Re: ??????need help on using libcurl

On Wed, 23 Oct 2019, ?? via curl-library wrote: I want to know if there is any option can set to not check the expire date of the certificate . Just like I can set CURLOPT_SSL_VERIFYHOST to not check the host name . No there isn't! It has been discussed in the past as a potential

Re: ARM test 323 failure

On Fri, 18 Oct 2019, Dan Fandrich via curl-library wrote: Secondly, I noticed a couple of test runs ended in segfaults (for example [3]; I thought there was another one but I can't find it now). That's a bit more worrisome, but it's going to be hard to produce without ready access to the

Re: Windows Aarch64 build

On Mon, 21 Oct 2019, Michał Janiszewski wrote: You can browse GitHub tags: * via the webui: https://github.com/janisozaur/curl-win-aarch64/blob/curl-7_66_0/curl.exe * for raw files for download: https://github.com/janisozaur/curl-win-aarch64/raw/curl-7_66_0/curl.exe * for automated release

Re: Windows Aarch64 build

On Sat, 19 Oct 2019, Michał Janiszewski wrote: I think I have addressed most of your suggestions. Almost! Do you have any suggestion on how my bot can find the curl version number of the package on page? "7.66.0" being the version number of the latest release. The file you provide isn't

Re: Looking for a HTTP test server which doesn't send Content-Length

On Sat, 19 Oct 2019, Andreas Falkenhahn via curl-library wrote: So unfortunately, my old test case HTTP server which sent HTTP responses without the Content-Length field seems to be gone. Does anybody happen to have a URL for a HTTP server which responds without setting Content-Length?

Re: ARM test 323 failure

On Sat, 19 Oct 2019, Jeffrey Walton wrote: The sad results is however that on this machine test 323 seems to work just fine even if I try to repeat it mant times! :-O The particular machine I used features a "AArch64 Processor rev 1 (aarch64)" in a board called "APM X-Gene Mustang". The

Re: ARM test 323 failure

On Sat, 19 Oct 2019, Jeffrey Walton wrote: The sad results is however that on this machine test 323 seems to work just fine even if I try to repeat it mant times! :-O The other ones to try is GCC117 or GCC118. They are roughly the same Aarch64 machines, but with CRC+Crypto. The Mustang

Re: file:// test URLs

On Fri, 18 Oct 2019, Marcel Raad via curl-library wrote: Virtually all file:// tests are failing. They use paths like file://localhost/%PWD%/log/test200.txt Is this format on purpose? Well, they worked so to that extent they're on purpose! =) But I will agree that it is a little sloppy to

Re: ARM test 323 failure

On Fri, 18 Oct 2019, Dan Fandrich via curl-library wrote: For one, test 323 consistently fails I got an account on cfarm, compiled curl on an aarch64 machine of there in order to reproduce and debug this case. The sad results is however that on this machine test 323 seems to work just

Re: ARM test 323 failure

On Fri, 18 Oct 2019, Jeffrey Walton wrote: GCC117 and GCC118 from the Compile Farm are ARMv8 machines with CRC and Crypto. If you don't have a CFarm account, you can signup at Neato. I've applied for an account, thanks! -- / daniel.haxx.se | Get the best commercial curl support there is

Re: ARM test 323 failure

On Fri, 18 Oct 2019, Dan Fandrich via curl-library wrote: I have a PR[1] to add an ARM build to Travis (ARM support is still in beta) and I've noticed a few problems. For one, test 323 consistently fails (see an example at [2]). curl is returning CURLE_GOT_NOTHING instead of the expected

Re: Sharing DNS cache between multi handles

On Fri, 18 Oct 2019, Denis Chapligin via curl-library wrote: "Note that when you use the multi interface, all easy handles added to the same multi handle will share DNS cache by default without using this option." But do several multi handles share same DNS cache or do they create their own

Re: Windows Aarch64 build

On Fri, 18 Oct 2019, Michał Janiszewski via curl-library wrote: I've prepared Windows Aarch64 build that you can find here: https://gist.github.com/janisozaur/584f5e04485b7d3a7b29b6fb4db3584b#file-curl-exe It is 7.66 release, built with WinSSL, using MSVC2019 and the CMake project. Is there any

