[Lldb-commits] [PATCH] D98482: [lldb] Support for multiprocess extension [WIP]

[Michał Górny via Phabricator via lldb-commits]

mgorny added a comment.

In D98482#2631817 <https://reviews.llvm.org/D98482#2631817>, @labath wrote:

> In D98482#2626237 <https://reviews.llvm.org/D98482#2626237>, @mgorny wrote:
>
>> Create a new `GDBRemoteError` class to pass gdb-remote `$E` codes through cleanly.
>
> The error codes we use right now are completely meaningless, so don't bother preserving them. I don't think we should be introducing a separate error class on their account, and I'm particularly unhappy about how this class has insinuated itself into the Status object.

Well, I found the ability to use different codes useful for debugging tests but I guess it doesn't matter much. So a plain `StringError` then?



================
Comment at: lldb/source/Utility/StringExtractorGDBRemote.cpp:625
+      uint64_t prev_index = m_index;
+      pid = GetHexMaxU64(false, 0);
+      if (m_index == prev_index || m_index == UINT64_MAX) {
----------------
labath wrote:
> This should be something like `view.consumeInteger(16, pid)`. At that point you can stop keeping m_index in sync, and just update it at the very end (maybe even via llvm::scope_exit).
Do you have some specific method of updating it mind? Comparing size before and after?


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D98482/new/

https://reviews.llvm.org/D98482