Re: [PATCH] D13407: [libcxx] Capture configuration information when installing the libc++ headers
EricWF accepted this revision. EricWF added a reviewer: EricWF. EricWF added a comment. This revision is now accepted and ready to land. Maintainers from Apple and FreeBSD gave this the thumbs up. http://reviews.llvm.org/D13407 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
Re: [PATCH] D13407: [libcxx] Capture configuration information when installing the libc++ headers
EricWF updated this revision to Diff 37013.
EricWF added a comment.
Make the generated config dependent on the "__config_site" file. It probably
didn't need this but it doesn't hurt to have.
http://reviews.llvm.org/D13407
Files:
CMakeLists.txt
cmake/Modules/HandleLibcxxFlags.cmake
docs/DesignDocs/CapturingConfigInfo.rst
docs/index.rst
include/CMakeLists.txt
include/__config_site.in
Index: include/__config_site.in
===
--- /dev/null
+++ include/__config_site.in
@@ -0,0 +1,20 @@
+//===--===//
+//
+// The LLVM Compiler Infrastructure
+//
+// This file is dual licensed under the MIT and the University of Illinois Open
+// Source Licenses. See LICENSE.TXT for details.
+//
+//===--===//
+
+#ifndef _LIBCPP_CONFIG_SITE
+#define _LIBCPP_CONFIG_SITE
+
+#cmakedefine _LIBCPP_HAS_NO_GLOBAL_FILESYSTEM_NAMESPACE
+#cmakedefine _LIBCPP_HAS_NO_STDIN
+#cmakedefine _LIBCPP_HAS_NO_STDOUT
+#cmakedefine _LIBCPP_HAS_NO_THREADS
+#cmakedefine _LIBCPP_HAS_NO_MONOTONIC_CLOCK
+#cmakedefine _LIBCPP_HAS_NO_THREAD_UNSAFE_C_FUNCTIONS
+
+#endif
Index: include/CMakeLists.txt
===
--- include/CMakeLists.txt
+++ include/CMakeLists.txt
@@ -1,10 +1,12 @@
if (NOT LIBCXX_INSTALL_SUPPORT_HEADERS)
set(LIBCXX_SUPPORT_HEADER_PATTERN PATTERN "support" EXCLUDE)
endif()
+
set(LIBCXX_HEADER_PATTERN
PATTERN "*"
PATTERN "CMakeLists.txt" EXCLUDE
PATTERN ".svn" EXCLUDE
+ PATTERN "__config_site.in" EXCLUDE
${LIBCXX_SUPPORT_HEADER_PATTERN}
)
@@ -22,4 +24,29 @@
${LIBCXX_HEADER_PATTERN}
PERMISSIONS OWNER_READ OWNER_WRITE GROUP_READ WORLD_READ
)
+
+ if (LIBCXX_NEEDS_SITE_CONFIG)
+set(UNIX_CAT cat)
+if (WIN32)
+ set(UNIX_CAT type)
+endif()
+# Generate and install a custom __config header. The new header is created
+# by prepending __config_site to the current __config header.
+add_custom_command(OUTPUT ${LIBCXX_BINARY_DIR}/__generated_config
+ COMMAND ${CMAKE_COMMAND} -E copy ${LIBCXX_BINARY_DIR}/__config_site ${LIBCXX_BINARY_DIR}/__generated_config
+ COMMAND ${UNIX_CAT} ${LIBCXX_SOURCE_DIR}/include/__config >> ${LIBCXX_BINARY_DIR}/__generated_config
+ DEPENDS ${LIBCXX_SOURCE_DIR}/include/__config
+ ${LIBCXX_BINARY_DIR}/__config_site
+)
+# Add a target that executes the generation commands.
+add_custom_target(generate_config_header ALL
+ DEPENDS ${LIBCXX_BINARY_DIR}/__generated_config)
+# Install the generated header as __config.
