| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Since commit c564a3e3ff (Ninja: Always compile sources using absolute
paths, 2021-05-19, v3.21.0-rc1~129^2), both the Ninja and Makefile
generators pass source files and include directories to the compiler as
absolute paths. However, in some other contexts within generated build
systems, we generate paths that may be relative or absolute. In these
contexts, we prefer relative paths, but avoid them when they contain a
`../` sequence that leaves both the build tree and the source tree:
* When the build tree is outside of the source tree, all paths to the
source tree are absolute.
* When the build tree is inside the source tree, we previously assumed
that it is a real directory such that exiting the build tree with
`../` enters the source tree. This allowed paths to the source
tree to be relative to the build tree.
In the latter case, we previously did not support using a symbolic link
inside the source tree to point at the build tree. This is because
relative paths to the source tree would be generated with `../`
sequences leaving the build tree, but they would jump to the parent of
the real build tree, which is not the source tree.
Fix this by requiring that `../` sequences stay inside the build tree
even if its path appears to be inside the source tree. When the build
tree is inside the source tree, all paths to the source tree are now
absolute. For consistency, this applies regardless of whether the
path to the build tree contains a symbolic link.
Fixes: #21819
|
|
|
|
|
| |
Refactor to set both at once so we have a single place in the code that
knows both have been set.
|
|
|
|
|
|
|
|
| |
Cache the Short Paths since we only convert the same few paths anyway
and calling `GetShortPathNameW` is really expensive.
Also, compile the code path only on Windows hosts since it only runs
when using a Windows Shell anyway.
|
|
|
|
| |
Fixes: #20070
|
| |
|
| |
|
|
|
|
|
| |
Helpers functions related to cmValue semantic are now part of
cmValue.h header.
|
|
|
|
| |
Move them up from cmLocalGenerator and out of cmStateDirectory.
|
| |
|
|
|
|
| |
Issue: #18870
|
| |
|
|
|
|
|
|
|
| |
In `ConvertToOutputForExisting` we convert paths with spaces to short
paths on Windows for use on command lines, e.g. for include directories.
Do the same for paths with `#` since tools like NMake do not have a way
to reliably put `#` in variable assignments.
|
|
|
|
| |
Co-Authored-by: vector-of-bool <vectorofbool@gmail.com>
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
Enable the check in .clang-tidy and fix all warnings.
|
| |
|
| |
|
|
|
|
| |
This allows the linker to entirely drop any symbols of the functions.
|
|
|
|
|
|
|
| |
Use std::string (with correct initial size) in cmOutputConverter::Shell__GetArgument
instead of ostringstream. This avoids several re-allocations of the
string buffer. In addition, convert some of the private static members into
inline free functions to avoid function calls.
|
|
|
|
|
|
|
|
|
|
| |
* Change some functions to take `std::string` instead of
`const char*` in the following classes: `cmMakeFile`, `cmake`,
`cmCoreTryCompile`, `cmSystemTools`, `cmState`, `cmLocalGenerator`
and a few others.
* Greatly reduce using of `const char*` overloads for
`cmSystemTools::MakeDirectory` and `cmSystemTools::RelativePath`.
* Remove many redundant `c_str()` conversions throughout the code.
|
|
|
|
|
|
| |
Changes done via `clang-tidy` with some manual fine-tuning
for the variable naming and `auto` type deduction
where appropriate.
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
| |
Port dependents to the new locations as needed.
Leave behind a cmState.h include in cmListFileCache to reduce noise. It
is removed in a following commit.
|
| |
|
| |
|
|
|
|
| |
Remove the need for method parameters to represent the distinction.
|
|
|
|
| |
They don't provide real value.
|
|
|
|
|
|
| |
The conditional early return can be moved to clients, which would have
many benefits, notably making cmOutputConverter independent of
directory-specific state.
|
|
|
|
|
| |
This means that we don't encounter the asserts in the case where we
early-return from here.
|
|
|
|
| |
This function delegates to another function which does the same check.
|
| |
|
|
|
|
| |
Remove comment made obsolete by them.
|
|
|
|
|
| |
Remove old ConvertToRelativePath function now that all clients have
migrated to the new signature.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Per-source copyright/license notice headers that spell out copyright holder
names and years are hard to maintain and often out-of-date or plain wrong.
Precise contributor information is already maintained automatically by the
version control tool. Ultimately it is the receiver of a file who is
responsible for determining its licensing status, and per-source notices are
merely a convenience. Therefore it is simpler and more accurate for
each source to have a generic notice of the license name and references to
more detailed information on copyright holders and full license terms.
Our `Copyright.txt` file now contains a list of Contributors whose names
appeared source-level copyright notices. It also references version control
history for more precise information. Therefore we no longer need to spell
out the list of Contributors in each source file notice.
Replace CMake per-source copyright/license notice headers with a short
description of the license and links to `Copyright.txt` and online information
available from "https://cmake.org/licensing". The online URL also handles
cases of modules being copied out of our source into other projects, so we
can drop our notices about replacing links with full license text.
Run the `Utilities/Scripts/filter-notices.bash` script to perform the majority
of the replacements mechanically. Manually fix up shebang lines and trailing
newlines in a few files. Manually update the notices in a few files that the
script does not handle.
|
| |
|
| |
|
|
|
|
| |
Make it more clear what is happening here.
|