Merge lp:~akopytov/percona-xtrabackup/bug1273207-2.1 into lp:percona-xtrabackup/2.1
Proposed by
Alexey Kopytov
Status: | Merged | ||||
---|---|---|---|---|---|
Approved by: | Sergei Glushchenko | ||||
Approved revision: | no longer in the source branch. | ||||
Merged at revision: | 731 | ||||
Proposed branch: | lp:~akopytov/percona-xtrabackup/bug1273207-2.1 | ||||
Merge into: | lp:percona-xtrabackup/2.1 | ||||
Diff against target: |
180 lines (+54/-29) 6 files modified
src/common.h (+24/-0) src/xbcrypt.c (+3/-3) src/xbcrypt.h (+1/-2) src/xbcrypt_read.c (+8/-8) src/xbstream_read.c (+6/-16) test/t/bug1273207.sh (+12/-0) |
||||
To merge this branch: | bzr merge lp:~akopytov/percona-xtrabackup/bug1273207-2.1 | ||||
Related bugs: |
|
Reviewer | Review Type | Date Requested | Status |
---|---|---|---|
Sergei Glushchenko (community) | g2 | Approve | |
Review via email: mp+208729@code.launchpad.net |
Description of the change
Bug #1273207: SIGPIPE handling of xbstream
Replaced my_read(..., MYF(MY_FULL_IO)) with a new wrapper around
my_read() which does the same, but bails out on EOF and errors returning
the total number of successfully read bytes. This allows to detect a
closed pipe and abort instead of waiting infinitely for more data.
http://
To post a comment you must log in.
xb_crypt_ read_callback returns ssize_t while xb_read_full returns size_t. Negative values do not make sense for both. read_callback definition changed by this patch anyway, maybe would be better to return size_t from read_callback? It would also allow to get rid of a number of type conversions in xbcrypt_read.c
Is there any reason to conversion size_t -> ssize_t in xb_read_full?
xb_crypt_
xb_crypt_