+install(FILES ${LIBCXX_BINARY_DIR}/__generated_config
+ DESTINATION include/c++/v1
+ PERMISSIONS OWNER_READ OWNER_WRITE GROUP_READ WORLD_READ
+ RENAME __config
+ COMPONENT libcxx)
+ endif()
+
endif()
Index: docs/index.rst
===
--- docs/index.rst
+++ docs/index.rst
@@ -124,6 +124,12 @@
Design Documents
+.. toctree::
+ :maxdepth: 1
+
+ DesignDocs/CapturingConfigInfo
+
+
* ` design `_
* ` design `_
* `Status of debug mode `_
Index: docs/DesignDocs/CapturingConfigInfo.rst
===
--- /dev/null
+++ docs/DesignDocs/CapturingConfigInfo.rst
@@ -0,0 +1,88 @@
+===
+Capturing configuration information during installation
+===
+
+.. contents::
+ :local:
+
+The Problem
+===
+
+Currently the libc++ supports building the library with a number of different
+configuration options. Unfortunately all of that configuration information is
+lost when libc++ is installed. In order to support "persistent"
+configurations libc++ needs a mechanism to capture the configuration options
+in the INSTALLED headers.
+
+
+Design Goals
+
+
+* The solution should not INSTALL any additional headers. We don't want an extra
+ #include slowing everybody down.
+
+* The solution should not unduly affect libc++ developers. The problem is limited
+ to installed versions of libc++ and the solution should be as well.
+
+* The solution should not modify any existing headers EXCEPT during installation.
+ It makes developers lives harder if they have to regenerate the libc++ headers
+ every time they are modified.
+
+* The solution should not make any of the libc++ headers dependant on
+ files generated by the build system. The headers should be able to compile
+ out of the box without any modification.
+
+* The solution should not have ANY effect on users who don't need special
+ configuration options. The vast majority of users will never n
Re: [PATCH] D13407: [libcxx] Capture configuration information when installing the libc++ headers
EricWF marked an inline comment as done.
Comment at: include/CMakeLists.txt:38
@@ +37,3 @@
+ COMMAND ${UNIX_CAT} ${LIBCXX_SOURCE_DIR}/include/__config >>
${LIBCXX_BINARY_DIR}/__generated_config
+ DEPENDS ${LIBCXX_SOURCE_DIR}/include/__config
+)
eugenis wrote:
> I think you need to depend on __config_site.in as well. Or on
> ${LIBCXX_BINARY_DIR}/__config_site
I don't think we need it. `__config_site` is generated with `configure_file`,
which runs at configuration time and re-runs whenever the input file is
changed. So `__config_site` should always be up to date.
However I'll see if we can actually make the change. It seems clearer and safer.
http://reviews.llvm.org/D13407
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
Re: [PATCH] D13407: [libcxx] Capture configuration information when installing the libc++ headers
eugenis added a comment.
Thanks. I'll upload the rebase patch on Monday.
Comment at: include/CMakeLists.txt:38
@@ +37,3 @@
+ COMMAND ${UNIX_CAT} ${LIBCXX_SOURCE_DIR}/include/__config >>
${LIBCXX_BINARY_DIR}/__generated_config
+ DEPENDS ${LIBCXX_SOURCE_DIR}/include/__config
+)
I think you need to depend on __config_site.in as well. Or on
${LIBCXX_BINARY_DIR}/__config_site
http://reviews.llvm.org/D13407
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
Re: [PATCH] D13407: [libcxx] Capture configuration information when installing the libc++ headers
EricWF marked an inline comment as done.
EricWF added a comment.
@eugenis I'll commit this tuesday morning If I don't hear anything over the
weekend (Monday is a holiday for me).
If you rebase the ABI patch to work with this I'll make sure to review it ASAP.
Comment at: include/CMakeLists.txt:9
@@ -7,2 +8,3 @@
PATTERN ".svn" EXCLUDE
+ PATTERN "__config_site.in" EXCLUDE
${LIBCXX_SUPPORT_HEADER_PATTERN}
vkalintiris wrote:
> Kind of silly but I believe that the files used for auto-generated config
> headers in LLVM have the extesions `.cmake` and `.in`, for cmake and
> autoconf, respectively.
libc++ uses `.cmake` for CMake scripts and `.in` for auto-generated files and I
think this is fine. libc++ doesn't support autoconf I'm not worried about it
causing confusion.
http://reviews.llvm.org/D13407
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
Re: [PATCH] D13407: [libcxx] Capture configuration information when installing the libc++ headers
vkalintiris added a subscriber: vkalintiris.
