Fixes for S3 uploading bugs.#1984
Conversation
First bug prevented the finalisation of the upload in AWS. This was caused by not removing two header entries which caused the signature to be miscalculated. Second was caused by failure to re-check the error state after a redirection.
hfile_s3.c
Outdated
| ret = -1; | ||
| cret = curl_easy_getinfo(fp->curl, CURLINFO_RESPONSE_CODE, &response_code); | ||
|
|
||
| fprintf(stderr, "*** GOT HERE %ld***\n", response_code); |
There was a problem hiding this comment.
Looks like this got left in by mistake...
hfile_s3.c
Outdated
There was a problem hiding this comment.
I think we should bail out here it ret != 0. If initialise_upload() has failed, the http request may not even have been made so there's not much point in trying to get the response code. It might also simplify the core below by removing a layer of if () statement.
Similarly, if initialise_upload() gets called again below then we should also bail out there if it fails.
hfile_s3.c
Outdated
| } | ||
|
|
||
| // reget the response code (may not have changed) | ||
| cret = curl_easy_getinfo(fp->curl, CURLINFO_RESPONSE_CODE, &response_code); |
There was a problem hiding this comment.
It probably isn't too harmful, but ideally we'd only do this if we've called initialise_upload() again (and it worked).
hfile_s3.c
Outdated
There was a problem hiding this comment.
While testing this, I got a 307 Temporary Redirect response. According to the docs this can happen "while the DNS server is being updated", probably by trying to access a bucket that's only just been created. Handling it like 301 Moved Permanently seems to work (i.e. just check for both response codes here, and also in s3_open_read()).
2 participants
First bug prevented the finalisation of the upload in AWS. This was caused by not removing two header entries which caused the signature to be miscalculated.
Second was caused by failure to re-check the error state after a redirection.
Hopefully this should fix #1979.