Skip to contentSkip to conversation

vte-native: reduce c++ requirements from c++17 -> c++14

PATCH In review Andreas Müller 3 participants 8 messages Apr 29, 2019 – Apr 30, 2019

lore
Share & tools
lore
#0 lore raw

cover vte-native: reduce c++ requirements from c++17 -> c++14

Let the autobuilders decide :)

Andreas Müller (1):
vte-native: reduce c++ requirements from c++17 -> c++14

 ...uce-gcc-required-support-downto-c-14.patch | 34 ++++++++++++++
 ...001-Revert-parser-glue-Use-std-clamp.patch | 44 +++++++++++++++++++
 ...replace-std-clamp-by-std-min-std-max.patch | 33 ++++++++++++++
 meta/recipes-support/vte/vte_0.56.1.bb        |  7 +++
 4 files changed, 118 insertions(+)
 create mode 100644 meta/recipes-support/vte/vte/0001-Reduce-gcc-required-support-downto-c-14.patch
 create mode 100644 meta/recipes-support/vte/vte/0001-Revert-parser-glue-Use-std-clamp.patch
 create mode 100644 meta/recipes-support/vte/vte/0001-vte.cc-replace-std-clamp-by-std-min-std-max.patch
Signature · 1 line
2.20.1
#1 lore raw

PATCH 1/1 vte-native: reduce c++ requirements from c++17 -> c++14

This should fix build on ancient hosts

---
 ...uce-gcc-required-support-downto-c-14.patch | 34 ++++++++++++++
 ...001-Revert-parser-glue-Use-std-clamp.patch | 44 +++++++++++++++++++
 ...replace-std-clamp-by-std-min-std-max.patch | 33 ++++++++++++++
 meta/recipes-support/vte/vte_0.56.1.bb        |  7 +++
 4 files changed, 118 insertions(+)
 create mode 100644 meta/recipes-support/vte/vte/0001-Reduce-gcc-required-support-downto-c-14.patch
 create mode 100644 meta/recipes-support/vte/vte/0001-Revert-parser-glue-Use-std-clamp.patch
 create mode 100644 meta/recipes-support/vte/vte/0001-vte.cc-replace-std-clamp-by-std-min-std-max.patch
meta/recipes-support/vte/vte/0001-Reduce-gcc-required-support-downto-c-14.patch new file +34 -0
new file mode 100644
index 0000000000..19379b4c9b
@@ -0,0 +1,34 @@
+From 6d33f959f442f1338846f6f80e826f1345951b57 Mon Sep 17 00:00:00 2001
+From: =?UTF-8?q?Andreas=20M=C3=BCller?= <[email protected]>
+Date: Sat, 27 Apr 2019 12:52:03 +0200
+Subject: [PATCH] Reduce gcc required support downto c++14
+MIME-Version: 1.0
+Content-Type: text/plain; charset=UTF-8
+Content-Transfer-Encoding: 8bit
+
+Native builds require host gcc and they might not support full c++17 but enough
+to build vte-native.
+
+Upstream-Status: Inappropriate [configuration]
+
+Signed-off-by: Andreas Müller <[email protected]>
+---
+ configure.ac | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+diff --git a/configure.ac b/configure.ac
+index 1c7a753b..698e795d 100644
+--- a/configure.ac
++++ b/configure.ac
+@@ -97,7 +97,7 @@ AC_HEADER_STDC
+ AM_PROG_CC_C_O
+ AC_PROG_SED
+ 
+-AX_CXX_COMPILE_STDCXX([17])
++AX_CXX_COMPILE_STDCXX([14])
+ 
+ AC_SYS_LARGEFILE
+ 
+-- 
+2.20.1
+
meta/recipes-support/vte/vte/0001-Revert-parser-glue-Use-std-clamp.patch new file +44 -0