Re: "curl_multi_unblock"

On Thu, 10 Oct 2019, Stefan Eissing wrote: In my curl use for ACME, I could make use of a curl_multi_abort(). By which I mean the abort of all contained easy handles and a subsequent unblock(). (I do not claim this to be *the* use case, just one which I can see for myself) Thanks. That's

"curl_multi_unblock"

Hi, The other day I mentioned the curl_easy_abort() idea on how to stop an easy transfer from another thread: https://github.com/curl/curl/wiki/curl_easy_abort Here's the similar idea but for "unblocking" a multi interface wait/poll call. I call it "curl_multi_unblock" as a working title.

Re: debug CURLOPT_POSTFIELDS url?

On Thu, 10 Oct 2019, ka sc via curl-library wrote: url_flags = "\"[\\\"INSERT INTO foo2 values(1,11)\\\"]\""; curl_easy_setopt(curl, CURLOPT_POSTFIELDS, url_flags.c_str()); Is there a way to see the output url so I can try that in commandline? The data you set with CURLOPT_POSTFIELDS is not

Re: Stoppable curl_easy_perform ?

On Tue, 8 Oct 2019, Christopher Head via curl-library wrote: Yes, that actually makes total sense. I hadn’t remembered that the easy interface is blocking, so really, they’re totally different ideas: an easy would have to be asynchronously aborted, while a multi would just be asynchronously

Re: [DKIM] Re: Stoppable curl_easy_perform ?

On Wed, 9 Oct 2019, Zakrzewski, Jakub via curl-library wrote: this is just a pass/fail of the "abort" verb, no? I guess it's not that new codes cost much, but does it buy anything? But even then curl_easy_abort(...) must return something that is not CURLE_OK. Since other codes do not seem

Freeze tomorrow

Hi, Yet again we reach this time in the cycle. Tomorrow we enter feature freeze and we will not merge "features" until the coming release - 7.67.0 - planned to happen on November 6. -- / daniel.haxx.se | Get the best commercial curl support there is - from me | Private

Re: Stoppable curl_easy_perform ?

On Mon, 7 Oct 2019, Dan Fandrich via curl-library wrote: curl_easy_stop(easy); Now it's time to bike shed :-) This name implies to me something like "pause" in that you might expect to be able to start it again later. I'm guessing once this is called, the transfer is toast, forever.

Re: Stoppable curl_easy_perform ?

On Mon, 7 Oct 2019, Christopher Head via curl-library wrote: How would this work with multi? First: not at all. I appreciate that you bring this up because I have been thinking about it and I want to offer something for that as well. Here's my thinking: In the multi interface case we

Stoppable curl_easy_perform ?

Hi all, A frequent request and question over the years has been how you stop curl_easy_perform() immediately from another thread. The answer of course has always been that there's no built-in way to do that but you have to make sure one of the callbacks return an error, which in the worst

Re: help me improve behavior on Windows?

On Sun, 6 Oct 2019, René Berber via curl-library wrote: There are implementations for Windows in * libevent (evutil.c: evutil_ersatz_socketpair()), and in * DBUS (dbus-sysdeps-win.c: _dbus_socketpair()). Both work fine under Windows (from W-95 to the present). They both work very similar to

Re: help me improve behavior on Windows?

On Sun, 6 Oct 2019, Patrick Monnerat via curl-library wrote: we already have a similar function: it is named inetsocketpair() and is located at lib/vtls/gskit.c line 526 (see comment before this line). In addition, inetsocketpair() checks for security that no other process connects to our

help me improve behavior on Windows?

