aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorBonfaceKilz2021-06-17 10:20:39 +0300
committerzsloan2021-06-17 18:51:27 +0000
commit4a6536e7a9b5830a91e17e9171ef699b68a8ddb5 (patch)
treeca7e9372c49b5ecfa9ed7c65f4ad751a30374d79
parentde645c34b18e14b0661b50410072f2f0f08355f2 (diff)
downloadgenenetwork2-4a6536e7a9b5830a91e17e9171ef699b68a8ddb5.tar.gz
doc: rpy2-performance: Add a section exploring possible solutions
-rw-r--r--doc/rpy2-performance.org9
1 files changed, 9 insertions, 0 deletions
diff --git a/doc/rpy2-performance.org b/doc/rpy2-performance.org
index 177c427e..8f917ca0 100644
--- a/doc/rpy2-performance.org
+++ b/doc/rpy2-performance.org
@@ -171,3 +171,12 @@ from locust import HttpUser, task, between
self.client.get("/test")
#+end_src
+
+** A possible solution
+
+From this [[https://github.com/rpy2/rpy2/issues/809#issuecomment-845923975][comment]], a possible reason for the above traces, is that
+from Flask's end, a [[https://tldp.org/LDP/lpg/node20.html][SIGPIPE]] is somehow generated by our Python
+code. However, at this particular point, the R thread just happens to
+be running, and R can't handle this correctly. This seems to have been
+fixed in this [[https://github.com/rpy2/rpy2/pull/810][PR]] with a this [[https://github.com/rpy2/rpy2/issues/809#issuecomment-851618215][explanation]]. On our end, to have these
+changes, we have to update our python-rpy2 version.