801437 - ensure_copy loads the entire file into memory before writing it out.

Status: RESOLVED FIXED

(Toolkit :: Application Update, defect)

Description

[Jeff Muizelaar [:jrmuizel]]

On Nightly with OS X this is a cause for sadness because XUL is 1GB in size.

I should be able to come up with a patch for this.

Comment 1

[Jeff Muizelaar [:jrmuizel]]

Attachment: Improve the ensure_copy code

Comment 2

[(no longer active)]

Comment on attachment 671520 [details] [diff] [review]
Improve the ensure_copy code

Review of attachment 671520 [details] [diff] [review]:
-----------------------------------------------------------------

::: toolkit/mozapps/update/updater/updater.cpp
@@ +620,5 @@
>    }
>    return 0;
>  #else
> +
> +#ifdef XP_UNIX

This change does not seem correct.

@@ +647,5 @@
>      return WRITE_ERROR;
>    }
>  
> +  // This block size was chosen pretty arbitrarily but seems like a reasonable
> +  // compramise. The optimal block size on my OS X machine is 100k */

I don't think that you need to mention things about your machine in the code comments.  Also, please fix the typo in "compromise."

@@ +648,5 @@
>    }
>  
> +  // This block size was chosen pretty arbitrarily but seems like a reasonable
> +  // compramise. The optimal block size on my OS X machine is 100k */
> +  const int blockSize = 32768;

Nit: 32 * 1024.

Comment 3

[Jeff Muizelaar [:jrmuizel]]

Attachment: Improve the ensure_copy code

Addresses review comments

Comment 4

[(no longer active)]

Comment on attachment 677372 [details] [diff] [review]
Improve the ensure_copy code

Review of attachment 677372 [details] [diff] [review]:
-----------------------------------------------------------------

Thanks!

::: toolkit/mozapps/update/updater/updater.cpp
@@ +644,5 @@
>    }
>  
> +  // This block size was chosen pretty arbitrarily but seems like a reasonable
> +  // compromise. For example, the optimal block size on a modern OS X machine
> +  // is 100k */

Nit: s#*/##

Comment 5

[Jeff Muizelaar [:jrmuizel]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/d6dd6d94c1a1

Comment 6

[Jeff Muizelaar [:jrmuizel]]

Breakage followup:
https://hg.mozilla.org/integration/mozilla-inbound/rev/81cb145bd52b

Comment 7

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/d6dd6d94c1a1
https://hg.mozilla.org/mozilla-central/rev/81cb145bd52b