Commit 1db86b0
committed
requirements from replacements do not come from the lock file
I think the core of the problem is that 725420e was a mistake. Before that PR we used `=` constraints to imply that it came from lock file so when we saw a `=` constraints being constructed for a replace it should use the new system for marking something as coming from lock file. But I don't think that is the semantics that an `=` constraint implies in this context.1 parent 6644828 commit 1db86b0
1 file changed
+1
-2
lines changed| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
1274 | 1274 | | |
1275 | 1275 | | |
1276 | 1276 | | |
1277 | | - | |
1278 | | - | |
| 1277 | + | |
1279 | 1278 | | |
1280 | 1279 | | |
1281 | 1280 | | |
| |||
0 commit comments