projects
/
m6w6
/
ext-http
/ blobdiff
commit
grep
author
committer
pickaxe
?
search:
re
summary
|
shortlog
|
log
|
commit
|
commitdiff
|
tree
|
github
raw
|
inline
| side by side
# you better don't look inside
[m6w6/ext-http]
/
KnownIssues.txt
diff --git
a/KnownIssues.txt
b/KnownIssues.txt
index db3b288059d8b3777baa349230724e3093fc86be..f4e945008aa2a6393f40225e73df0a4e180b1c2d 100644
(file)
--- a/
KnownIssues.txt
+++ b/
KnownIssues.txt
@@
-7,13
+7,18
@@
HttpResponse class is only available for PHP >= 5.1
Not all places where files are handled check for open_basedir and/or safe_mode.
Throttling with the FastCGI SAPI may behave unexpected, because libfcgi seems
Not all places where files are handled check for open_basedir and/or safe_mode.
Throttling with the FastCGI SAPI may behave unexpected, because libfcgi seems
-to really write on flush only and every ~20kb.
+not to practically flush its buffers when calling sapi_flush().
+See also http://bugs.php.net/bug.php?id=34429
If you keep getting "SSL connect error" when trying to issue requests on
Windows, try another (newer) libeay32.dll/ssleay32.dll pair.
If you keep getting "SSL connect error" when trying to issue requests on
Windows, try another (newer) libeay32.dll/ssleay32.dll pair.
+Inflating compressed HttpRequest responses happens twice if libcurl
+was built with zlib support.
Internals:
- the request bodies created in http_request_pool_attach() are not
Internals:
- the request bodies created in http_request_pool_attach() are not
- destroyed in http_request_pool_detach(); may be a memory problem
- in long running scripts
+ destroyed in http_request_pool_detach() but on reset;
+ may be a memory problem in long running scripts which reuse one
+ request pool several times
+