labath (Pavel Labath)
User

Projects

User does not belong to any projects.

User Details

User Since
Jun 4 2013, 6:02 AM (246 w, 6 h)

Recent Activity

Today

labath created D43506: Fix a couple of more tests to not create files in the source tree.
Tue, Feb 20, 3:26 AM
labath committed rL325570: Avoid dirtying the source tree in breakpoint command tests.
Avoid dirtying the source tree in breakpoint command tests
Tue, Feb 20, 2:28 AM
labath closed D43464: Avoid dirtying the source tree in breakpoint command tests.
Tue, Feb 20, 2:28 AM
labath added inline comments to D43345: [LLDB] Initial version of PPC64 InstEmulation.
Tue, Feb 20, 2:18 AM

Yesterday

labath committed rL325519: Two more dosep-paralellization fallout fixes.
Two more dosep-paralellization fallout fixes
Mon, Feb 19, 9:25 AM
labath added a comment to D43471: Handle typeof() expressions.

Seems straight-forward enough, but technically Jim is the owner of the expression evaluator these days, so I'll leave the honours to him.

Mon, Feb 19, 9:24 AM
labath added a comment to D43345: [LLDB] Initial version of PPC64 InstEmulation.

How would you feel about replacing the custom debug printfs with standard logging messages?

That way, you can always have them available without recompiling (and it's always nice to have more uniformity).

Ok. That was my initial idea, but I was not sure about which category to log the messages to, and looking into the other instruction emulation plugins I found no log messages, so I ended up leaving only debug printfs.

I guess the best category would be LIBLLDB_LOG_UNWIND, right?

Mon, Feb 19, 9:01 AM
labath committed rL325516: [CodeGen] Refactor AppleAccelTable.
[CodeGen] Refactor AppleAccelTable
Mon, Feb 19, 8:14 AM
labath closed D43285: [CodeGen] Refactor AppleAccelTable.
Mon, Feb 19, 8:14 AM
labath added a comment to D43345: [LLDB] Initial version of PPC64 InstEmulation.

How would you feel about replacing the custom debug printfs with standard logging messages?

Mon, Feb 19, 7:55 AM
labath committed rL325511: Fix TestStopReplyContainsThreadPcs on 32-bit x86 (pr36013).
Fix TestStopReplyContainsThreadPcs on 32-bit x86 (pr36013)
Mon, Feb 19, 7:44 AM
labath closed D43376: Fix TestStopReplyContainsThreadPcs on 32-bit x86 (pr36013).
Mon, Feb 19, 7:44 AM
labath committed rL325504: Add SBDebugger::GetBuildConfiguration and use it to skip an XML test.
Add SBDebugger::GetBuildConfiguration and use it to skip an XML test
Mon, Feb 19, 7:08 AM
labath closed D43333: Add SBDebugger::GetBuildConfiguration and use it to skip an XML test.
Mon, Feb 19, 7:08 AM
labath created D43464: Avoid dirtying the source tree in breakpoint command tests.
Mon, Feb 19, 7:07 AM
labath committed rL325495: Make gdb-client tests generate binaries in the build tree.
Make gdb-client tests generate binaries in the build tree
Mon, Feb 19, 5:56 AM

Sat, Feb 17

labath requested changes to D43419: Fix TestBreakpointInGlobalConstructor for Windows.

Yeah, setting this to zero would break other platforms that are able to locate shared libraries before running the executable. On linux, we try to locate them as well, but it's kind of on a best-effort basis -- it's quite hard to figure out what library will get loaded with absolute precision, as it can e.g. depend on the value of LD_LIBRARY_PATH env var that you run the process with (and you don't know that until the actual "process launch" command). In fact, it wouldn't be hard to come up with examples where this static resolution finds the wrong library.

Sat, Feb 17, 4:39 AM

Fri, Feb 16

labath updated the diff for D43285: [CodeGen] Refactor AppleAccelTable.

