loadMetaInfos tried to use the documents checksum before loading. The empty checksum results in it always failing to load metainfo and returning false, thus not using any of the stored metainfo.
BUG: 384087
loadMetaInfos tried to use the documents checksum before loading. The empty checksum results in it always failing to load metainfo and returning false, thus not using any of the stored metainfo.
BUG: 384087
No Linters Available |
No Unit Test Coverage |
Buildable 3981 | |
Build 3999: arc lint + arc unit |
I'm assuming (based on the previous docstring) that loadMetaInfos at some point also opened the url and the old doc->openUrl(u); was a fallback in case it it did not not (feature disabled, first time to open url, etc). That seems to have been removed at some point, so now it tried to load metainfo based on an empty document (and empty checksum), which of course always failed. I can't find any drawbacks to always loading the file in one place, but please tell me if there was a reason for the two different methods!
Hmm, actually the readSessionConfig call in the loadMetaInfos should do an openUrl with the URL stored in the meta data.
But yes, it can't work like it is.
I am ok with the patch, beside that one should use the SkipUrl flag for the readSessionConfig to avoid double loading.
Thank you for the quick review! Could you also land this patch for me? I don't have commit rights yet.
Yes indeed! If you're open to ideas, might I suggest any of the following?