Commit d20c7d0
committed
fix webpack resolver state leakage
I found a case where we were accidentally leaking state between WebpackModuleRequests. This is hard to avoid since webpack *really* pushes us toward using mutation on the shared ResolveData object.
I solved it by making our mutations happen only at the precise moment when we're letting a request bubble back out to webpack.
This is expected to fail tests however because this bug was masking another bug. Specifically, we have several test apps that use things like the `page-title` helper. That helper is (1) used with ambiguous syntax, (2) published in a v2 addon. It turns out we cannot resolve the helper due to #1686, and this was only "working" by accident since the failed attempt at resolving the helper got leaked back to webpack and webpack *does* understand package.json exports and would find it.1 parent 014abac commit d20c7d0
1 file changed
+22
-0
lines changed| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
176 | 176 | | |
177 | 177 | | |
178 | 178 | | |
| 179 | + | |
| 180 | + | |
| 181 | + | |
| 182 | + | |
| 183 | + | |
| 184 | + | |
| 185 | + | |
| 186 | + | |
| 187 | + | |
| 188 | + | |
| 189 | + | |
| 190 | + | |
| 191 | + | |
| 192 | + | |
| 193 | + | |
| 194 | + | |
| 195 | + | |
| 196 | + | |
| 197 | + | |
| 198 | + | |
| 199 | + | |
| 200 | + | |
179 | 201 | | |
180 | 202 | | |
181 | 203 | | |
| |||
0 commit comments