Message321889
| Author | vstinner |
|---|---|
| Recipients | njs, pitrou, vstinner, yselivanov |
| Date | 2018-07-18.15:16:02 |
| SpamBayes Score | -1.0 |
| Marked as misclassified | Yes |
| Message-id | <1531926962.66.0.56676864532.issue34130@psf.upfronthosting.co.za> |
| In-reply-to |
| Content | |
|---|---|
I tried to modify the test_warn_on_full_buffer() code filling the socketpair buffer: * iterate on multiple chunk sizes: 2**16, 2**10, 2**5, 1 bytes * first try in non-blocking mode then retry with a timeout of 100 ms * I also tried to play with SO_SNDBUF and SO_RCVBUF socket options The problem is that I failed to fail a reliable way to fill the socketpair buffer. I will stick with Nathaniel's very simple idea: retry once if the test fails. I implemented exactly this idea in my PR 8327 (with additional minor changes to enhance the test). |
|
| History | |||
|---|---|---|---|
| Date | User | Action | Args |
| 2018-07-18 15:16:02 | vstinner | set | recipients: + vstinner, pitrou, njs, yselivanov |
| 2018-07-18 15:16:02 | vstinner | set | messageid: <1531926962.66.0.56676864532.issue34130@psf.upfronthosting.co.za> |
| 2018-07-18 15:16:02 | vstinner | link | issue34130 messages |
| 2018-07-18 15:16:02 | vstinner | create | |