Hi friends, In 7.66.0 we merged code into curl that makes it provide a socket for applications to wait for during the threaded name resolve phase, removing the need for polling for status and thus figuring out how often to poll or even in some cases end up busy-looping like crazy. That

Re: CURLOPT_HEADERFUNCTION and buffer modifications

On Thu, 3 Oct 2019, Clément Gregoire via curl-library wrote: I'm using CURLOPT_HEADERFUNCTION to read/parse headers, and it'd be useful for me if I could replace some characters with for example the '\0'. Typical usage would be for strtok. Is this allowed ? I know the header callback buffer

TeamViewer sponsors more AppVeyor power

Hi, I just want to announce that TeamViewer has graciously and most generously stepped up and now sponsors additional AppVeyor CI CPU power for the curl project. On the behalf of the curl projet, I thank TeamViewer and our other awesome sponsors. It is thanks to these companies that we can

Re: [RELEASE] curl 7.66.0

On Mon, 30 Sep 2019, Dennis Clarke via curl-library wrote: On an Oracle Solaris 10 server I was surprised to see : "asyn-thread.c", line 248: error: undefined symbol: AF_LOCAL It surprised me/us as well as the Solaris CI builds were supposed to catch problems like that. Turns out they

Re: Adding/removing handles to the multi_handle produces unexpected result of curl_multi_perform

On Mon, 30 Sep 2019, Denis Chapligin wrote: Thank you very much, it works as expected now. I think using curl_multi_info_read for that case should be documented, cause at the moment it is non-obvious, so i'll make documentation PR this week. Cool, thanks! -- / daniel.haxx.se | Get the

Re: Adding/removing handles to the multi_handle produces unexpected result of curl_multi_perform

On Fri, 27 Sep 2019, Denis Chapligin via curl-library wrote: This could be a reason, but how can i check the state of that handle and prove that? Will curl_multi_info_read return data for that handle, for example? Yes it will. It returns information about all added handles, even if they

Re: Managing application data fetched from DNS (eg for ESNI)

On Fri, 20 Sep 2019, Niall O'Reilly via curl-library wrote: Dedicate additional slots, perhaps per (data-type, origin) rather than per QTYPE, as some QTYPEs are re-used for different purposes using distinct QNAME prefixes. More thought seems needed here. I'm minded to prototype a solution

Re: Adding/removing handles to the multi_handle produces unexpected result of curl_multi_perform

On Fri, 27 Sep 2019, Denis Chapligin via curl-library wrote: //Here goes the main loop curl_multi_perform(multi_handle, _handles); while(running_handles>0){ curl_multi_wait(multi_handle) while(curl_multi_info_read(multi_handle) { if(CURLMSG_DONE) { //success

Re: Ignore SSH private key if password is provided

On Mon, 23 Sep 2019, Christian Schmitz via curl-library wrote: Could CURL be changed to ignore wrong key for ssh file if password is provided for SSH and password is available as authentication type? Sounds completely sensible to me! -- / daniel.haxx.se | Get the best commercial curl

Re: Get cookies for specific URL

On Tue, 24 Sep 2019, Hölzl, Dominik via curl-library wrote: get all cookies for a specific URL? No, we don't provide such an API. Yet. Maybe we should? Ideas on how it would look like? CURLINFO_COOKIELIST/Curl_cookie_list skips cookies without domain property as stated in the docs, so

Re: bypassing name resolve hangs

On Mon, 23 Sep 2019, Dan Fandrich via curl-library wrote: I wrote a test program that resolves a name found in /etc/hosts and it does 70K lookups per second. If you disable curl's threaded resolver and configure the system resolver to only look in /etc/hosts (and not to go to the network for

Re: surprising call of pop3_done() when doing http fuzzing

On Mon, 23 Sep 2019, Paul Dreik via curl-library wrote: https://github.com/curl/curl/blob/41db01a39f88d05f43344d0ea1d1b588b3441403/lib/doh.c#L242 the timeout_ms can become negative, and in that case I believe the doh should return an error (perhaps CURLE_OPERATION_TIMEDOUT?) It should

Re: surprising call of pop3_done() when doing http fuzzing

On Mon, 23 Sep 2019, Paul Dreik via curl-library wrote: The decoded contents of the test data means this is what happens: - set hostname to "A" - set doh url to "pop3:/tA" - start transferring Ah, this a bug but a pretty harmless one: The code:

Re: Crashes in inflate_fast

On Fri, 20 Sep 2019, Christian Schmitz via curl-library wrote: I rebuild the libraries and made sure we all use recent zlib 1.2.11 so everywhere data structures are the same. Anyone sees this, too? We have not received any such reports from others, no. Can you reproduce with the curl command

