Re: Review Request 39518: Changed write api.

[Jojy Varghese]

> On Oct. 23, 2015, 8:02 a.m., Timothy Chen wrote:
> > src/slave/containerizer/provisioner/docker/registry_client.cpp, line 638
> > 
> >
> > Did we ever find out why before we just use another interface?

I realized that since the registry client refactor already takes care of 
buffered write (and inadvertantly uses the '''write(fd, char* , size_t) ''' 
API, I dont need this patch for now and will wait for the refactor patches. The 
patch that does this is https://reviews.apache.org/r/39340/


- Jojy


---
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/39518/#review103740
---


On Oct. 21, 2015, 6:13 p.m., Jojy Varghese wrote:
> 
> ---
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/39518/
> ---
> 
> (Updated Oct. 21, 2015, 6:13 p.m.)
> 
> 
> Review request for mesos and Timothy Chen.
> 
> 
> Bugs: MESOS-3773
> https://issues.apache.org/jira/browse/MESOS-3773
> 
> 
> Repository: mesos
> 
> 
> Description
> ---
> 
> See MESOS-3773.
> 
> Change:
> io::write(fd, string) has been having issues with junk being written to the
> file. Using io::write(fd, void*, size_t) does not have this issue.
> 
> 
> Diffs
> -
> 
>   src/slave/containerizer/provisioner/docker/registry_client.cpp 
> 24aa95c5800ff4dfb37465b71421b014b5dd5998 
> 
> Diff: https://reviews.apache.org/r/39518/diff/
> 
> 
> Testing
> ---
> 
> Make check.
> 
> 
> Thanks,
> 
> Jojy Varghese
> 
>

Re: Review Request 39518: Changed write api.

[Jojy Varghese]

> On Oct. 23, 2015, 7:59 a.m., Benjamin Bannier wrote:
> > src/slave/containerizer/provisioner/docker/registry_client.cpp, line 642
> > 
> >
> > Wouldn't this potentially cause data loss if the called `write` 
> > function cannot dump everything in the `fd`?

Good point. Didnt realize that write could return unfinished. Addressed in 
https://reviews.apache.org/r/39340/


- Jojy


---
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/39518/#review103739
---


On Oct. 21, 2015, 6:13 p.m., Jojy Varghese wrote:
> 
> ---
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/39518/
> ---
> 
> (Updated Oct. 21, 2015, 6:13 p.m.)
> 
> 
> Review request for mesos and Timothy Chen.
> 
> 
> Bugs: MESOS-3773
> https://issues.apache.org/jira/browse/MESOS-3773
> 
> 
> Repository: mesos
> 
> 
> Description
> ---
> 
> See MESOS-3773.
> 
> Change:
> io::write(fd, string) has been having issues with junk being written to the
> file. Using io::write(fd, void*, size_t) does not have this issue.
> 
> 
> Diffs
> -
> 
>   src/slave/containerizer/provisioner/docker/registry_client.cpp 
> 24aa95c5800ff4dfb37465b71421b014b5dd5998 
> 
> Diff: https://reviews.apache.org/r/39518/diff/
> 
> 
> Testing
> ---
> 
> Make check.
> 
> 
> Thanks,
> 
> Jojy Varghese
> 
>

Re: Review Request 39518: Changed write api.

[Benjamin Bannier]

---
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/39518/#review103739
---



src/slave/containerizer/provisioner/docker/registry_client.cpp (line 615)


I find it surprising that this would need to be handled here since I would 
have expected a proper `write` to deal with that correctly.

Are you trying to work-around a bug elsewhere?



src/slave/containerizer/provisioner/docker/registry_client.cpp (line 633)


This seems to be happen in `write(int fd, void* data, size_t size)` already 
anyway.



src/slave/containerizer/provisioner/docker/registry_client.cpp (line 641)


Wouldn't this potentially cause data loss if the called `write` function 
cannot dump everything in the `fd`?