44 lines not on this page · raw message

meta/recipes-support/vte/vte/0001-vte.cc-replace-std-clamp-by-std-min-std-max.patch new file +33 -0
new file mode 100644
index 0000000000..ef53458858
@@ -0,0 +1,33 @@
+From db3b9cceb55130b0a4b3414be804530c368ececc Mon Sep 17 00:00:00 2001
+From: =?UTF-8?q?Andreas=20M=C3=BCller?= <[email protected]>
+Date: Sat, 27 Apr 2019 13:22:52 +0200
+Subject: [PATCH] vte.cc: replace std:clamp by std:min(std:max(..)..)
+MIME-Version: 1.0
+Content-Type: text/plain; charset=UTF-8
+Content-Transfer-Encoding: 8bit
+
+std::clamp requires c++17
+
+Upstream-Status: Inappropriate [configuration]
+
+Signed-off-by: Andreas Müller <[email protected]>
+---
+ src/vte.cc | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+diff --git a/src/vte.cc b/src/vte.cc
+index 0aa9a3fc..c9faab49 100644
+--- a/src/vte.cc
++++ b/src/vte.cc
+@@ -5125,7 +5125,7 @@ Terminal::resolve_selection_endpoint(vte::grid::halfcoords const& rowcolhalf, bo
+                 /* Just find the nearest cell boundary within the line, not caring about CJKs, unused
+                  * cells, or wrapping at EOL. The @after parameter is unused in this mode. */
+                 col += half;
+-                col = std::clamp (col, 0L, m_column_count);
++                col = std::min (std::max(col, 0L), m_column_count);
+         } else {
+                 switch (m_selection_type) {
+                 case selection_type_char:
+-- 
+2.20.1
+
meta/recipes-support/vte/vte_0.56.1.bb +7 -0
index 3585e9ce57..b261c03191 100644
@@ -19,6 +19,13 @@ SRC_URI += "file://0001-Don-t-enable-stack-protection-by-default.patch \
            file://0001-app.cc-use-old-school-asignment-to-avoid-gcc-4.8-err.patch \
            file://0001-Add-W_EXITCODE-macro-for-non-glibc-systems.patch \
            "
+
+SRC_URI_append_class-native = " \
+    file://0001-Reduce-gcc-required-support-downto-c-14.patch \
+    file://0001-Revert-parser-glue-Use-std-clamp.patch \
+    file://0001-vte.cc-replace-std-clamp-by-std-min-std-max.patch \
+"
+
 SRC_URI[archive.md5sum] = "a8984cd5a101dbff0b0c875d1de3f692"
 SRC_URI[archive.sha256sum] = "02fa8ecc02a9332e47f486795494527b5687b3bd448e73e6b67285f2f326dc7c"
 
Signature · 1 line
2.20.1
#2 lore raw
Quoted 3 lines · Andreas Müller

This should fix build on ancient hosts

This fixed debian9 but didn't fix centos7 :(

https://autobuilder.yoctoproject.org/typhoon/#/builders/79/builds/150

| checking for a sed that does not truncate output... (cached) sed
| checking whether g++  supports C++14 features by default... no
| checking whether g++  supports C++14 features with -std=gnu++14... no
| checking whether g++  supports C++14 features with -std=gnu++1y... no
| checking whether g++  supports C++14 features with -std=c++14... no
| checking whether g++  supports C++14 features with +std=c++14... no
| checking whether g++  supports C++14 features with -h std=c++14... no
| checking whether g++  supports C++14 features with -std=c++1y... no
| checking whether g++  supports C++14 features with +std=c++1y... no
| checking whether g++  supports C++14 features with -h std=c++1y... no
| configure: error: *** A compiler with support for C++14 language

features is required.

So we may have to look at reverting, or perhaps just adding a vte- native recipe with an older version of vte?

Cheers,

Richard

#3 lore raw
Quoted 21 lines · Richard Purdie
Quoted 3 lines · Andreas Müller

This should fix build on ancient hosts

This fixed debian9 but didn't fix centos7 :(

https://autobuilder.yoctoproject.org/typhoon/#/builders/79/builds/150

| checking for a sed that does not truncate output... (cached) sed
| checking whether g++  supports C++14 features by default... no
| checking whether g++  supports C++14 features with -std=gnu++14... no
| checking whether g++  supports C++14 features with -std=gnu++1y... no
| checking whether g++  supports C++14 features with -std=c++14... no
| checking whether g++  supports C++14 features with +std=c++14... no
| checking whether g++  supports C++14 features with -h std=c++14... no
| checking whether g++  supports C++14 features with -std=c++1y... no
| checking whether g++  supports C++14 features with +std=c++1y... no
| checking whether g++  supports C++14 features with -h std=c++1y... no
| configure: error: *** A compiler with support for C++14 language
Quoted 2 lines

features is required.

I expected that - turns nightmarish..

Quoted 2 lines

So we may have to look at reverting, or perhaps just adding a vte- native recipe with an older version of vte?

I have no problem with this but is it a good idea to run self test with a different version as the target one?

What about Ross' suggestion:

Quoted 3 lines

3. Don't use vte in qemu when building under virgl. The gtk PACKAGECONFIG turns on vte support unconditionally but from glancing at the qemu code that isn't a requirement.

Question: If this is accepted and works properly - could we get rid of vte-native then?

Let's decide some option: I really would like to close this and I have no problem with either way - gnome-terminal can wait. Gnome stuff just hobby to me and for test I can add recent vte in my personal meta-verse.