Comment at: include/CMakeLists.txt:9
@@ -7,2 +8,3 @@
PATTERN ".svn" EXCLUDE
+ PATTERN "__config_site.in" EXCLUDE
${LIBCXX_SUPPORT_HEADER_PATTERN}
Kind of silly but I believe that the files used for auto-generated config
headers in LLVM have the extesions `.cmake` and `.in`, for cmake and autoconf,
respectively.
http://reviews.llvm.org/D13407
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
Re: [PATCH] D13407: [libcxx] Capture configuration information when installing the libc++ headers
EricWF added a comment. Cool. I've reached out to some platform maintainers to make sure they don't see any problems. I'll commit this once I hear back. http://reviews.llvm.org/D13407 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
Re: [PATCH] D13407: [libcxx] Capture configuration information when installing the libc++ headers
eugenis added a comment. In http://reviews.llvm.org/D13407#261228, @EricWF wrote: > Use `type` instead of `cat` on windows as suggested by @rnk. @eugenis does > this address your concern? Absolutely! LGTM http://reviews.llvm.org/D13407 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
Re: [PATCH] D13407: [libcxx] Capture configuration information when installing the libc++ headers
EricWF updated this revision to Diff 36682.
EricWF added a comment.
Use `type` instead of `cat` on windows as suggested by @rnk. @eugenis does this
address your concern?
http://reviews.llvm.org/D13407
Files:
CMakeLists.txt
cmake/Modules/HandleLibcxxFlags.cmake
docs/DesignDocs/CapturingConfigInfo.rst
docs/index.rst
include/CMakeLists.txt
include/__config_site.in
Index: include/__config_site.in
===
--- /dev/null
+++ include/__config_site.in
@@ -0,0 +1,20 @@
+//===--===//
+//
+// The LLVM Compiler Infrastructure
+//
+// This file is dual licensed under the MIT and the University of Illinois Open
+// Source Licenses. See LICENSE.TXT for details.
+//
+//===--===//
+
+#ifndef _LIBCPP_CONFIG_SITE
+#define _LIBCPP_CONFIG_SITE
+
+#cmakedefine _LIBCPP_HAS_NO_GLOBAL_FILESYSTEM_NAMESPACE
+#cmakedefine _LIBCPP_HAS_NO_STDIN
+#cmakedefine _LIBCPP_HAS_NO_STDOUT
+#cmakedefine _LIBCPP_HAS_NO_THREADS
+#cmakedefine _LIBCPP_HAS_NO_MONOTONIC_CLOCK
+#cmakedefine _LIBCPP_HAS_NO_THREAD_UNSAFE_C_FUNCTIONS
+
+#endif
Index: include/CMakeLists.txt
===
--- include/CMakeLists.txt
+++ include/CMakeLists.txt
@@ -1,10 +1,12 @@
if (NOT LIBCXX_INSTALL_SUPPORT_HEADERS)
set(LIBCXX_SUPPORT_HEADER_PATTERN PATTERN "support" EXCLUDE)
endif()
+
set(LIBCXX_HEADER_PATTERN
PATTERN "*"
PATTERN "CMakeLists.txt" EXCLUDE
PATTERN ".svn" EXCLUDE
+ PATTERN "__config_site.in" EXCLUDE
${LIBCXX_SUPPORT_HEADER_PATTERN}
)
@@ -22,4 +24,28 @@
${LIBCXX_HEADER_PATTERN}
PERMISSIONS OWNER_READ OWNER_WRITE GROUP_READ WORLD_READ
)
+
+ if (LIBCXX_NEEDS_SITE_CONFIG)
+set(UNIX_CAT cat)
+if (WIN32)
+ set(UNIX_CAT type)
+endif()
+# Generate and install a custom __config header. The new header is created
+# by prepending __config_site to the current __config header.
+add_custom_command(OUTPUT ${LIBCXX_BINARY_DIR}/__generated_config
+ COMMAND ${CMAKE_COMMAND} -E copy ${LIBCXX_BINARY_DIR}/__config_site ${LIBCXX_BINARY_DIR}/__generated_config
+ COMMAND ${UNIX_CAT} ${LIBCXX_SOURCE_DIR}/include/__config >> ${LIBCXX_BINARY_DIR}/__generated_config
+ DEPENDS ${LIBCXX_SOURCE_DIR}/include/__config
+)
+# Add a target that executes the generation commands.
+add_custom_target(generate_config_header ALL
+ DEPENDS ${LIBCXX_BINARY_DIR}/__generated_config)
+# Install the generated header as __config.
