| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
It's best to pick a more local owner.
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
| Code-Review | +1 |
Nice find, lgtm with comments (not an owner).
if (this->flags().is_reparse()) {Please add a brief comment. (And consider renaming the flags parameter to regexp_error).
// TODO(jgruber): If already validated in the preparser, skip validation inIs this comment addressed? Note it's about the preparser, not reparsing.
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
| Code-Review | +1 |
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
// TODO(jgruber): If already validated in the preparser, skip validation inIs this comment addressed? Note it's about the preparser, not reparsing.
Actually you're right; this should probably be is_lazy_compile instead.
I presume this isn't the only place where regexp is checked for syntax correctness wrt v8 sandboxing?
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
| Code-Review | +0 |
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
// TODO(jgruber): If already validated in the preparser, skip validation inToon VerwaestIs this comment addressed? Note it's about the preparser, not reparsing.
Actually you're right; this should probably be is_lazy_compile instead.
I presume this isn't the only place where regexp is checked for syntax correctness wrt v8 sandboxing?
I suppose in your case you _do_ want is_reparse since that's what we use for lazy source positions. But we should probably extend that flag to include what's currently covered by is_lazy_compile so we inlude all forms of reparses.
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
// TODO(jgruber): If already validated in the preparser, skip validation inToon VerwaestIs this comment addressed? Note it's about the preparser, not reparsing.
Actually you're right; this should probably be is_lazy_compile instead.
I presume this isn't the only place where regexp is checked for syntax correctness wrt v8 sandboxing?
I think that we can bail out in both cases. I’ve verified that `is_lazy_compile` implies a pre-parsed RegExp (https://chromium.googlesource.com/c/v8/v8/+/7665887), while `is_reparse` correctly identifies the re-parse phase triggered source position collection.
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
// TODO(jgruber): If already validated in the preparser, skip validation inToon VerwaestIs this comment addressed? Note it's about the preparser, not reparsing.
Marco VitaleActually you're right; this should probably be is_lazy_compile instead.
I presume this isn't the only place where regexp is checked for syntax correctness wrt v8 sandboxing?
I think that we can bail out in both cases. I’ve verified that `is_lazy_compile` implies a pre-parsed RegExp (https://chromium.googlesource.com/c/v8/v8/+/7665887), while `is_reparse` correctly identifies the re-parse phase triggered source position collection.
I think so, what I wanted to say was that a is_lazy_compile (should) imply a reparse? So we can still use is_reparse if we also correctly set the flag where we set is_lazy_compile
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |