labath added a comment.
In D76188#1924840 <https://reviews.llvm.org/D76188#1924840>, @mib wrote:
> @labath Do you object if I land this patch to revert our CIs to green and
> work on matching the symbol by address instead of name on a separate one ?
I think this is fine -- a vector is definitely cleaner than two member
variables, and I can imagine there being a situation where you'd need to match
multiple symbols that aren't just aliases of one another.
I just have one question about the regex handling.
> Also do you know what's the "default" symbol I should look for with the
> assert frame recognizer ?
`raise` is the canonical name. __GI_raise is an internal name you'll only get
if you have debug info (.symtab) for libc installed. `gsignal` is an alias to
`raise` only due to a "historical mistake". I am not sure if anyone is actually
using it (in fact, I didn't know the thing exists until yesterday).
================
Comment at: lldb/source/Commands/CommandObjectFrame.cpp:880
auto func =
- RegularExpressionSP(new RegularExpression(m_options.m_function));
+ RegularExpressionSP(new
RegularExpression(m_options.m_symbols.front()));
StackFrameRecognizerManager::AddRecognizer(recognizer_sp, module, func);
----------------
mib wrote:
> labath wrote:
> > Is there something which ensure that m_symbols contains at least one
> > element here? (i.e., that we do not silently drop the extra symbols
> > arguments)
> Few lines above, we make sure the symbols list is not empty.
Not empty is one thing, but what about the case when it contains 2 or more
elements (`frame recognizer add --name foo --name bar --regex`)? Will that
produce some kind of an error, or will it just ignore the second regex ?
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D76188/new/
https://reviews.llvm.org/D76188
_______________________________________________
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