+install(FILES ${LIBCXX_BINARY_DIR}/__generated_config
+ DESTINATION include/c++/v1
+ PERMISSIONS OWNER_READ OWNER_WRITE GROUP_READ WORLD_READ
+ RENAME __config
+ COMPONENT libcxx)
+ endif()
+
endif()
Index: docs/index.rst
===
--- docs/index.rst
+++ docs/index.rst
@@ -124,6 +124,12 @@
Design Documents
+.. toctree::
+ :maxdepth: 1
+
+ DesignDocs/CapturingConfigInfo
+
+
* ` design `_
* ` design `_
* `Status of debug mode `_
Index: docs/DesignDocs/CapturingConfigInfo.rst
===
--- /dev/null
+++ docs/DesignDocs/CapturingConfigInfo.rst
@@ -0,0 +1,88 @@
+===
+Capturing configuration information during installation
+===
+
+.. contents::
+ :local:
+
+The Problem
+===
+
+Currently the libc++ supports building the library with a number of different
+configuration options. Unfortunately all of that configuration information is
+lost when libc++ is installed. In order to support "persistent"
+configurations libc++ needs a mechanism to capture the configuration options
+in the INSTALLED headers.
+
+
+Design Goals
+
+
+* The solution should not INSTALL any additional headers. We don't want an extra
+ #include slowing everybody down.
+
+* The solution should not unduly affect libc++ developers. The problem is limited
+ to installed versions of libc++ and the solution should be as well.
+
+* The solution should not modify any existing headers EXCEPT during installation.
+ It makes developers lives harder if they have to regenerate the libc++ headers
+ every time they are modified.
+
+* The solution should not make any of the libc++ headers dependant on
+ files generated by the build system. The headers should be able to compile
+ out of the box without any modification.
+
+* The solution should not have ANY effect on users who don't need special
+ configuration options. The vast majority of users will never need this so it
+ shouldn't cost them.
+
+
+The Solution
+
Re: [PATCH] D13407: [libcxx] Capture configuration information when installing the libc++ headers
rnk added a subscriber: rnk.
Comment at: include/CMakeLists.txt:31
@@ +30,3 @@
+# by prepending __config_site to the current __config header.
+# TODO(EricWF) Is it portable to use "cat" and ">>"?
+add_custom_command(OUTPUT ${LIBCXX_BINARY_DIR}/__generated_config
EricWF wrote:
> eugenis wrote:
> > EricWF wrote:
> > > eugenis wrote:
> > > > I don't think >> would work on windows.
> > > > Do you really need __generated_config to be created at build time (as
> > > > opposed to configure time)? You could use file(read) and file(append)
> > > > then.
> > > >
> > > I would strongly prefer the file got generated at build time so that it
> > > contains any changes made in the source tree. Any other behavior is
> > > developer hostile and non-obvious. In order to keep the installed headers
> > > consistent we need to do this. Although I hope there is a better way to
> > > achieve this.
> > Right, good point. Then you could go back to the approach in D11963 where
> > you called cmake in a custom command with a small script that used file(*)
> > commands.
> >
> That approach had a problem because the cmake INSTALL command used to invoke
> the script doesn't take a "COMPONENT" argument which would break the
> "install-libcxx" rule. On windows we could do it with a call out to python or
> a shell script. Would that work?
The >> operator should work on Windows. It's supported by cmd. However, cat
generally isn't available. If you use 'type' in place of 'cat' it should work.
http://reviews.llvm.org/D13407
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
Re: [PATCH] D13407: [libcxx] Capture configuration information when installing the libc++ headers
EricWF added inline comments.
Comment at: include/CMakeLists.txt:31
@@ +30,3 @@
+# by prepending __config_site to the current __config header.
+# TODO(EricWF) Is it portable to use "cat" and ">>"?
+add_custom_command(OUTPUT ${LIBCXX_BINARY_DIR}/__generated_config
eugenis wrote:
> EricWF wrote:
> > eugenis wrote:
> > > I don't think >> would work on windows.
> > > Do you really need __generated_config to be created at build time (as
> > > opposed to configure time)? You could use file(read) and file(append)
> > > then.
> > >
> > I would strongly prefer the file got generated at build time so that it
> > contains any changes made in the source tree. Any other behavior is
> > developer hostile and non-obvious. In order to keep the installed headers
> > consistent we need to do this. Although I hope there is a better way to
> > achieve this.
> Right, good point. Then you could go back to the approach in D11963 where you
> called cmake in a custom command with a small script that used file(*)
> commands.