Re: HTTPS using my own TLS session

On Fri, 13 Sep 2019, David Woodhouse via curl-library wrote: So: what would it take to use curl for HTTP while basically abusing it from both sides? Not only do I need it to use my own underlying TLS connection, but I also need in some cases to make a CONNECT or even GET request which

Re: extending the fuzzers

On Tue, 17 Sep 2019, Paul Dreik via curl-library wrote: There is more in the pipe, but not yet fit for public discussion. I'm grateful for your help and contributions (both past and future). Stuff like this strengthen curl a lot! Thanks! I am running the existing fuzzers, to look for under

Re: Curl coredump after re-use connection

On Tue, 17 Sep 2019, Daniel Stenberg via curl-library wrote: It looks like you're not telling curl how large your post content is and you're not asking for it to get done with Chunked encoding. Sorry, that was wrong. You use CURLOPT_POSTFIELDS and of course curl knows how long

Re: Curl coredump after re-use connection

On Tue, 17 Sep 2019, Nguyen Hoang Tuan (Hubble Vietnam) via curl-library wrote: Please help to review what I miss. It looks like you're not telling curl how large your post content is and you're not asking for it to get done with Chunked encoding. I think curl can only close the connection

Re: http/3 and libcurl-multi

On Mon, 16 Sep 2019, Paul Oliver via curl-library wrote: I'm adding support for http/3 to my libcurl-based library and have come across a strange issue. I add a curl_multi_add_handle as usual and the first request works fine, however when I submit the second I receive no callbacks to my

Re: Two curl pull request for Windows

On Sun, 15 Sep 2019, Gilles Vollant wrote: https://github.com/curl/curl/pull/4346/commits I see a green mark "pass" only on the 5d01a00 commit But not at the latest commit c6af726 c6af726 only change calling a Windows api Unicode to call ansi api to do the same thing both 5d01a00 and c6af726

Re: Getting CURLE_OK with wrong pinned public key after using the correct.

On Fri, 13 Sep 2019, Sebastian Haglund wrote: This patch fixed it for me Lovely, thanks a lot. I've ammended my PR with your update just now and will merge as soon as the CI agrees. -- / daniel.haxx.se | Get the best commercial curl support there is - from me | Private

Re: Verification of :authority in push promise requests too strict?

On Wed, 11 Sep 2019, Christoph M. Becker via curl-library wrote: As of curl 7.64.1 one test case[1] in the PHP test suite is failing, which is caused by . The problem is that the server sends :authority

Re: Getting CURLE_OK with wrong pinned public key after using the correct.

On Fri, 13 Sep 2019, Daniel Stenberg via curl-library wrote: I'll write up a PR for this and get back. Done. Can you try this and see if it fixes the behavior for you? https://github.com/curl/curl/pull/4347 -- / daniel.haxx.se | Get the best commercial curl support there is - from me

Re: Getting CURLE_OK with wrong pinned public key after using the correct.

On Fri, 13 Sep 2019, Sebastian Haglund via curl-library wrote: While adding public key pinning to a cURL c++ wrapper, I discovered that setting the wrong public key after using the correct still yields OK result (expected CURLE_SSL_PINNEDPUBKEYNOTMATCH). It seems to be related to re-using the

Re: Adding a feature to curl: disabling partial range verification

On Thu, 12 Sep 2019, Daniel Stenberg via curl-library wrote: My attempt to work around this kind of rubbish in the response can be found in this PR I just filed: https://github.com/curl/curl/pull/4339 Philippe, can you check if that patch fixes the problem for you? I wrote up a test case

Re: Adding a feature to curl: disabling partial range verification

On Thu, 12 Sep 2019, philippe marguinaud via curl-library wrote: Well, you are right, there is probably a bug somewhere in libcurl, as with -v the FTP server returns the file size : Or a bug in the server that responds like this: SIZE ICMSHFCSTINIT < 213 File: ICMSHFCSTINIT, Size = 860160

Re: Adding a feature to curl: disabling partial range verification

On Thu, 12 Sep 2019, philippe marguinaud via curl-library wrote: Well, from what curl says, it returns a size of zero: $ curl -o ICMSHFCSTINIT -r 100-200 -n ftp://hendrix.meteo.fr/tmp/ICMSHFCSTINIT ** Resuming transfer from byte position 100 % Total% Received % Xferd Average Speed Time

