Add proper Content-Length based off of the request body's length #85

hello-josh · 2013-07-03T15:48:28Z

Make sure to use serialized_payload length as Content-Length if the request has a body

nategood · 2013-07-03T15:57:00Z

src/Httpful/Request.php

+        if (isset($this->payload)) {
+            $this->serialized_payload = $this->_serializePayload($this->payload);
+            curl_setopt($ch, CURLOPT_POSTFIELDS, $this->serialized_payload);
+            $this->headers['Content-Length'] = strlen($this->serialized_payload);


Haven't checked the HTTP spec yet, but should this be multi byte string length mb_strlen?

Nope, I think you are correct. We want strlen even for multibyte characters.

Correct. It is the size in bytes not number of characters.
On Jul 3, 2013 12:00 PM, "Nate Good" [email protected] wrote:

In src/Httpful/Request.php:

@@ -784,6 +784,14 @@ public function _curlPrep()
curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, $this->strict_ssl);
curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);

// https://github.com/nategood/httpful/issues/84

// set Content-Length to the size of the payload if present

if (isset($this->payload)) {

$this->serialized_payload = $this->_serializePayload($this->payload);

curl_setopt($ch, CURLOPT_POSTFIELDS, $this->serialized_payload);

$this->headers['Content-Length'] = strlen($this->serialized_payload);

Nope, I think you are correct. We want strlen even for multibyte
characters.

—
Reply to this email directly or view it on GitHubhttps://github.com//pull/85/files#r5012797
.

Yep. Silly question. Brain is moving slow this morning.

Result:
string 'test' (length=4)
int 4
int 4
int 4
string 'тест' (length=8)
int 4
int 4
int 8

That fix is incorrect:
http://www.w3.org/Protocols/rfc2616/rfc2616-sec14.html#sec14.13

More specifically: Applications SHOULD use this field to indicate the
transfer-length of the message-body, unless this is prohibited by the rules
in section 4.4.

On Fri, Sep 13, 2013 at 11:12 AM, wackyfrog [email protected]:

In src/Httpful/Request.php:

@@ -784,6 +784,14 @@ public function _curlPrep()
curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, $this->strict_ssl);
curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);

// https://github.com/nategood/httpful/issues/84

// set Content-Length to the size of the payload if present

if (isset($this->payload)) {

$this->serialized_payload = $this->_serializePayload($this->payload);

curl_setopt($ch, CURLOPT_POSTFIELDS, $this->serialized_payload);

$this->headers['Content-Length'] = strlen($this->serialized_payload);

Result:
string 'test' (length=4)
int 4
int 4
int 4
string 'тест' (length=8)
int 4
int 4
int 8

—

Reply to this email directly or view it on GitHubhttps://github.com//pull/85/files#r6349942
.

racingDeveloper · 2013-07-03T16:05:11Z

It looks correct to me! It should contains the number of OCTETs (8bit), so the number of bytes, not chars!

Add proper Content-Length based off of the request body's length

racingDeveloper · 2013-07-04T11:04:16Z

Hi!

Could you update it on packagist?
Composer still download the old version to me :-(

nategood · 2013-07-05T00:08:43Z

Sorry about that. Forgot the version bump in composer.json. Try 0.2.6 via
composer/packagist.

On Thu, Jul 4, 2013 at 7:04 AM, Gabriele Tondi [email protected]:

Hi!

Could you update it on packagist?
Composer still download the old version to me :-(

—
Reply to this email directly or view it on GitHubhttps://github.com//pull/85#issuecomment-20471491
.

Nate Good
Director of Software Engineering
ShowClix, Inc.

Ph: 412.259.5869
Fax: 724.261.5332
http://www.showclix.com

hello-josh · 2013-07-05T15:59:10Z

https://packagist.org/packages/nategood/httpful is still showing dev-master, dev-dev, and 0.2.5 as the latest versions even though composer.json shows 0.2.6 as the version info. Do you have to force packagist to update?

nategood · 2013-07-05T16:24:36Z

Okay figured it out. My git tag and composer.json update weren't aligned.
Packagist didn't like that. Fixed and confirmed.

On Fri, Jul 5, 2013 at 11:59 AM, Josh Johnston [email protected]:

https://packagist.org/packages/nategood/httpful is still showing
dev-master, dev-dev, and 0.2.5 as the latest versions even though
composer.json shows 0.2.6 as the version info. Do you have to force
packagist to update?

