@@ -6,15 +6,24 @@
include(CMakePushCheckState)
-function(lldb_check_curses_tinfo CURSES_LIBRARIES CURSES_HAS_TINFO)
+function(lldb_check_curses_tinfo CURSES_HEADER CURSES_LIBRARIES
CURSES_HAS_TINFO)
cmake_reset_check_state()
set(CMAKE_REQUIRED_LIBRARIES "${CURSES_LIBRA
mysterymath wrote:
The reland broke the Fuchsia Windows build with a similar error to the LLVM
Windows bots, but I wasn't able to cleanly revert this commit without detailed
knowledge of the contents. Could you revert this (preferentially) or fix
forward?
https://github.com/llvm/llvm-project/
mysterymath wrote:
@JDevlieghere , we just landed a patch that adds packaging for Fuchsia. Should
be good to go ahead.
https://github.com/llvm/llvm-project/pull/93712
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi
mysterymath wrote:
> We have other packages the test suite relies on (pexpect, psutil, etc) so it
> seemed reasonable, but if folks feel strongly about it we can maintain our
> own "version parsing".
Notably the other packages are optional; the lion's share of the test suite can
be run withou
mysterymath wrote:
Did a bit of digging around; it looks like at the very least the Arch Linux
python package and the Python docker container don't contain the `packaging`
library.
https://github.com/llvm/llvm-project/pull/93712
___
lldb-commits mai
mysterymath wrote:
We don't have `pexpect` either; instead the tests that require it are disabled.
Our build is as hermetic as possible; even if the package were available in the
Python on the bots, we wouldn't be able to use it, since we don't make use of
the Python on the bots, for the most
mysterymath wrote:
At least at present, this will probably break Fuchsia's build; the Python we
ship includes `pkg_resources` but not `packaging`. Is the default `cpython`
build from source intended to build and install `packaging`? Is this something
that changed recently?
https://github.com/
mysterymath wrote:
I did some brief investigation on our end. The pkg-config file for libedit
includes a private dependency on libncurses, which provides the terminfo
functions. The issue here seems to be one uncovered in LLVM's
FindLibEdit.cmake. We're building against a static libedit, but t
https://github.com/mysterymath closed
https://github.com/llvm/llvm-project/pull/93094
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
https://github.com/mysterymath created
https://github.com/llvm/llvm-project/pull/93094
Reverts llvm/llvm-project#92572 due to Fuchsia CI breakages (using CLI tools in
tests that weren't necessarily built).
>From daa63e1870999e8ca8db454a788a0720740d194a Mon Sep 17 00:00:00 2001
From: Daniel Tho
mysterymath wrote:
This broke the Fuchsia CI builders' LLDB test suite:
https://luci-milo.appspot.com/ui/p/fuchsia/builders/toolchain.ci/clang-linux-x64/b8747217173107569041/overview
```
FAIL: LLDB (/b/s/w/ir/x/w/llvm_build/bin/clang-x86_64) ::
test_normal_stripped_split_with_dwp
(TestDebugin
https://github.com/mysterymath approved this pull request.
Thanks!
https://github.com/llvm/llvm-project/pull/84860
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
mysterymath wrote:
> @mysterymath do you run the API test suites with that custom python, i.e.
> `ninja check-lldb-api`? If so, would you be able to add bundle `pexpect` for
> your tests even if it's not in what you ship?
We do run the tests with the Python we ship, since we don't have direct
mysterymath wrote:
Chiming in on behalf of Fuchsia, since this did break us.
We build and ship a custom Python with our LLDB, since we're intended to be a
portable distribution. That Python doesn't have pexpect. I'm not very familiar
with what pexpect is used for, or the states of the world b
https://github.com/mysterymath approved this pull request.
https://github.com/llvm/llvm-project/pull/78605
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
@@ -58,15 +61,25 @@ Expected getDefaultDebuginfodCacheDirectory();
/// DEBUGINFOD_TIMEOUT environment variable, default is 90 seconds (9 ms).
std::chrono::milliseconds getDefaultDebuginfodTimeout();
+/// Get the full UrlPath for a source request of a given BuildID and file
@@ -58,15 +61,25 @@ Expected getDefaultDebuginfodCacheDirectory();
/// DEBUGINFOD_TIMEOUT environment variable, default is 90 seconds (9 ms).
std::chrono::milliseconds getDefaultDebuginfodTimeout();
+/// Get the full UrlPath for a source request of a given BuildID and file
@@ -58,15 +61,25 @@ Expected getDefaultDebuginfodCacheDirectory();
/// DEBUGINFOD_TIMEOUT environment variable, default is 90 seconds (9 ms).
std::chrono::milliseconds getDefaultDebuginfodTimeout();
+/// Get the full UrlPath for a source request of a given BuildID and file
@@ -46,6 +46,9 @@ bool canUseDebuginfod();
/// environment variable.
SmallVector getDefaultDebuginfodUrls();
+/// Creates the cache-key for a given Debuginfod UrlPath
mysterymath wrote:
```suggestion
/// Returns the cache key for a given debuginfod URL path.
https://github.com/mysterymath edited
https://github.com/llvm/llvm-project/pull/78605
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
https://github.com/mysterymath commented:
LGTM for debuginfod
https://github.com/llvm/llvm-project/pull/78605
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
https://github.com/mysterymath closed
https://github.com/llvm/llvm-project/pull/71112
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
mysterymath wrote:
> > Out of curiosity, why did you choose the delimiter as ' ' instead of
> > something like ';'?
