Bug 167596 - [GTK][EFL] Avoid using a thin directory to create the built product on the archive-built-product step.
Summary: [GTK][EFL] Avoid using a thin directory to create the built product on the ar...
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: Tools / Tests (show other bugs)
Version: WebKit Nightly Build
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Carlos Alberto Lopez Perez
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2017-01-30 11:02 PST by Carlos Alberto Lopez Perez
Modified: 2017-01-30 12:07 PST (History)
6 users (show)

See Also:


Attachments
Patch (3.94 KB, patch)
2017-01-30 11:16 PST, Carlos Alberto Lopez Perez
no flags Details | Formatted Diff | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Carlos Alberto Lopez Perez 2017-01-30 11:02:49 PST
When the build-only GTK+ bots create the built product for uploading to the test bots they do:

 * Remove WebKitBuild/Configuration/thin directory
 * Create WebKitBuild/Configuration/thin directory again
 * Copy the needed directories from WebKitBuild/Configuration/ to WebKitBuild/Configuration/thin/
 * Remove the from WebKitBuild/Configuration/thin all *.o files
 * Invoke the zip program to make an archive of WebKitBuild/Configuration/thin/* 

I think this can be optimized with

 * Just invoke the zip program with the right arguments on WebKitBuild/Configuration/

This will not only speed up the built-product step, but will also reduce the disk space requirements. On the GTK+ debug build-only bot this thin directory requires 16GB of disk space.
Comment 1 Carlos Alberto Lopez Perez 2017-01-30 11:16:58 PST
Created attachment 300126 [details]
Patch
Comment 2 Daniel Bates 2017-01-30 11:41:47 PST
Comment on attachment 300126 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=300126&action=review

> Tools/BuildSlaveSupport/built-product-archive:107
> +        if e.errno != 2:

We use used the names errno constant ENOENT (*) instead of hardocding the integer value 2.

(*) https://docs.python.org/2/library/errno.html

> Tools/BuildSlaveSupport/built-product-archive:115
> +    else:

We follow the WebKit Code Style Guidelines and avoid using an else statement when the preceeding if case has a return statement.
Comment 3 Daniel Bates 2017-01-30 11:43:37 PST
(In reply to comment #2)
> Comment on attachment 300126 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=300126&action=review
> 
> > Tools/BuildSlaveSupport/built-product-archive:107
> > +        if e.errno != 2:
> 
> We use used the names errno constant ENOENT (*) instead of hardocding the
> integer value 2.
> 

I meant to write:

We should use the named errno constant ENOENT (*) instead of hardcoding the integer value 2.
Comment 4 Carlos Alberto Lopez Perez 2017-01-30 12:07:05 PST
Committed r211380: <http://trac.webkit.org/changeset/211380>