—
Reply to this email directly or view it on GitHubhttps://github.com//pull/85#issuecomment-20526069
.

Nate Good
Director of Software Engineering
ShowClix, Inc.

Ph: 412.259.5869
Fax: 724.261.5332
http://www.showclix.com

wackyfrog · 2013-09-13T09:08:22Z

Wrong length for mutlibyte body (utf-8). So, form submits incorrect.

hello-josh · 2013-09-13T12:16:15Z

Are you sure? Content length should be the number of bytes not the number
of characters. Double check the receiving end.
On Sep 13, 2013 5:08 AM, "wackyfrog" [email protected] wrote:

Wrong length for mutlibyte body (utf-8). So, form submits incorrect.

—
Reply to this email directly or view it on GitHubhttps://github.com//pull/85#issuecomment-24382090
.

wackyfrog · 2013-09-13T12:23:40Z

I'm sure.
with ini param 'mbstring.func_overload' = 2 (overload strlen, http://php.net/manual/en/mbstring.overload.php)

Test example:

Result:
string 'test' (length=4)
int 4
int 4
int 4
string 'тест' (length=8)
int 4
int 4
int 8

So, header 'content-length' contains invalid length with strlen (shorter)

nategood · 2013-09-13T16:45:24Z

Ah. Yes. This makes sense if mbstring.func_overload = 2 is set. In that case it will be dependent your mb_internal_encoding default too. We may want to be explicit and use mb_strlen($var, '8bit') for users that are overloading strlen. Agreement?

hello-josh · 2013-09-13T17:00:08Z

I misunderstood the problem. I thought you were suggesting that it should be sending the number of characters. Disregard my previous response(s).

I agree, the mb_strlen($var, '8bit') is the way to go. I completely overlooked the fact that you can overload strlen()

Add phpunit to require-dev

d766061

Make sure to use serialized_payload length as Content-Length if the request has a body

hello-josh mentioned this pull request Jul 3, 2013

Empty body #84

Closed

nategood reviewed Jul 3, 2013
View reviewed changes

nategood added a commit that referenced this pull request Jul 3, 2013

Merge pull request #85 from Trii/proper-content-length

fe72af0

Add proper Content-Length based off of the request body's length

nategood merged commit fe72af0 into nategood:master Jul 3, 2013

nategood mentioned this pull request Jul 3, 2013

Better HTTP protocol compliance under Squid #83

Closed

hello-josh deleted the proper-content-length branch July 5, 2013 15:53

hello-josh mentioned this pull request Sep 13, 2013

Fix for overloaded strlen #93

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add proper Content-Length based off of the request body's length #85

Add proper Content-Length based off of the request body's length #85

hello-josh commented Jul 3, 2013

nategood Jul 3, 2013

nategood Jul 3, 2013

hello-josh Jul 3, 2013

nategood Jul 3, 2013

wackyfrog Sep 13, 2013

hello-josh Sep 13, 2013

racingDeveloper commented Jul 3, 2013

racingDeveloper commented Jul 4, 2013

nategood commented Jul 5, 2013

hello-josh commented Jul 5, 2013

nategood commented Jul 5, 2013

wackyfrog commented Sep 13, 2013

hello-josh commented Sep 13, 2013

wackyfrog commented Sep 13, 2013

nategood commented Sep 13, 2013

hello-josh commented Sep 13, 2013

Add proper Content-Length based off of the request body's length #85

Add proper Content-Length based off of the request body's length #85

Conversation

hello-josh commented Jul 3, 2013

nategood Jul 3, 2013

Choose a reason for hiding this comment

nategood Jul 3, 2013

Choose a reason for hiding this comment

hello-josh Jul 3, 2013

Choose a reason for hiding this comment

nategood Jul 3, 2013

Choose a reason for hiding this comment

wackyfrog Sep 13, 2013

Choose a reason for hiding this comment

hello-josh Sep 13, 2013

Choose a reason for hiding this comment

racingDeveloper commented Jul 3, 2013

racingDeveloper commented Jul 4, 2013

nategood commented Jul 5, 2013

hello-josh commented Jul 5, 2013

nategood commented Jul 5, 2013

wackyfrog commented Sep 13, 2013

hello-josh commented Sep 13, 2013

wackyfrog commented Sep 13, 2013

nategood commented Sep 13, 2013

hello-josh commented Sep 13, 2013