::move (ctx)] ()
^~~
...
Fix this by capturing a copy instead when using -std=c++11:
...
= gdb::thread_pool::g_thread_pool->post_task ([this, per_bfd, ctx] ()
...
Tested by building with and without -stdc=++11 on x86_64-linux.
Reported-By: Tom Tromey <tom@tromey.com>
---
gdb/dwarf2/cooked-index.c | 7 ++++++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/gdb/dwarf2/cooked-index.c b/gdb/dwarf2/cooked-index.c
index df82e86b690..92ef0e35c5e 100644
--- a/gdb/dwarf2/cooked-index.c
+++ b/gdb/dwarf2/cooked-index.c
@@ -466,7 +466,12 @@ cooked_index::start_writing_index (dwarf2_per_bfd *per_bfd)
started, because it may call 'wait'. */
m_write_future
= gdb::thread_pool::g_thread_pool->post_task ([this, per_bfd,
- ctx = std::move (ctx)] ()
+#if __cplusplus >= 201402L
+ ctx = std::move (ctx)
+#else
+ ctx
+#endif
+ ] ()
{
maybe_write_index (per_bfd, ctx);
});
base-commit: 8d27b09d0877982dc2252c01a3f1d9c631b2d6eb
--
2.35.3
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH] [gdb/build] Fix build breaker with -std=c++11
2023-08-07 22:16 ` Tom Tromey
@ 2023-08-08 0:30 ` Tom de Vries
2023-08-08 22:19 ` Tom Tromey
2023-08-08 15:21 ` Simon Marchi
1 sibling, 1 reply; 10+ messages in thread
From: Tom de Vries @ 2023-08-08 0:30 UTC (permalink / raw)
To: Tom Tromey, Simon Marchi; +Cc: Tom de Vries via Gdb-patches
[-- Attachment #1: Type: text/plain, Size: 548 bytes --]
On 8/8/23 00:16, Tom Tromey wrote:
> Simon> When we encounter things like this, should we add some greppable
> Simon> TODO comment to record an actionable item for when we switch to a
> Simon> more recent C++ version?
>
> FWIW I know of one other instance of this, in gdb_demangle:
>
> std::string copy = name;
> run_on_main_thread ([=] ()
> {
> report_failed_demangle (copy.c_str (), core_dump_allowed,
>
> Here, the name is copied twice, but with move captures there could be
> just a single copy.
Like so?
Thanks,
- Tom
[-- Attachment #2: 0001-gdb-build-Use-move-capture-in-gdb_demangle.patch --]
[-- Type: text/x-patch, Size: 1141 bytes --]
From 7a6f3e7e39701b2b0e2f3b8f761e7c32f46af4a6 Mon Sep 17 00:00:00 2001
From: Tom de Vries <tdevries@suse.de>
Date: Tue, 8 Aug 2023 01:34:45 +0200
Subject: [PATCH] [gdb/build] Use move capture in gdb_demangle
Use move capture in gdb_demangle when compiling for c++14 or higher, to save a
std::string copy.
Tested on x86_64-linux.
Reported-By: Tom Tromey <tom@tromey.com>
---
gdb/cp-support.c | 8 +++++++-
1 file changed, 7 insertions(+), 1 deletion(-)
diff --git a/gdb/cp-support.c b/gdb/cp-support.c
index 2fb9d9ef0b2..0300727434d 100644
--- a/gdb/cp-support.c
+++ b/gdb/cp-support.c
@@ -1656,7 +1656,13 @@ gdb_demangle (const char *name, int options)
we might be in a background thread. Instead, arrange for
the reporting to happen on the main thread. */
std::string copy = name;
- run_on_main_thread ([=] ()
+ run_on_main_thread ([
+#if __cplusplus >= 201402L
+ =, copy = std::move (copy)
+#else
+ =
+#endif
+ ] ()
{
report_failed_demangle (copy.c_str (), core_dump_allowed,
crash_signal);
base-commit: ddbe8c3c8a8665e77f9ebb108c5627a6ebecb81f
--
2.35.3
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH] [gdb/build] Fix build breaker with -std=c++11
2023-08-07 23:11 ` Tom de Vries
@ 2023-08-08 5:37 ` John Baldwin
0 siblings, 0 replies; 10+ messages in thread
From: John Baldwin @ 2023-08-08 5:37 UTC (permalink / raw)
To: Tom de Vries, Simon Marchi, Tom de Vries via Gdb-patches; +Cc: Tom Tromey
On 8/7/23 4:11 PM, Tom de Vries wrote:
> On 8/7/23 20:01, John Baldwin wrote:
>> On 8/7/23 10:15 AM, Tom de Vries via Gdb-patches wrote:
>>> On 8/7/23 16:35, Simon Marchi wrote:
>>>>
>>>>
>>>> On August 6, 2023 7:59:15 p.m. EDT, Tom de Vries via Gdb-patches
>>>> <gdb-patches@sourceware.org> wrote:
>>>>> When building with -std=c++11 I run into:
>>>>> ...
