Attachment 160458 Details for Bug 249632 – Summary of Rationale for and implementation of RHEL 1.5 Firefox Strategy

Summary of Rationale for and implementation of RHEL 1.5 Firefox Strategy

ff2.0-rationale.html (text/html), 3.65 KB, created by wdc on 2007-08-01 20:08:24 UTC

(hide)

Description:

Filename:

MIME Type:

Creator: wdc

Created: 2007-08-01 20:08:24 UTC

Size: 3.65 KB

patch

obsolete

><html>
><head></head>
><body>
><h4>Firefox 1.5 not 2.0</h4>
> Mozilla.org made a big push to get everyone to switch from
> Firefox 1.5 to 2.0, and de-supported Firefox 1.5 in May
> 2007. Red Hat, SuSE, and other distributions established a
> collaboration to continue supporting Firefox 1.5 after Mozilla
> discontinued support. 
> The rationale for sticking with 1.5 was:
> <ul>
> <li>There were costs involved in dealing with subsystems that
> had started using libraries from Firefox (devhelp, epiphany,
> galeon, gnome-python2-gtkmozembed, liferea and yelp), and in
> dealing with migration issues for plug-ins and extensions
> based on Firefox 1.5. </li>
> <li>Firefox 2.0 had re-introduced bugs that had been fixed in
> Firefox 1.5 for rendering international characters, printing,
> cascading style sheets, security and random application
> freeze-up.</li>
> <li>RSS handling took a step backwards in functionality. </li>
> <li>The new functionality in 2.0 was deemed not significant
> enough to deal with the costs and regressions. </li>
> <li>A consortium of Linux distributions including Red Hat,
> SuSE, Ubuntu and others formed to continue supporting Firefox
> 1.5 after Mozilla.org support ended.</li>
> <li>Significant improvements to behavior under Linux and
> cleaner support of applications wanting to use the
> Mozilla-supplied libraries were slated for Firefox 3.0. </li>
> <li>Red Hat made the primary focus of Enterprise Client to be
> customers with large deployments that could not easily take an
> upgrade to Firefox.</li>
> </ul>
> With a support consortium in place, a pathway to good code
> and significant functionality improvement established, and the
> committment to not forcing big changes on customers, Red Hat
> locked itself in to Firefox 1.5. 
> It is unfortunate that the other members of the Linux Firefox 1.5 support consortium seem to have caved to the Firefox 2.0 marketing push. All the other distributions offer 2.0 either by default or as an option. This makes Red Hat's carefully reasoned decision look like a bad one.
> There are options for installing Firefox 2.0 on Red RHEL but they create problems:
> <ul>
> <li>Firefox 2.0 and 1.5 profiles are incompatible, so you need
> to use the profile manager and be very careful never to run
> 1.5 on your 2.0 profile. When the time comes to update to the
> new version of Firefox from Red Hat, you will need to be very
> careful that the right profile migrates over and becomes your
> one and only.</li>
> <li>Removing Firefox 1.5 completely would eliminate the
> profile conflict, but then the applications that depend upon
> the Firefox 1.5 libraries (devhelp, epiphany, galeon,
> gnome-python2-gtkmozembed, liferea and yelp) would break.</li>
> <li>Adopting an early release of xulrunner would eliminate the
> library conflict, but that release is not maintained, and
> contains known security problems.</li>
> <li>MIT could craft shadow RPMs that replace the dependent
> subsystems and Firefox, but dealing with any updates by Red
> Hat to the dependent subsystems or Firefox would be messy. If
> both MIT and Red Hat did not agree perfectly on the version
> numbering, systems could get into a state where nothing could
> be installed until a careful removal of components was made by
> a wizard.</li>
> </ul>
> Bottom Line: No Firefox 2.0 for Red Hat Enterprise Linux. 
></body>
></html>

<html>
<head></head>
<body>
<h4>Firefox 1.5 not 2.0</h4>
      <p>Mozilla.org made a big push to get everyone to switch from
      Firefox 1.5 to 2.0, and de-supported Firefox 1.5 in May
      2007. Red Hat, SuSE, and other distributions established a
      collaboration to continue supporting Firefox 1.5 after Mozilla
      discontinued support. </p>
      <p>The rationale for sticking with 1.5 was:</p>
      <ul>
        <li>There were costs involved in dealing with subsystems that
        had started using libraries from Firefox (devhelp, epiphany,
        galeon, gnome-python2-gtkmozembed, liferea and yelp), and in
        dealing with migration issues for plug-ins and extensions
        based on Firefox 1.5. </li>
        <li>Firefox 2.0 had re-introduced bugs that had been fixed in
        Firefox 1.5 for rendering international characters, printing,
        cascading style sheets, security and random application
        freeze-up.</li>
        <li>RSS handling took a step backwards in functionality. </li>
        <li>The new functionality in 2.0 was deemed not significant
        enough to deal with the costs and regressions. </li>
        <li>A consortium of Linux distributions including Red Hat,
        SuSE, Ubuntu and others formed to continue supporting Firefox
        1.5 after Mozilla.org support ended.</li>
        <li>Significant improvements to behavior under Linux and
        cleaner support of applications wanting to use the
        Mozilla-supplied libraries were slated for Firefox 3.0. </li>
        <li>Red Hat made the primary focus of Enterprise Client to be
        customers with large deployments that could not easily take an
        upgrade to Firefox.</li>
      </ul>
      <p>With a support consortium in place, a pathway to good code
      and significant functionality improvement established, and the
      committment to not forcing big changes on customers, Red Hat
      locked itself in to Firefox 1.5. </p>
      <p>It is unfortunate that the other members of the Linux Firefox 1.5 support consortium seem to have caved to the Firefox 2.0 marketing push. All the other distributions offer 2.0 either by default or as an option. This makes Red Hat's carefully reasoned decision look like a bad one.</p>
      <p>There are options for installing Firefox 2.0 on Red RHEL but they create problems:</p>
      <ul>
        <li>Firefox 2.0 and 1.5 profiles are incompatible, so you need
        to use the profile manager and be very careful never to run
        1.5 on your 2.0 profile. When the time comes to update to the
        new version of Firefox from Red Hat, you will need to be very
        careful that the right profile migrates over and becomes your
        one and only.</li>
        <li>Removing Firefox 1.5 completely would eliminate the
        profile conflict, but then the applications that depend upon
        the Firefox 1.5 libraries (devhelp, epiphany, galeon,
        gnome-python2-gtkmozembed, liferea and yelp) would break.</li>
        <li>Adopting an early release of xulrunner would eliminate the
        library conflict, but that release is not maintained, and
        contains known security problems.</li>
        <li>MIT could craft shadow RPMs that replace the dependent
        subsystems and Firefox, but dealing with any updates by Red
        Hat to the dependent subsystems or Firefox would be messy. If
        both MIT and Red Hat did not agree perfectly on the version
        numbering, systems could get into a state where nothing could
        be installed until a careful removal of components was made by
        a wizard.</li>
      </ul>
      <p>Bottom Line: No Firefox 2.0 for Red Hat Enterprise Linux. </p>
</body>
</html>

Actions: View

Attachments on bug 249632: 160458