Re: Managing application data fetched from DNS (eg for ESNI)

On Wed, 11 Sep 2019, Niall.oReilly+lists--- via curl-library wrote: Hi Niall, thanks for the status report and it is certainly encouraging to hear that you're progressing nicely! So that libcurl can support ESNI, it needs to be able to fetch and manage data from the DNS beyond the address

RE: curl --data slow with big file

On Thu, 12 Sep 2019, Gilles Vollant wrote: Here is a modified file2string Thanks. I edited it ever so slightly and submitted it as a PR here: https://github.com/curl/curl/pull/4336 -- / daniel.haxx.se | Get the best commercial curl support there is - from me | Private

Re: curl --data slow with big file

On Thu, 12 Sep 2019, Gilles Vollant via curl-library wrote: And in src/tool_paramhlp.c, the function file2string does very often realloc Oh right. Once *per line* of the file. For a 100MB json file, I bet that can end up in quite a lot of reallocs! So I suggest modifying file2string to

Re: Curl and SSL in an IMB's OnDemand environment

On Wed, 11 Sep 2019, Michael Rellstab wrote: Do you mean, OnDemand itself has libcurl linked (statically?) into its binaries? And my code uses this binary instead of the libcurl that is installed on the Linux? That was my question too =) I'm not really sure if we simply can focus on the

Re: [RELEASE] curl 7.66.0

On Wed, 11 Sep 2019, Daniel Stenberg via curl-library wrote: Usng the winbuid system I am running into the error that nmake cannot find vquic\ngtcp2.obj. I tried a lot of command-line argumeents. Something has changed, but what? Does https://github.com/curl/curl/pull/4323 fix the build

Re: [RELEASE] curl 7.66.0

On Wed, 11 Sep 2019, Jan Ehrhardt via curl-library wrote: Usng the winbuid system I am running into the error that nmake cannot find vquic\ngtcp2.obj. I tried a lot of command-line argumeents. Something has changed, but what? Curious, because that bug was previously reported in

Re: Autobuilds broken

On Wed, 11 Sep 2019, Marcel Raad via curl-library wrote: on https://curl.haxx.se/dev/builds.html, I see that one of my autobuilds throws 12 compiler warnings. When I click on it, the page says "Replace this text with the error page you would like to serve to clients if your origin is

[SECURITY ADVISORY] curl: TFTP small blocksize heap buffer overflow

TFTP small blocksize heap buffer overflow = Project curl Security Advisory, September 11th 2019 - [Permalink](https://curl.haxx.se/docs/CVE-2019-5482.html) VULNERABILITY - libcurl contains a heap buffer overflow in the function

[SECURITY ADVISORY] curl: FTP-KRB double-free

porter on September 3, 2019. We contacted distros@openwall on September 5. This advisory was posted on September 11th 2019. CREDITS --- Reported by Thomas Vegas. Patch by Daniel Stenberg. Thanks a lot! -- / daniel.haxx.se | Get the best commercial curl support there is - f

[RELEASE] curl 7.66.0

esniffer13 on github, Daniel Gustafsson, Daniel Stenberg, Dominik Hölzl, Eric Wong, Felix Hädicke, Gergely Nagy, Gisle Vanem, Igor Makarov, Ironbars13 on github, Jason Lee, Jeremy Lainé, Jonathan Cardoso Machado, Junho Choi, Kamil Dudka, Kyle Abramowitz, Kyohei Kadota, Lance Ware, Marcel Raad,

Re: Curl and SSL in an IMB's OnDemand environment

On Tue, 10 Sep 2019, Michael Rellstab via curl-library wrote: Since several days I'm trying out to get my project to work, but I don't have any success. Giving a short overview: I have to implement a UserExit (callback routine) for the IBM's OnDemand Software. Inside this UserExit I'm using

Re: max header size?