Remove the detail namespace, and rename the EmissionContext class.

Fri, Feb 16, 6:18 AM
labath added inline comments to D43285: [CodeGen] Refactor AppleAccelTable.
Fri, Feb 16, 6:15 AM
labath committed rL325340: Shorten socket names in TestPlatformProcessConnect.
Shorten socket names in TestPlatformProcessConnect
Fri, Feb 16, 5:01 AM
labath committed rL325334: Fix paralelization of remote tests.
Fix paralelization of remote tests
Fri, Feb 16, 3:43 AM
labath updated subscribers of D43345: [LLDB] Initial version of PPC64 InstEmulation.

This FileCheck-style test seems to be easier to write and understand.
In fact, I wrote a gtest-style test because:
1- I didn't know I could write a FileCheck-style test instead.
2- This was inspired by the ARM64 plugin, that seemed like the simplest one to understand and follow.

Let me try to convert it to a FileCheck-style test.

Fri, Feb 16, 3:36 AM
labath created D43376: Fix TestStopReplyContainsThreadPcs on 32-bit x86 (pr36013).
Fri, Feb 16, 2:07 AM
labath abandoned D43292: [dotest] Add ability to skip tests based on build configuration.

It looks like we'll do something like D43333 instead...

Fri, Feb 16, 1:36 AM
labath committed rL325324: [lldb][PPC64] Fixed vector and struct return value.
[lldb][PPC64] Fixed vector and struct return value
Fri, Feb 16, 1:32 AM
labath closed D42468: [lldb][PPC64] Fixed vector and struct return value.
Fri, Feb 16, 1:31 AM
labath committed rL325322: [dosep] Run tests in a more parallel fashion.
[dosep] Run tests in a more parallel fashion
Fri, Feb 16, 1:25 AM
labath closed D43335: [dosep] Run tests in a more parallel fashion.
Fri, Feb 16, 1:25 AM
labath updated the diff for D43333: Add SBDebugger::GetBuildConfiguration and use it to skip an XML test.

I've updated the format to include a description of the item. I don't we need to
worry about the expensiveness of this data structure too much, as it should
never be used in a hot loop (the only use case I can think of is querying the
configuration at start-up).

Fri, Feb 16, 1:16 AM
labath added a comment to D43345: [LLDB] Initial version of PPC64 InstEmulation.

Some nits, but nothing major from me.

Fri, Feb 16, 12:50 AM
labath accepted D43344: [LLDB][PPC64] Fixed next blocked forever at same line.
Fri, Feb 16, 12:02 AM

Thu, Feb 15

labath added a comment to D42468: [lldb][PPC64] Fixed vector and struct return value.

It doesn't look like there is. Do you need someone to check this in?

Thu, Feb 15, 7:36 AM
labath committed rL325250: @skipIfRemote TestTargetXMLArch.
@skipIfRemote TestTargetXMLArch
Thu, Feb 15, 7:26 AM
labath created D43335: [dosep] Run tests in a more parallel fashion.
Thu, Feb 15, 7:06 AM
labath added a comment to D43330: [gtest] Add PrintTo overload for StringRef..

This is unfortunate, but it's kind of a consequence of how googletest implements it's universal printers. The reason why I don't think it's a good idea is that this is not a problem specific to StringRef, as any class can come out as giberrish if the user forgets to include the PrintTo function. I am not adamant about this, but I think the bar for modifying gtest code should be higher than this (it's trivial to add the right include if you see you're values aren't coming out right).

Thu, Feb 15, 5:02 AM
labath created D43333: Add SBDebugger::GetBuildConfiguration and use it to skip an XML test.
Thu, Feb 15, 4:15 AM
labath updated subscribers of D43330: [gtest] Add PrintTo overload for StringRef..

This code does not have to live directly inside gtest. As such, maybe a better place for it would be somewhere under include/llvm/Testing ? (We already have PrintTo functions for Error and Expected there..)