>
> Because that's how the environment variable works. It was less a choice and
> more 🤷
>
> Also, ChatGPT tells me that URL's can include semicolons, so maybe it's not a
> gre
https://github.com/mysterymath edited
https://github.com/llvm/llvm-project/pull/70996
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
https://github.com/mysterymath edited
https://github.com/llvm/llvm-project/pull/70996
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
https://github.com/mysterymath edited
https://github.com/llvm/llvm-project/pull/70996
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
@@ -258,6 +258,8 @@ class TargetProperties : public Properties {
bool GetDebugUtilityExpression() const;
+ Args GetDebugInfoDURLs() const;
mysterymath wrote:
I tend to think of it as agglutanitive, like "blackbird."
https://github.com/llvm/llvm-project/p
https://github.com/mysterymath edited
https://github.com/llvm/llvm-project/pull/70996
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
https://github.com/mysterymath edited
https://github.com/llvm/llvm-project/pull/70996
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
https://github.com/mysterymath edited
https://github.com/llvm/llvm-project/pull/70996
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
@@ -396,8 +398,22 @@ Symbols::LocateExecutableSymbolFile(const ModuleSpec
&module_spec,
}
}
}
-
- return LocateExecutableSymbolFileDsym(module_spec);
+ FileSpec dsym_bundle = LocateExecutableSymbolFileDsym(module_spec);
+ if (dsym_bundle)
+return dsym_bundle
mysterymath wrote:
> The nodes on GreenDragon run under a UI session so that might explain why we
> don't see that issue there. Is there any harm with using
> `packages/Python/lldbsuite/test/make/entitlements.plist` for macOS too? Does
> `com.apple.private.security.no-sandbox` cause issues?
A
https://github.com/mysterymath edited
https://github.com/llvm/llvm-project/pull/71112
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
https://github.com/mysterymath edited
https://github.com/llvm/llvm-project/pull/71112
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
https://github.com/mysterymath created
https://github.com/llvm/llvm-project/pull/71112
Running the LLDB test suite in a GUI-less CI macOS environment requires that
the debugged tasks be signed with the get-task-allow entitlement.
NOTE: I'm not entirely sure why the upstream macOS builders don'
@@ -258,6 +258,8 @@ class TargetProperties : public Properties {
bool GetDebugUtilityExpression() const;
+ Args GetDebugInfoDURLs() const;
mysterymath wrote:
General naming nit:
`debuginfod` is all lower-case, and capitalized as if it were a single word:
@@ -396,8 +398,22 @@ Symbols::LocateExecutableSymbolFile(const ModuleSpec
&module_spec,
}
}
}
-
- return LocateExecutableSymbolFileDsym(module_spec);
+ FileSpec dsym_bundle = LocateExecutableSymbolFileDsym(module_spec);
+ if (dsym_bundle)
+return dsym_bundle
Author: Daniel Thornburgh
Date: 2023-09-20T14:30:24-07:00
New Revision: 7f3467412ed52d6749b7c53f7934c4374c6d2a00
URL:
https://github.com/llvm/llvm-project/commit/7f3467412ed52d6749b7c53f7934c4374c6d2a00
DIFF:
https://github.com/llvm/llvm-project/commit/7f3467412ed52d6749b7c53f7934c4374c6d2a00.d
mysterymath wrote:
It looks like this is breaking buildbots:
https://lab.llvm.org/buildbot/#/builders/68/builds/60267
The test can't find `main.cpp`, and it looks like the immediately preceding
test issues an `os.chdir`. It has a hook to reset it, but maybe there's some
problem with it?
http
https://github.com/mysterymath approved this pull request.
https://github.com/llvm/llvm-project/pull/66447
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
@@ -781,6 +781,9 @@ SBStructuredData SBDebugger::GetBuildConfiguration() {
AddBoolConfigEntry(
*config_up, "editline", LLDB_ENABLE_LIBEDIT,
"A boolean value that indicates if editline support is enabled in LLDB");
+ AddBoolConfigEntry(
+ *config_up, "editlin
mysterymath wrote:
I dug into this a little bit; it looks like the autodetection for
`LLDB_EDITLINE_USE_WCHAR` was broken on our build, so it was turned `OFF`, even
though our libedit supports the wchar API. I can fix this on our end, but more
generally, is there a way to condition this test o
Author: Daniel Thornburgh
Date: 2023-05-26T11:30:33-07:00
New Revision: b22965df1304a752af2724f7c2c3c58347532e71
URL:
https://github.com/llvm/llvm-project/commit/b22965df1304a752af2724f7c2c3c58347532e71
DIFF:
https://github.com/llvm/llvm-project/commit/b22965df1304a752af2724f7c2c3c58347532e71.d
Author: Daniel Thornburgh
Date: 2023-03-16T16:16:21-07:00
New Revision: 88eb4cbbfe7ded9556131c3bd4228ee9e9d01b4d
URL:
https://github.com/llvm/llvm-project/commit/88eb4cbbfe7ded9556131c3bd4228ee9e9d01b4d
DIFF:
https://github.com/llvm/llvm-project/commit/88eb4cbbfe7ded9556131c3bd4228ee9e9d01b4d.d
44 matches
Mail list logo