Andreas

#4 lore raw
Quoted 32 lines · Andreas Müller
Quoted 21 lines · Richard Purdie
Quoted 3 lines · Andreas Müller

This should fix build on ancient hosts

This fixed debian9 but didn't fix centos7 :(

https://autobuilder.yoctoproject.org/typhoon/#/builders/79/builds/150

Quoted 12 lines

checking for a sed that does not truncate output... (cached) sed checking whether g++ supports C++14 features by default... no checking whether g++ supports C++14 features with -std=gnu++14... no
checking whether g++ supports C++14 features with -std=gnu++1y... no
checking whether g++ supports C++14 features with -std=c++14... no
checking whether g++ supports C++14 features with +std=c++14... no
checking whether g++ supports C++14 features with -h std=c++14... no

Quoted 9 lines
Quoted 7 lines

checking whether g++ supports C++14 features with -std=c++1y... no
checking whether g++ supports C++14 features with +std=c++1y... no
checking whether g++ supports C++14 features with -h std=c++1y... no
configure: error: *** A compiler with support for C++14 language

features is required.

Quoted 6 lines

I expected that - turns nightmarish..

Quoted 2 lines

So we may have to look at reverting, or perhaps just adding a vte- native recipe with an older version of vte?

I have no problem with this but is it a good idea to run self test with a different version as the target one?

They're totally different uses so whilst I'd normally be cautious, I'm not sure it matters so much here.

Quoted 7 lines

What about Ross' suggestion:

Quoted 4 lines

3. Don't use vte in qemu when building under virgl. The gtk PACKAGECONFIG turns on vte support unconditionally but from glancing
at the qemu code that isn't a requirement.

Question: If this is accepted and works properly - could we get rid of vte-native then?

I've cc'd Alex, I'm sure there was a good reason we did this. I think it was to avoid having to have a GL enabled X display so I suspect this isn't an option.

Quoted 4 lines

Let's decide some option: I really would like to close this and I have no problem with either way - gnome-terminal can wait. Gnome stuff just hobby to me and for test I can add recent vte in my personal meta-verse.

Agreed, we need to make a decision, I'm just making sure we have all the options and then we "just" need to figure out which one to go for.

vte-native of an older version is looking like the best option unless Alex says I'm mis-remembering...

Cheers,

Richard

#5 lore raw
Quoted 10 lines · On Mon, 29 Apr 2019 at 23:22, <[email protected]> wrote:
Quoted 4 lines

Let's decide some option: I really would like to close this and I have no problem with either way - gnome-terminal can wait. Gnome stuff just hobby to me and for test I can add recent vte in my personal meta-verse.

Agreed, we need to make a decision, I'm just making sure we have all the options and then we "just" need to figure out which one to go for.

vte-native of an older version is looking like the best option unless Alex says I'm mis-remembering...

I checked what --enable-vte actually does in combination with the gtk UI in qemu, and it makes the serial console tab in the gtk ui behave like a real gnome-terminal tab (selection, scrolling etc). In the absence of vte support, the serial console tab is still there and you can still log in and issue commands, but it is more "dumb" regarding terminal capabilities. This serial console is an addition to the main tab where the output from the virtual graphical card goes.

So to conclude: we can split the vte option into its own PACKAGECONFIG, and leave it disabled for native. I would rather do this, than make vte and vte-native versions diverge.

Alex

#6 lore raw
Quoted 14 lines · Alexander Kanavin
Quoted 3 lines · On Mon, 29 Apr 2019 at 23:22, <[email protected]> wrote:

vte-native of an older version is looking like the best option unless
Alex says I'm mis-remembering...

I checked what --enable-vte actually does in combination with the gtk UI in qemu, and it makes the serial console tab in the gtk ui behave like a real gnome-terminal tab (selection, scrolling etc). In the absence of vte support, the serial console tab is still there and you can still log in and issue commands, but it is more "dumb" regarding terminal capabilities. This serial console is an addition to the main tab where the output from the virtual graphical card goes.

Quoted 3 lines

So to conclude: we can split the vte option into its own PACKAGECONFIG, and leave it disabled for native. I would rather do this, than make vte and vte-native versions diverge.

Sounds like a plan, can someone send me a patch please! :)

Cheers,

Richard

#7 lore raw
Quoted 20 lines · On Tue, Apr 30, 2019 at 12:53 PM <[email protected]> wrote:
Quoted 14 lines · Alexander Kanavin
Quoted 3 lines · On Mon, 29 Apr 2019 at 23:22, <[email protected]> wrote:

vte-native of an older version is looking like the best option unless
Alex says I'm mis-remembering...

I checked what --enable-vte actually does in combination with the gtk UI in qemu, and it makes the serial console tab in the gtk ui behave like a real gnome-terminal tab (selection, scrolling etc). In the absence of vte support, the serial console tab is still there and you can still log in and issue commands, but it is more "dumb" regarding terminal capabilities. This serial console is an addition to the main tab where the output from the virtual graphical card goes.

Quoted 3 lines

So to conclude: we can split the vte option into its own PACKAGECONFIG, and leave it disabled for native. I would rather do this, than make vte and vte-native versions diverge.

Quoted 6 lines

Sounds like a plan, can someone send me a patch please! :)

Cheers,

Richard

As the origin of all this I'll try to be the someone

Andreas