On Sun, 8 Sep 2019, Felipe Gasper via curl-library wrote: Is there an option in libcurl to limit the size of HTTP headers that are accepted? No, libcurl has an internal fixed limit for this: CURL_MAX_HTTP_HEADER, set to 102400. -- / daniel.haxx.se | Get the best commercial curl support

Re: Specify specific IP address?

On Fri, 6 Sep 2019, Felipe Gasper via curl-library wrote: Never mind, I think I have it w/ the open-socket callback. I once blogged about different options: https://daniel.haxx.se/blog/2018/04/05/curl-another-host/ -- / daniel.haxx.se | Get the best commercial curl support there is -

Re: curl crashes when internet switched off

On Wed, 4 Sep 2019, Salman Ahmed wrote: Please stop top-posting. It makes it really hard to follow this discussion/thread. I updated my libcurl to 7.65.3 and at the same time I found out that the 3rd party sdk i am using does use curl_easy_setopt(curl, CURLOPT_NOSIGNAL, 1). After I set that

Re: curl crashes when internet switched off

On Wed, 4 Sep 2019, Salman Ahmed wrote: CURL VER libcurl/7.47.0 OpenSSL/1.0.2g zlib/1.2.8 libidn/1.32 librtmp/2.3 Linux VirtualBox 4.15.0-58-generic #64~16.04.1-Ubuntu SMP Wed Aug 7 14:10:35 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux I guess its 3 year old libcurl?I dont remember manually

Re: curl crashes when internet switched off

On Wed, 4 Sep 2019, Salman Ahmed via curl-library wrote: I am facing this problem where my application just terminates when I manually switch off my WiFi while it is waiting on curl_easy_perform() in a separate thread from the main thread. What libcurl version is this? On Linux, right?

Re: Random "couldn't resolve host" failures in curl_multi

On Sun, 1 Sep 2019, Sinus via curl-library wrote: Also, it's always the final handles that fail - n first handles work, n+1 till the end fail to resolve, with n sometimes being all 12, sometimes 8, a few times 4, but I also caught it on 11. "couldn't resolve host" means that the name

Re: Documentation of CURLOPT_HEADERFUNCTION

On Tue, 27 Aug 2019, Christopher Head via curl-library wrote: Done! Excellent. I've merged it! -- / daniel.haxx.se | Get the best commercial curl support there is - from me | Private help, bug fixes, support, ports, new features

Re: Informal 2019/2020 curl roadmap poll

Hi, I've closed the poll. 20 people responded and here are the results. I counted share of "positive votes" (ie voted 4 or 5). If ending on the same precentage, more 5s makes you a winner. Here's how the ten features rank: 1. HSTS (70%) 2. DNS-over-TLS (70%) 3. ESNI (65%) 4. tiny-curl

[poll comment] "grpc"

Hi, Someone mentioned "grpc" as a suggested feature to work on in curl going forward. I don't see the fit. I don't think grpc even has its own URL syntax. If someone wants to argue for this, please help me understand. -- / daniel.haxx.se | Get the best commercial curl support there is -

Re: curl_easy_perform() crash?

On Tue, 27 Aug 2019, Daniel Polski via curl-library wrote: I'm trying to figure out why my application crash sometimes, and ended up with the following stack: Thread #19 [backend_ser] 14800 [core: 3] (Suspended : Signal : SIGABRT:Aborted) __GI_raise() at raise.c:50 0x7783ced7

Re: response reading failed

On Tue, 27 Aug 2019, Christian Schmitz via curl-library wrote: I see that Curl_pp_readresp is the only place this error message is made. else if(gotbytes <= 0) { keepon = FALSE; result = CURLE_RECV_ERROR; failf(data, "response reading failed"); Is it possible we could have

Re: libunbound as name resolver?

On Fri, 23 Aug 2019, Felipe Gasper wrote: That patch's purpose was primarily to enable DNSSEC and DANE/TLSA for libcurl, but yes, it didn't get enough interest and energy from the team or the original developer (Björn) so it rotted pretty fast. DNSSEC and DANE/TLSA are still often mentioned

Re: libunbound as name resolver?

