Age | Commit message (Collapse) | Author |
|
|
|
|
|
Also removed a TODO suggesting caching the precheck results. Tests
showed this would save about 0.1 sec on the total test run time on a
relatively modern system, an unnoticeable gain at the cost of longer and
more complicated code. There would also be a danger that a cached test
result would be inappropriately returned, such as when other test
dependencies (like environment variables) are different or when the
precheck causes side effects (like filesystem changes).
|
|
@MarcelRaad noted that `test1399` causes infinite loop on MinGW.
Looking into this, seems like it is related to how Windows handles
CRLF. See https://github.com/curl/curl/commit/9e093f by @mback2k.
Removing `test1399` as it's identical to `test1326` then with such a
fix.
Test 1399 was broughy by commit 862b02f8947039e
Closes #1478
|
|
Include the test number in the names of files written out by tests to
reduce the chance of accidental duplication and to make it more clear
which test is associated with which file.
|
|
This is already added by the test suite; it's not clear why all these
tests had it, unless it's cargo-culting.
|
|
|
|
... using the docs/cmdline-opts/gen.pl script, so that we get all the
command line option documentation from the same source.
The generation of the list has to be done manually and pasted into the
source code.
Closes #1465
|
|
|
|
Otherwise, subsequent uses of stunnel overwrite the configuration file
of previous invocations so they can no longer be inspected.
|
|
alarm's argument is unsigned.
|
|
long is 32 bits while size_t is 64 bits on MinGW-w64, so
typecheck-gcc.h complains when using size_t for a long option.
Also, curl_socket_t is unsigned long long rather than int.
|
|
|
|
Avoids narrowing conversion warnings because rlim_t is usually
unsigned long.
Closes https://github.com/curl/curl/pull/1469
|
|
Also, document that numbered datacheck sections are possible.
|
|
Avoid casting string literals to non-const char *.
|
|
|
|
calloc and ai_addrlen expect different (usually unsigned) types.
|
|
Suggested-by: Dan Fandrich
|
|
... instead of numerical order.
Closes #1466
|
|
|
|
assign string literals to const char * instead of char * in order to
avoid a lot of these warnings:
cast from 'const char *' to 'char *' drops const qualifier
[-Wcast-qual]
|
|
Windows does not allow setting the locale with environment variables (as
the test attempted to do), so the test failed when run with a user
locale that has a comma as radixchar. Changed the test to call
setlocale() explicitly to ensure that a known working locale is set even
on Windows.
|
|
|
|
Previous TODO wanting to write in chunks. We should support writing more
at once since some TELNET servers may respond immediately upon first
byte written such as WHOIS servers.
Closes #1389
|
|
Support for _MPRINTF_REPLACE in mprintf.h was removed in
55452ebdff47f98bf3cc383f1dfc3623fcaefefd, replaced with curl_printf.h.
|
|
It was a wrong assumption that it could do that!
|
|
This fixes the following clang warning:
getpart.c:201:17: warning: cast from function call of type 'CURLcode'
to non-matching type 'int' [-Wbad-function-cast]
|
|
This fixes missing-variable-declarations warnings when building with
clang.
|
|
|
|
|
|
It's dangerous to continue to run the test when a memory alloc fails.
|
|
|
|
|
|
|
|
Test command 'time curl http://localhost/80GB -so /dev/null' on a Debian
Linux.
Before (middle performing run out 9):
real 0m28.078s
user 0m11.240s
sys 0m12.876s
After (middle performing run out 9)
real 0m26.356s (93.9%)
user 0m5.324s (47.4%)
sys 0m8.368s (65.0%)
Also, doing SFTP over a 200 millsecond latency link is now about 6 times
faster.
Closes #1446
|
|
The 'list element' struct now has to be within the data that is being
added to the list. Removes 16.6% (tiny) mallocs from a simple HTTP
transfer. (96 => 80)
Also removed return codes since the llist functions can't fail now.
Test 1300 updated accordingly.
Closes #1435
|
|
|
|
A few random typos, and minor whitespace cleanups, found in comments
while reading code.
Closes #1423
|
|
This makes test 1135 pass with CRLF checkouts.
Ref: https://github.com/curl/curl/pull/1344#issuecomment-289243166
Closes https://github.com/curl/curl/pull/1422
|
|
MinGW-w64 complains:
warning: conversion to 'long int' from 'time_t {aka long long int}' may
alter its value [-Wconversion]
Fix this by using the correct type.
|
|
Ref: https://github.com/curl/curl/issues/1408
Closes https://github.com/curl/curl/pull/1412
|
|
Follow-up to aa573c3c55cda72ec5ef677d87f6f46a53385f0c
Ref: https://github.com/curl/curl/pull/1406
|
|
If the existing timer is still in there but has expired, the new timer
should be added.
Reported-by: Rainer Canavan
Bug: https://curl.haxx.se/mail/lib-2017-04/0030.html
Closes #1407
|
|
|
|
... the sizes and the formatting strings are what's really important and
avoids problems with int64_t vs "long long".
Bug: https://curl.haxx.se/mail/lib-2017-04/0019.html
|
|
This checks the new behavior of Curl_splaygetbest, so that the smallest
node not larger than the key is removed, and FIFO behavior is kept even
when there are multiple nodes with the same key.
Closes #1358
|
|
Multi handles repeatedly invert the queue of pending easy handles when
used with CURLMOPT_MAX_TOTAL_CONNECTIONS. This is caused by a multistep
process involving Curl_splaygetbest and violates the FIFO property of
the multi handle.
This patch fixes this issue by redefining the "best" node in the
context of timeouts as the "smallest not larger than now", and
implementing the necessary data structure modifications to do this
effectively, namely:
- splay nodes with the same key are now stored in a doubly-linked
circular list instead of a non-circular one to enable O(1)
insertion to the tail of the list
- Curl_splayinsert inserts nodes with the same key to the tail of
the same list
- in case of multiple nodes with the same key, the one on the head of
the list gets selected
|
|
No longer allocate the curl_llist head struct for lists separately.
Removes 17 (15%) tiny allocations in a normal "curl localhost" invoke.
closes #1381
|
|
system.h is aimed to replace curlbuild.h at a later point in time when
we feel confident system.h works sufficiently well.
curl/system.h is currently used in parallel with curl/curlbuild.h
curl/system.h determines a data sizes, data types and include file
status based on available preprocessor defines instead of getting
generated at build-time. This, in order to avoid relying on a build-time
generated file that makes it complicated to do 32 and 64 bit bields from
the same installed set of headers.
Test 1541 verifies that system.h comes to the same conclusion that
curlbuild.h offers.
Closes #1373
|