Adds formula for C++ graphql parser

[hwhelchel]

This will make it easier to install the libgraphqlparser and use bindings for it in higher level languages. This is my first formula so I may have missed something. Thanks!

/cc @dylanahsmith @rmosolgo @eapache

[bfontaine]

Isn’t there a make install we could use?

[dylanahsmith]

this is just an example program. I doesn't need to be installed.

[hwhelchel]

@dylanahsmith do you have any recommendations on how to test a successful install? This was the easiest way I could see.

[DomT4]

If it's just a test executable, you can keep & use it, but leave it in libexec or pkgshare so users don't bump into it unnecessarily.

[hwhelchel]

@bfontaine looks like there is no make install. /cc @eapache

2015-09-27 20:51:45 -0400

make
install

make: *** No rule to make target `install'.  Stop.

[dylanahsmith]

Since you aren't using make install you are probably missing headers.

Try to compile a simple program that uses libgraphqlparser with it installed. E.g.

$ cat > main.c <<EOF
#include "graphqlparser/c/GraphQLParser.h"
#include "graphqlparser/c/GraphQLAstNode.h"

int main(int argc, char **argv) {
    struct GraphQLAstNode *node = graphql_parse_string("{ field }", NULL);
    if (node != NULL) {
        graphql_node_free(node);
        return 0;
    }
    return 1;
}
EOF
$ cc -lgraphqlparser main.c -o main

[dylanahsmith]

looks like there is no make install

There is in the latest master, but not in version 0.1.0. Maybe we should ask them to make a release so that it can more easily be packaged.

[hwhelchel]

Yep the graphqlparser/c/GraphQLParser.h header file wasn't found. I've submitted a request on the original library for an updated release. Thank you for your help.

[bfontaine]

Unless we need sample_ast later in the test we should probably set the variable directly:

ast = {
  "kind" => "Document",
  "loc" => {
    "start" => 1,
    "end" => 2,
    ...
  }
  ...
}

[hwhelchel]

great idea

[hwhelchel]

@bfontaine there is so much activity on the project I'm having a hard time getting a green build on travis. We're hitting Github's API rate limit. Can you all add a token to the travis environment? Should I create an issue?

Error: GitHub API rate limit exceeded for 208.78.110.199. (But here's the good news: Authenticated requests get a higher rate limit. Check out the documentation for more details.)
Try again in 40 minutes 45 seconds, or create an personal access token:
  https://github.com/settings/tokens
and then set the token as: HOMEBREW_GITHUB_API_TOKEN

[DomT4]

Just ignore Travis here. Jenkins is the primary CI still and won't have that problem.

[ablyler]

@Homebrew/owners the mavericks brew bot machine is stuck working on this, this job has been running for 9+ hours: http://bot.brew.sh/job/Homebrew%20Pull%20Requests/version=mavericks/33886/

[bfontaine]

@ablyler Thanks, I’m killing the build for now.

[ablyler]

@bfontaine thanks, looks like the box might need rebooted. it is odd that it aborted after 180 minutes but stopped when it could delete a file:

Build timed out (after 180 minutes). Marking the build as aborted. FATAL: Unable to delete script file /var/folders/_z/11_kwl954xvch_n92gnzdjzw0000gp/T/hudson1289499836159169147.sh

[ablyler]

@bfontaine looks like it is currently hung up at: Started calculate disk usage of build Finished Calculation of disk usage of build in 0 seconds

[ablyler]

@bfontaine looks like the build has been killed, but the mavericks box isn't coming back up?

[bfontaine]

@ablyler I’m on it but I’m at work so I don’t have access to the credentials necessary to connect to the VM.

[bfontaine]

@ablyler update: I woke up other team members and we’re on it.

[DomT4]

It's alive again now.

[bfontaine]

Thanks @DomT4 ❤️

[hwhelchel]

@DomT4 I wanted to let you know that I re-kicked off the build by closing/re-opening the PR. I did this because the last one failed on jenkins but due to a git error. The new build has been hanging for awhile. My apologies but thought you would like to know in case closing/re-opening a PR exposes a bug in the build process. Thanks

[DomT4]

You can ping one of us to restart builds as well, or we generally get around to going back through bot-failed builds without obvious cause and restarting them throughout the day. The new build hasn't actually started yet, so not hanging per se, it's just in the queue :)

[hwhelchel]

Awesome thanks

On Tue, Sep 29, 2015 at 10:39 PM Dominyk Tiller notifications@github.com
wrote:

You can ping one of us to restart builds as well, or we generally get
around to going back through bot-failed builds without obvious cause and
restarting them throughout the day. The new build hasn't actually started
yet, so not hanging per se, it's just in the queue :)

—
Reply to this email directly or view it on GitHub
#44389 (comment).

[DomT4]

To be a little annoying, if both of these are required only for the test do stage, let's move them to the top of that block.

[hwhelchel]

Good call on simplifying the sample query

[hwhelchel]

Looks like we've got a passing Jenkins build! What can I do to help get this merged?

[dunn]

Just double-checking, this can't use the system bison and flex?

[hwhelchel]

Would that make the dependency implicit? Do all targeted systems have bison and flex? The library comes with a checked in parser and lexer it fallsback to using if the appropriate version of bison and flex aren't present. That's why I was thinking recommended made the most sense.

https://github.com/graphql/libgraphqlparser#building-libgraphqlparser

[dunn]

Xcode 7 has bison 2.3 and flex 2.5.35; what's the minimum version we need here?

[tdsmith]

looks like Bison 3 https://github.com/graphql/libgraphqlparser/blob/master/CMakeLists.txt#L13 -- but if we can just use the distributed ones, that seems fine, too

[hwhelchel]

cool i've removed the depends_on bison and flex

[dunn]

Merged in 81c9da5, thanks for all your work here!