Thu, Feb 15, 3:02 AM

Wed, Feb 14

labath committed rL325137: Remove vestigial remnants of the test crash info hook.
Remove vestigial remnants of the test crash info hook
Wed, Feb 14, 8:12 AM
labath created D43292: [dotest] Add ability to skip tests based on build configuration.
Wed, Feb 14, 8:07 AM
labath updated the diff for D43067: Implement equal_range for the DWARF v5 accelerator table.

Make base Entry destructor protected and nonvirtual (the class is not owned polymorphically right now).

Wed, Feb 14, 5:58 AM
labath added inline comments to D42955: Make Module::GetSectionList output consistent.
Wed, Feb 14, 5:49 AM
labath updated the diff for D43285: [CodeGen] Refactor AppleAccelTable.

delete copy constructors of AccelTableBase and move one comment a bit.

Wed, Feb 14, 5:47 AM
labath added a comment to D43285: [CodeGen] Refactor AppleAccelTable.

I've also put up D43286 (which is still WIP, but nearing completion...) so you can see the context of this change.

Wed, Feb 14, 5:28 AM
labath abandoned D42420: WIP: DO NOT SUBMIT: .debug_names generation.

This is now superseeded by D43286.

Wed, Feb 14, 5:23 AM
labath created D43286: WIP: .debug_names generation.
Wed, Feb 14, 5:21 AM
labath created D43285: [CodeGen] Refactor AppleAccelTable.
Wed, Feb 14, 5:17 AM
labath reopened D42740: Implement a case-folding version of DJB hash.
Wed, Feb 14, 3:51 AM
labath committed rL325115: Revert r325107 (case folding DJB hash) and subsequent build fix.
Revert r325107 (case folding DJB hash) and subsequent build fix
Wed, Feb 14, 3:09 AM
labath committed rL325109: Fix build broken by r325107.
Fix build broken by r325107
Wed, Feb 14, 2:27 AM
labath committed rL325107: Implement a case-folding version of DJB hash.
Implement a case-folding version of DJB hash
Wed, Feb 14, 2:07 AM
labath closed D42740: Implement a case-folding version of DJB hash.
Wed, Feb 14, 2:07 AM

Mon, Feb 12

labath updated the diff for D43067: Implement equal_range for the DWARF v5 accelerator table.

Make the base class copy/move operations protected and remote the hand-coded implementations from the derived class.

Mon, Feb 12, 3:48 AM
labath added a comment to D43048: [lldb-test/WIP] Allow a way to test autocompletion.

(Btw, if you're looking for things to FileCheck-ify, I think the stuff under lldb/unittests/UnwindAssembly is a prime candidate and has a much higher bang/buck ratio.)

Mon, Feb 12, 3:36 AM
labath added inline comments to D43096: [lit] Update how clang and other binaries are found in per-configuration directories.
Mon, Feb 12, 3:03 AM
labath added a comment to D43099: Make LLDB's clang module cache path customizable.

There are several components in the "command line parsing"

  1. argparse: parsing a single string into the individual words (processing quotes, backslashes, etc.)
  2. resolving the "command" part of the command line (the "frame variable" part of "frame variable --raw-output FOO")
  3. resolving the arguments of a command (the "--raw-output FOO" part)
Mon, Feb 12, 2:31 AM
labath committed rL324870: Skip TestTargetXMLArch on non-darwin OSs.
Skip TestTargetXMLArch on non-darwin OSs
Mon, Feb 12, 1:47 AM

Fri, Feb 9

labath added inline comments to D43073: [CodeGen] Optimize AccelTable.
Fri, Feb 9, 4:24 AM
labath committed rL324730: Fix some warnings in SymbolFilePDB.cpp.
Fix some warnings in SymbolFilePDB.cpp
Fri, Feb 9, 3:40 AM
labath added inline comments to D43067: Implement equal_range for the DWARF v5 accelerator table.
Fri, Feb 9, 3:15 AM
labath updated the diff for D43067: Implement equal_range for the DWARF v5 accelerator table.