>
That approach had a problem because the cmake INSTALL command used to invoke
the script doesn't take a "COMPONENT" argument which would break the
"install-libcxx" rule. On windows we could do it with a call out to python or a
shell script. Would that work?
http://reviews.llvm.org/D13407
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
Re: [PATCH] D13407: [libcxx] Capture configuration information when installing the libc++ headers
eugenis added inline comments.
Comment at: include/CMakeLists.txt:31
@@ +30,3 @@
+# by prepending __config_site to the current __config header.
+# TODO(EricWF) Is it portable to use "cat" and ">>"?
+add_custom_command(OUTPUT ${LIBCXX_BINARY_DIR}/__generated_config
EricWF wrote:
> eugenis wrote:
> > I don't think >> would work on windows.
> > Do you really need __generated_config to be created at build time (as
> > opposed to configure time)? You could use file(read) and file(append) then.
> >
> I would strongly prefer the file got generated at build time so that it
> contains any changes made in the source tree. Any other behavior is developer
> hostile and non-obvious. In order to keep the installed headers consistent we
> need to do this. Although I hope there is a better way to achieve this.
Right, good point. Then you could go back to the approach in D11963 where you
called cmake in a custom command with a small script that used file(*) commands.
http://reviews.llvm.org/D13407
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
Re: [PATCH] D13407: [libcxx] Capture configuration information when installing the libc++ headers
EricWF added inline comments.
Comment at: include/CMakeLists.txt:31
@@ +30,3 @@
+# by prepending __config_site to the current __config header.
+# TODO(EricWF) Is it portable to use "cat" and ">>"?
+add_custom_command(OUTPUT ${LIBCXX_BINARY_DIR}/__generated_config
eugenis wrote:
> I don't think >> would work on windows.
> Do you really need __generated_config to be created at build time (as opposed
> to configure time)? You could use file(read) and file(append) then.
>
I would strongly prefer the file got generated at build time so that it
contains any changes made in the source tree. Any other behavior is developer
hostile and non-obvious. In order to keep the installed headers consistent we
need to do this. Although I hope there is a better way to achieve this.
http://reviews.llvm.org/D13407
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
Re: [PATCH] D13407: [libcxx] Capture configuration information when installing the libc++ headers
eugenis added a comment.
Looks great!
Comment at: include/CMakeLists.txt:31
@@ +30,3 @@
+# by prepending __config_site to the current __config header.
+# TODO(EricWF) Is it portable to use "cat" and ">>"?
+add_custom_command(OUTPUT ${LIBCXX_BINARY_DIR}/__generated_config
I don't think >> would work on windows.
Do you really need __generated_config to be created at build time (as opposed
to configure time)? You could use file(read) and file(append) then.
http://reviews.llvm.org/D13407
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
Re: [PATCH] D13407: [libcxx] Capture configuration information when installing the libc++ headers
jroelofs added a comment. Design doc looks good! I think it captures all the design constraints well, and the proposed solution satisfies those constraints as far as I am concerned. Thumbs up from me... but I'll still defer the final 'LGTM' to @mclow.lists. http://reviews.llvm.org/D13407 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
Re: [PATCH] D13407: [libcxx] Capture configuration information when installing the libc++ headers
EricWF updated this revision to Diff 36457.
EricWF added a comment.
I removed all non-essential changes in order to make this easy and fast to
review. This patch is blocking a number of people so I would like to have this
reviewed soon.
http://reviews.llvm.org/D13407
Files:
CMakeLists.txt
cmake/Modules/HandleLibcxxFlags.cmake
docs/DesignDocs/CapturingConfigInfo.rst
docs/index.rst
include/CMakeLists.txt
include/__config_site.in
Index: include/__config_site.in
===
--- /dev/null
+++ include/__config_site.in
@@ -0,0 +1,20 @@
+//===--===//
+//
+// The LLVM Compiler Infrastructure
+//
+// This file is dual licensed under the MIT and the University of Illinois Open
+// Source Licenses. See LICENSE.TXT for details.
