[Lldb-commits] [lldb] [lldb][progress] Add discrete boolean flag to progress reports (PR #69516)

[Greg Clayton via lldb-commits]

================
@@ -2225,7 +2225,8 @@ void ObjectFileMachO::ParseSymtab(Symtab &symtab) {
   const char *file_name = file.GetFilename().AsCString("<Unknown>");
   LLDB_SCOPED_TIMERF("ObjectFileMachO::ParseSymtab () module = %s", file_name);
   LLDB_LOG(log, "Parsing symbol table for {0}", file_name);
-  Progress progress(llvm::formatv("Parsing symbol table for {0}", file_name));
+  Progress progress(llvm::formatv("Parsing symbol table for {0}", file_name),
+                    Progress::ProgressReportType::eAggregateProgressReport);
----------------
clayborg wrote:

So we do want all of the reports to be sent, this seems like a GUI / display issue that can be solved with category names. For you guys the DWARF isn't a big contributor and we want to always see those since on linux a lot of DWARF is not indexed and that is a major contributor to our startup times. Symbol table parsing takes time on mac and linux as well. So everyone is going to classify what progress is interesting in their own way. With a single enum for what Apple users might consider "aggregate" doesn't match up with what Linux or other users consider aggregate and we are now going to be hard coded this into the progress objectsd and that is what is worrying me about the current approach. 

The user interface is what needs to figure out how to display these things efficiently and the category string seems to allow user interfaces to do more grouping in more efficient ways. The boolean of is_aggregate is a bit too coarse. With the category strings like "Parsing symbol tables" or "Manually Indexing DWARF" or "Loading DWARF accelerator tables", it would allow spammy stuff to be noticed on the, and then instead of displaying individual entries, it would display a single entry with the category name and a spinning progress GUI. Of course you could hard code things based on the category name all the time in Xcode so that "Parsing symbol tables" and "Loading DWARF accelerator tables" would always be grouped if desired to keep the noise down.

If no one else agrees with me here, please let me know

https://github.com/llvm/llvm-project/pull/69516