Resolve David's comments (apart from one, which I have a question on).

Fri, Feb 9, 3:14 AM
labath updated the diff for D42740: Implement a case-folding version of DJB hash.

Modify the code generator script to produce patterns suggested by joerg. This
will make the function return quickly for lower-valued characters, which are
probably more likely to appear in input.

Fri, Feb 9, 2:52 AM
labath committed rL324724: [CodeGen] Optimize AccelTable.
[CodeGen] Optimize AccelTable
Fri, Feb 9, 2:10 AM
labath closed D43073: [CodeGen] Optimize AccelTable.
Fri, Feb 9, 2:10 AM
labath committed rL324722: llgs-test: Parse and store register info recieved from lldb-server.
llgs-test: Parse and store register info recieved from lldb-server
Fri, Feb 9, 1:42 AM
labath closed D43076: llgs-test: Parse and store register info recieved from lldb-server.
Fri, Feb 9, 1:42 AM

Thu, Feb 8

labath added inline comments to D42740: Implement a case-folding version of DJB hash.
Thu, Feb 8, 10:00 AM
labath updated the diff for D43067: Implement equal_range for the DWARF v5 accelerator table.

Fix issues pointed out by Jonas + clang-format.

Thu, Feb 8, 8:52 AM
labath created D43076: llgs-test: Parse and store register info recieved from lldb-server.
Thu, Feb 8, 8:38 AM
labath created D43073: [CodeGen] Optimize AccelTable.
Thu, Feb 8, 7:34 AM
labath added a comment to D43048: [lldb-test/WIP] Allow a way to test autocompletion.

The change seems fine to me, and I don't really have anything to add to the things that were said already.

Thu, Feb 8, 3:59 AM
labath added inline comments to D43067: Implement equal_range for the DWARF v5 accelerator table.
Thu, Feb 8, 3:56 AM
labath added a comment to D42740: Implement a case-folding version of DJB hash.

I believe I have resolved all outstanding issues. If there's anything else you'd like me to do, let me know (this is mainly addressed at @probinson, as his initial NACK is keeping this review red).

Thu, Feb 8, 3:53 AM
labath created D43067: Implement equal_range for the DWARF v5 accelerator table.
Thu, Feb 8, 3:46 AM
labath committed rL324590: Rewrite the flaky test_restart_bug test in a more deterministic way.
Rewrite the flaky test_restart_bug test in a more deterministic way
Thu, Feb 8, 2:41 AM
This revision was not accepted when it landed; it landed in state Needs Review.
Thu, Feb 8, 2:41 AM
labath added inline comments to D42994: Only throw -fPIC when building a shared library.
Thu, Feb 8, 2:03 AM
labath requested changes to D43059: Recognize MSVC style mangling in CPlusPlusLanguage::IsCPPMangledName.

That seems extremely dodgy. Surely whether something is a mangled name does not depend on the compiler we built lldb with. I am aware that we have many places with _Z hardcoded, but this seems to be making the situation only worse.

Thu, Feb 8, 1:47 AM

Wed, Feb 7

labath added inline comments to D43024: [test] Enable test category for inline tests..
Wed, Feb 7, 8:16 AM
labath added a comment to D42845: Add an option 'allow-all-hosts' to permit lldb debugging inside a Docker container.

I'm confused here. Can you share the exact commands you use to setup the debug session?

Wed, Feb 7, 6:23 AM · Restricted Project
labath added inline comments to D42082: Add DWARF for discriminated unions.
Wed, Feb 7, 3:22 AM · debug-info
labath committed rL324474: [BinaryFormat] Remove dangling declaration of DiscriminantString.
[BinaryFormat] Remove dangling declaration of DiscriminantString
Wed, Feb 7, 3:21 AM
labath committed rL324472: Remove function DW_DSC_value_to_name.
Remove function DW_DSC_value_to_name
Wed, Feb 7, 3:15 AM
labath added a comment to D42994: Only throw -fPIC when building a shared library.