- Benjamin Bannier


On Oct. 21, 2015, 6:13 p.m., Jojy Varghese wrote:
> 
> ---
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/39518/
> ---
> 
> (Updated Oct. 21, 2015, 6:13 p.m.)
> 
> 
> Review request for mesos and Timothy Chen.
> 
> 
> Bugs: MESOS-3773
> https://issues.apache.org/jira/browse/MESOS-3773
> 
> 
> Repository: mesos
> 
> 
> Description
> ---
> 
> See MESOS-3773.
> 
> Change:
> io::write(fd, string) has been having issues with junk being written to the
> file. Using io::write(fd, void*, size_t) does not have this issue.
> 
> 
> Diffs
> -
> 
>   src/slave/containerizer/provisioner/docker/registry_client.cpp 
> 24aa95c5800ff4dfb37465b71421b014b5dd5998 
> 
> Diff: https://reviews.apache.org/r/39518/diff/
> 
> 
> Testing
> ---
> 
> Make check.
> 
> 
> Thanks,
> 
> Jojy Varghese
> 
>

Re: Review Request 39518: Changed write api.

[Timothy Chen]

---
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/39518/#review103740
---



src/slave/containerizer/provisioner/docker/registry_client.cpp (line 637)


Did we ever find out why before we just use another interface?


- Timothy Chen


On Oct. 21, 2015, 6:13 p.m., Jojy Varghese wrote:
> 
> ---
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/39518/
> ---
> 
> (Updated Oct. 21, 2015, 6:13 p.m.)
> 
> 
> Review request for mesos and Timothy Chen.
> 
> 
> Bugs: MESOS-3773
> https://issues.apache.org/jira/browse/MESOS-3773
> 
> 
> Repository: mesos
> 
> 
> Description
> ---
> 
> See MESOS-3773.
> 
> Change:
> io::write(fd, string) has been having issues with junk being written to the
> file. Using io::write(fd, void*, size_t) does not have this issue.
> 
> 
> Diffs
> -
> 
>   src/slave/containerizer/provisioner/docker/registry_client.cpp 
> 24aa95c5800ff4dfb37465b71421b014b5dd5998 
> 
> Diff: https://reviews.apache.org/r/39518/diff/
> 
> 
> Testing
> ---
> 
> Make check.
> 
> 
> Thanks,
> 
> Jojy Varghese
> 
>

Review Request 39518: Changed write api.

[Jojy Varghese]

---
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/39518/
---

Review request for mesos and Timothy Chen.


Bugs: MESOS-3773
https://issues.apache.org/jira/browse/MESOS-3773


Repository: mesos


Description
---

See MESOS-3773.

Change:
io::write(fd, string) has been having issues with junk being written to the
file. Using io::write(fd, void*, size_t) does not have this issue.


Diffs (updated)
-

  src/slave/containerizer/provisioner/docker/registry_client.cpp 
24aa95c5800ff4dfb37465b71421b014b5dd5998 

Diff: https://reviews.apache.org/r/39518/diff/


Testing
---

Make check.


Thanks,

Jojy Varghese

Re: Review Request 39518: Changed write api.

[Mesos ReviewBot]

---
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/39518/#review103480
---


Patch looks great!

Reviews applied: [39518]

All tests passed.

- Mesos ReviewBot


On Oct. 21, 2015, 6:13 p.m., Jojy Varghese wrote:
> 
> ---
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/39518/
> ---
> 
> (Updated Oct. 21, 2015, 6:13 p.m.)
> 
> 
> Review request for mesos and Timothy Chen.
> 
> 
> Bugs: MESOS-3773
> https://issues.apache.org/jira/browse/MESOS-3773
> 
> 
> Repository: mesos
> 
> 
> Description
> ---
> 
> See MESOS-3773.
> 
> Change:
> io::write(fd, string) has been having issues with junk being written to the
> file. Using io::write(fd, void*, size_t) does not have this issue.
> 
> 
> Diffs
> -
> 
>   src/slave/containerizer/provisioner/docker/registry_client.cpp 
> 24aa95c5800ff4dfb37465b71421b014b5dd5998 
> 
> Diff: https://reviews.apache.org/r/39518/diff/
> 
> 
> Testing
> ---
> 
> Make check.
> 
> 
> Thanks,
> 
> Jojy Varghese
> 
>