>>>>> gdb/dwarf2/cooked-index.c: In member function \
>>>>> ‘void cooked_index::start_writing_index(dwarf2_per_bfd*)’:
>>>>> gdb/dwarf2/cooked-index.c:469:10: error: lambda capture initializers
>>>>> only \
>>>>> available with -std=c++14 or -std=gnu++14 [-Werror]
>>>>> ctx = std::move (ctx)] ()
>>>>> ^~~
>>>>> ...
>>>>>
>>>>> Fix this by capturing a copy instead:
>>>>> ...
>>>>> = gdb::thread_pool::g_thread_pool->post_task ([this, per_bfd,
>>>>> ctx] ()
>>>>> ...
>>>>>
>>>>> Tested on x86_64-linux.
>>>>>
>>>>> Reported-By: Tom Tromey <tom@tromey.com>
>>>>
>>>> When we encounter things like this, should we add some greppable TODO
>>>> comment to record an actionable item for when we switch to a more
>>>> recent C++ version?
>>>
>>> Hi Simon,
>>>
>>> thanks for the review.
>>>
>>> Fine by me, added comment.
>>
>> An alternative that is slightly more work but might be nicer long term
>> would be
>> to #if on the C++ version, so for example:
>>
>> #if __cplusplus >= 201402L
>> = gdb::thread_pool::g_thread_pool->post_task ([this, per_bfd,
>> ctx = std::move (ctx)] ()
>> #else
>> = gdb::thread_pool::g_thread_pool->post_task ([this, per_bfd, ctx]) ()
>> #endif
>>
>
> Thanks for the suggestion. Updated patch attached.
LGTM
Reviewed-by: John Baldwin <jhb@FreeBSD.org>
--
John Baldwin
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH] [gdb/build] Fix build breaker with -std=c++11
2023-08-07 22:16 ` Tom Tromey
2023-08-08 0:30 ` Tom de Vries
@ 2023-08-08 15:21 ` Simon Marchi
1 sibling, 0 replies; 10+ messages in thread
From: Simon Marchi @ 2023-08-08 15:21 UTC (permalink / raw)
To: Tom Tromey; +Cc: Tom de Vries, Tom de Vries via Gdb-patches
On August 7, 2023 6:16:19 p.m. EDT, Tom Tromey <tom@tromey.com> wrote:
>Simon> When we encounter things like this, should we add some greppable
>Simon> TODO comment to record an actionable item for when we switch to a
>Simon> more recent C++ version?
>
>FWIW I know of one other instance of this, in gdb_demangle:
>
> std::string copy = name;
> run_on_main_thread ([=] ()
> {
> report_failed_demangle (copy.c_str (), core_dump_allowed,
>
>Here, the name is copied twice, but with move captures there could be
>just a single copy.
>
>I wonder when we can switch to C++14.
If I recall, Tom de Vries said they have to support some SLES version that has gcc 4.x as the system compiler for a while still. Other than that I am not aware of anything holding back.
Simon
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH] [gdb/build] Fix build breaker with -std=c++11
2023-08-08 0:30 ` Tom de Vries
@ 2023-08-08 22:19 ` Tom Tromey
0 siblings, 0 replies; 10+ messages in thread
From: Tom Tromey @ 2023-08-08 22:19 UTC (permalink / raw)
To: Tom de Vries; +Cc: Tom Tromey, Simon Marchi, Tom de Vries via Gdb-patches
>> Here, the name is copied twice, but with move captures there could be
>> just a single copy.
Tom> Like so?
Yes, thank you.
Approved-By: Tom Tromey <tom@tromey.com>
Tom
^ permalink raw reply [flat|nested] 10+ messages in thread
end of thread, other threads:[~2023-08-08 22:19 UTC | newest]
Thread overview: 10+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-08-06 23:59 [PATCH] [gdb/build] Fix build breaker with -std=c++11 Tom de Vries
2023-08-07 14:35 ` Simon Marchi
2023-08-07 17:15 ` Tom de Vries
2023-08-07 18:01 ` John Baldwin
2023-08-07 23:11 ` Tom de Vries
2023-08-08 5:37 ` John Baldwin
2023-08-07 22:16 ` Tom Tromey
2023-08-08 0:30 ` Tom de Vries
2023-08-08 22:19 ` Tom Tromey
2023-08-08 15:21 ` Simon Marchi
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).