On Fri, 23 Aug 2019, Felipe Gasper via curl-library wrote: Hello! I’d like to have libcurl do recursive name resolutions, toward which end it would be great if libunbound were supported as a name resolution backend. I'm not sure I even I understand what recursive means when done from a

Re: [poll comment] "SNMP request/reply support"

On Wed, 21 Aug 2019, Sergei Nikulov wrote: The first thing to decide with snmp support is: which side of the snmp protocoll you want to support? - client side: to gather information via snmp from devices supporting snmp? This is the primary usage. Curl is a cross-platform multiprotocol

Re: Question about SSL Session Tickets

On Mon, 19 Aug 2019, Joachim Mammele via curl-library wrote: According to https://vincent.bernat.ch/en/blog/2011-ssl-session-reuse-rfc5077 it is necessary to send and empty "session_ticket" extension in the Client Hello message. Is this possible to do with CURL? I don't think it is. I

[poll comment] "SNMP request/reply support"

Another suggestion from the poll. (SNMP is not currently supported by curl.) I'm not personally fluent in SNMP so I would need to do this together with someone who is and someone who has an existing use case that could drive it - so that it becomes a less theoretical and more practical

[poll comment] "Bandwidth limit settings on the multi interface"

Suggested in the free-form field in the roadmap poll. Years ago implemented this feature as an experiment but it turned out pretty complicated. If you do X transfers and they have N bytes/sec allowance in aggregate, it is a hard challange to distribute the available bandwidth on the transfers

[poll comment] "curl needs more tests"

Hi, I totally agree with this, since as long as there are protocols and code in the project that isn't tested and covered by at least one test case we're not good enough. That said, it is also a question of getting return on invested time and energy so I'm very curious which current areas

Re: Informal 2019/2020 curl roadmap poll

On Mon, 19 Aug 2019, Aleksandar Lazic wrote: This ^^ service for example only offers a very basic poll functionality and can't host an 11-question poll the way this roadmap thing is. What's your requirements for the polling solution. I will then try to find another solution instead of

Re: Informal 2019/2020 curl roadmap poll

On Sat, 17 Aug 2019, Daniel Stenberg via curl-library wrote: [the poll is still open] https://docs.google.com/forms/d/e/1FAIpQLSe-T5IrWq73d6QXvdsn8zVhCWKHCwq97J7KrNktOPL0g2Y7pA/viewform Just an idea: If you fill in a suggestion in the free form field at the end of the poll, do consider

Re: Informal 2019/2020 curl roadmap poll

On Sun, 18 Aug 2019, Aleksandar Lazic wrote:  - ESNI First: +1  - MQTT Second: +1 Going by the responses we've seen so far, that seems to match what others think as well... Maybe you can switch from google forms to a more privacy oriented poll service like. Sure! I've tried out

RE: Loosing connection on iOS hangs request

On Wed, 14 Aug 2019, Vincas Razma via curl-library wrote: Getting back to this, any ideas what would help resolving first iOS request hanging after network connection is lost? Still issue on latter Never got answer about DNS setting here This top-post really made me not understand your mail

Informal 2019/2020 curl roadmap poll

Hi team! About six months ago I wrote a document to myself listing twelve features and functionality I wanted to work on for curl. This is the status of them right now: - four of the larger ones are already well underway (HTTP/3, alt-svc, parallel transfers in curl, tiny-curl) - one is

Re: Username and password in URL stopped working with proxy

On Thu, 15 Aug 2019, Mike Crowe wrote: If I make that change to master (and don't modify the GET lines because master doesn't supply the username and password to the proxy that way) then no Authorization header is sent, so the test fails. Right, I saw that as well. Test 168 doesn't appear

Re: Username and password in URL stopped working with proxy

On Thu, 15 Aug 2019, Mike Crowe wrote: I'll have a dig into the unit test infrastructure to see if I can find a way to add a test case for this. I'd previously assumed that the use of a proxy would make creating a test case difficult. I believe we can reproduce this problem with a modified

Re: Username and password in URL stopped working with proxy

On Thu, 15 Aug 2019, Mike Crowe via curl-library wrote: That seems to be using basic authentication, so there's no 401 round trip required to get the nonce. Please can you post the full output when using digest authentication? You didn't say that was a requirement... =) I don't have a digest

curl up 2020: May 9-10 in Berlin, Germany

Hi friends, I'm happy to say that we have a country, a city, a date and a venue all set for curl up 2020! Country: Germany City: Berlin Date: May 9-10, 2020 Venue: http://co-up.de/ Mark you calendars. Plan something for your family to do without you. Prepare for the fun. This day