In the future when you upload diffs can you include context? (i.e. git diff -U999999). It's nice to be able to see the surrounding code when I'm looking at a diff.

Is there ever a case where you would want to build a shared library without -fPIC? I'm wondering if we should just update the common Makefile.rules and if DYLIB_NAME is set (or something else indicating that this is a shared library), then we put the logic about -fPIC in that common file. Then people wouldn't have to remember to get this right in the future.

Wed, Feb 7, 3:13 AM
labath accepted D42990: Add CMAKE_CFG_INTDIR as part of the include path for the default test compiler. .

looks good.

Wed, Feb 7, 2:21 AM

Tue, Feb 6

labath created D42959: Rewrite the flaky test_restart_bug test in a more deterministic way.
Tue, Feb 6, 6:09 AM
labath added inline comments to D42955: Make Module::GetSectionList output consistent.
Tue, Feb 6, 4:54 AM
labath created D42955: Make Module::GetSectionList output consistent.
Tue, Feb 6, 4:50 AM
labath accepted D42950: LLDB fails to build.

Looks good, however I should point out that the state of the lldb+python3+linux combo is unknown..

Tue, Feb 6, 2:53 AM
labath added a comment to D42914: Rewrite TestTargetSymbolsBuildidCase to be more focused.

Heh, no good deed goes unpunished, right.. :)

Tue, Feb 6, 2:48 AM
labath added a comment to D42931: ELF GetSectionHeaderInfo: process SHT_NOTE even without .shstrtab.

As I mentioned in the patch above, I don't think it's worth it trying to tiptoe here, as for 99% of files (basically, anything that is not yaml2obj's output, I think), we will end up reading the whole file anyway. It just increases the number of things that can go wrong.

Tue, Feb 6, 2:31 AM
labath accepted D42763: Build each testcase variant in its own subdirectory and remove the srcdir lock file.

This looks much better. Thanks.

Tue, Feb 6, 2:21 AM
labath accepted D42939: More correct handling of error cases C++ name parser.

Being more resilient when handling demangler outputs seems like a good thing, but I think it is important to understand what made the demangler produce that output in the first place, to make sure we aren't missing anything.

Tue, Feb 6, 1:59 AM

Mon, Feb 5

labath committed rL324256: Add a comment explaining how the input for….
Add a comment explaining how the input for…
Mon, Feb 5, 10:05 AM
labath committed rL324254: Fix parsing of object files with "early" section headers.
Fix parsing of object files with "early" section headers
Mon, Feb 5, 9:27 AM
labath created D42914: Rewrite TestTargetSymbolsBuildidCase to be more focused.
Mon, Feb 5, 9:23 AM
labath accepted D42912: Sync PlatformNetBSD.cpp with Linux.
Mon, Feb 5, 9:00 AM
labath added a comment to D42908: Fix Bug 36237 - TestTargetSymbolsSepDebugSymlink fails when run remotely.

> I think that a more reasonable behavior here would be to copy the file contents instead.

Mon, Feb 5, 7:47 AM
labath committed rL324236: Skip TestTargetSymbolsSepDebugSymlink on remote targets.
Skip TestTargetSymbolsSepDebugSymlink on remote targets
Mon, Feb 5, 6:09 AM
labath committed rL324230: TestLinuxCore -- add a check for thread name.
TestLinuxCore -- add a check for thread name
Mon, Feb 5, 4:36 AM
labath added a comment to D42870: Correct recognition of NetBSD images.

Extending lldb-test's dumpModules() to also dump out the module's triple sounds like a reasonable thing to do (I am assuming that the Module class will do something reasonable when given an object file with no sections, like a core file -- if not we could make a separate command for dumping this out).

Mon, Feb 5, 4:33 AM