+//
+//===--===//
+
+#ifndef _LIBCPP_CONFIG_SITE
+#define _LIBCPP_CONFIG_SITE
+
+#cmakedefine _LIBCPP_HAS_NO_GLOBAL_FILESYSTEM_NAMESPACE
+#cmakedefine _LIBCPP_HAS_NO_STDIN
+#cmakedefine _LIBCPP_HAS_NO_STDOUT
+#cmakedefine _LIBCPP_HAS_NO_THREADS
+#cmakedefine _LIBCPP_HAS_NO_MONOTONIC_CLOCK
+#cmakedefine _LIBCPP_HAS_NO_THREAD_UNSAFE_C_FUNCTIONS
+
+#endif
Index: include/CMakeLists.txt
===
--- include/CMakeLists.txt
+++ include/CMakeLists.txt
@@ -1,10 +1,12 @@
if (NOT LIBCXX_INSTALL_SUPPORT_HEADERS)
set(LIBCXX_SUPPORT_HEADER_PATTERN PATTERN "support" EXCLUDE)
endif()
+
set(LIBCXX_HEADER_PATTERN
PATTERN "*"
PATTERN "CMakeLists.txt" EXCLUDE
PATTERN ".svn" EXCLUDE
+ PATTERN "__config_site.in" EXCLUDE
${LIBCXX_SUPPORT_HEADER_PATTERN}
)
@@ -22,4 +24,25 @@
${LIBCXX_HEADER_PATTERN}
PERMISSIONS OWNER_READ OWNER_WRITE GROUP_READ WORLD_READ
)
+
+ if (LIBCXX_NEEDS_SITE_CONFIG)
+# Generate and install a custom __config header. The new header is created
+# by prepending __config_site to the current __config header.
+# TODO(EricWF) Is it portable to use "cat" and ">>"?
+add_custom_command(OUTPUT ${LIBCXX_BINARY_DIR}/__generated_config
+ COMMAND ${CMAKE_COMMAND} -E copy ${LIBCXX_BINARY_DIR}/__config_site ${LIBCXX_BINARY_DIR}/__generated_config
+ COMMAND cat ${LIBCXX_SOURCE_DIR}/include/__config >> ${LIBCXX_BINARY_DIR}/__generated_config
+ DEPENDS ${LIBCXX_SOURCE_DIR}/include/__config
+)
+# Add a target that executes the generation commands.
+add_custom_target(generate_config_header ALL
+ DEPENDS ${LIBCXX_BINARY_DIR}/__generated_config)
+# Install the generated header as __config.
+install(FILES ${LIBCXX_BINARY_DIR}/__generated_config
+ DESTINATION include/c++/v1
+ PERMISSIONS OWNER_READ OWNER_WRITE GROUP_READ WORLD_READ
+ RENAME __config
+ COMPONENT libcxx)
+ endif()
+
endif()
Index: docs/index.rst
===
--- docs/index.rst
+++ docs/index.rst
@@ -124,6 +124,12 @@
Design Documents
+.. toctree::
+ :maxdepth: 1
+
+ DesignDocs/CapturingConfigInfo
+
+
* ` design `_
* ` design `_
* `Status of debug mode `_
Index: docs/DesignDocs/CapturingConfigInfo.rst
===
--- /dev/null
+++ docs/DesignDocs/CapturingConfigInfo.rst
@@ -0,0 +1,89 @@
+===
+Capturing configuration information during installation
+===
+
+.. contents::
+ :local:
+
+The Problem
+===
+
+Currently the libc++ supports building the library with a number of different
+configuration options. Unfortunately all of that configuration information is
+lost when libc++ is installed. In order to support "persistent"
+configurations libc++ needs a mechanism to capture the configuration options
+in the INSTALLED headers.
+
+
+Design Goals
+
+
+* The solution should not INSTALL any additional headers. We don't want an extra
+ #include slowing everybody down.
+
+* The solution should not unduly affect libc++ developers. The problem is limited
+ to installed versions of libc++ and the solution should be as well.
+
+* The solution should not modify any existing headers EXCEPT during installation.
+ It makes developers lives harder if they have to regenerate the libc++ headers
+ every time they are modified.
+
+* The solution should not make any of the libc++ headers dependant on
+ files generated by the build system. The headers should be able to compile
+ out of the box without any modification.
+
+* The solution should not have ANY effect on users who don't need special
+ configuration options. The vast majority of users will never need this so it
+ shouldn't cos
Re: [PATCH] D13407: [libcxx] Capture configuration information when installing the libc++ headers
EricWF added subscribers: eugenis, asl, espositofulvio, ed. EricWF added a comment. Adding subscribes from the old patch to this one. http://reviews.llvm.org/D13407 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
Re: [PATCH] D13407: [libcxx] Capture configuration information when installing the libc++ headers
EricWF added a comment. Oh, but don't review the design doc too closely. I'm still working on it. http://reviews.llvm.org/D13407 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