Re: Username and password in URL stopped working with proxy

On Wed, 14 Aug 2019, Mike Crowe wrote: If this doesn't help point at the problem, please can you post the full output from when you successfully ran: $ ./src/curl http://daniel:he...@www.haxx.se -x localhost:80 --trace-ascii - -o /dev/null It seems to work? : GET http://www.haxx.se/

Re: HTTP/3 with curl

On Fri, 9 Aug 2019, Daniel Stenberg via curl-library wrote: To get everyone up to speed, I'm doing a dedicated "HTTP/3 with curl" live-stream on Wednesday Done. The recorded version is now available here: https://youtu.be/dKT3vlIw6bQ -- / daniel.haxx.se | Get the best comme

Re: CURL Sending FINS without curl_multi_cleanup, and ways to avoid it

On Tue, 13 Aug 2019, Aravindhan Krishnan via curl-library wrote: I have enabled pipeline with pipe-length of 5 (default) I would advice against enabling pipelining. It had several issues and it has been removed in later versions. and have set CURLMOPT_MAX_TOTAL_CONNECTIONS and

Re: [PATCH] asyn-thread: issue CURL_POLL_REMOVE before closing socket

On Sun, 11 Aug 2019, Daniel Stenberg via curl-library wrote: Thanks, I made it into a pull-request: https://github.com/curl/curl/pull/4211 And this is now merged into master. Thanks! -- / daniel.haxx.se | Get the best commercial curl support there is - from me | Private

It is freeze time again

Hi team, I just want to highlight the fact that we close the feature window on Wednesday and then go only-bug-fixing mode until the release, scheduled to happen on September 11. I don't think we have any more features or changes in queue to get merged before this date. (Features that are

Re: [PATCH] asyn-thread: issue CURL_POLL_REMOVE before closing socket

On Sat, 10 Aug 2019, Eric Wong via curl-library wrote: This avoids EBADF errors from EPOLL_CTL_DEL operations in the ephiperfifo.c example. EBADF is dangerous in multi-threaded applications where I rely on epoll_ctl to operate on the same epoll description from different threads. Thanks, I

HTTP/3 with curl

Hi team, To get everyone up to speed, I'm doing a dedicated "HTTP/3 with curl" live-stream on Wednesday, Aug 14th on twitch [1]. A short abstract: HTTP/3 support in curl is still EXPERIMENTAL. It will surely change a little over time! 1. A brief recap of how HTTP/3 differs from HTTP/1

Re: setting Interface precedence in libcurl

On Wed, 7 Aug 2019, Praveen Pvs via curl-library wrote: Is there way to set the precedence of interface(media/network) in libcurl? I wanted to set the order as ETH0, WLAN0, GPRS0,BT/WLAN libcurl only lets you set a single interface to use so if you want to try a set in an order, you need to

First successful HTTP/3 tests

FYI: In the afternoon of August 5 2019, I successfully made curl request a document over HTTP/3, retrieve it and then exit cleanly again. https://daniel.haxx.se/blog/2019/08/05/first-http-3-with-curl/ -- / daniel.haxx.se | Get the best commercial curl support there is - from me

Re: Username and password in URL stopped working with proxy

On Sat, 3 Aug 2019, Mike Crowe wrote: I've attached two logs. One from my build of master, and one from the Debian Stretch version of curl (7.52.1.) It looks like v7.52.1 passes the username and password on to the proxy in the URL, whereas master does not: I ran cd tests && ./runtests.pl

Re: Username and password in URL stopped working with proxy

On Sat, 3 Aug 2019, Daniel Stenberg via curl-library wrote: curl --proxy http://proxy:3128/ --anyauth http://unittest:password@host/ And you're using curl 7.65.3 ? I just tried this with git master: $ ./src/curl http://daniel:he...@www.haxx.se -x localhost:80 --trace-ascii - -o /dev

<    3   4   5   6   7   8   9   10   11   12   >