Comment Re:super user (Score 1) 58
(But then again I wouldn't believe that home routers could be sold with an internet-facing backdoor open by default in their stock firmware, until that happened.)
(But then again I wouldn't believe that home routers could be sold with an internet-facing backdoor open by default in their stock firmware, until that happened.)
(cmake is probably the best, since it's more portable than autoconf).
As a user of autoconfed packages, I find autoconf superior to cmake. Packages built with autoconf have standardized mechanisms for uninstallation (a cmake package may generate an install-manifest file, an uninstall target, or none of the two), to specify where to put documentation, for cross-compilation, and to fine-tune the build and the installation. With cmake, I can't even tell the package where to install libraries (most packages will allow you to do it, but each package has a different standard about the way to be told); with autoconf, I can even specify a sed to be run on the name of the installed binaries (useful if different packages provide different implementations of the same binary) and still have the installed package work. Also, with cmake packages building both static and dynamic libraries at the same time is usually impossible.
Moreover, modern autoconf scripts are (relatively) easy to debug and patch when they don't work; cmake scripts are more scattered and they're written in an obscure mainframish language.
That said, I imagine that using autoconf on non-posix systems might be less funny.
But in this case, well, Apple does something wrong (not even remotely comparable to the trainwreck that Microsoft did with Metro, I'll concede) that devalues the largest part of its already expensive product line, with the exception of the most expensive products, and without adding any value to those either, but Apple fan are happy nonetheless because... it's good to be shown how Apple does not care about who doesn't spend the most?
What is this, an exercise of asceticism in the path of the true Apple worship?
From reading the linked discussion (before people started having shitfits), a dev suggested removing extFS support as "an unnecessary feature"
A dev announced that extFS support had been removed in beta. To people protesting, they replied that the feature was going away full stop. They even modified the ChromeOS feature page stating that ChromeOS had ext support.
because of theoretical security issues
Because of FUD. Stating that supporting the ext file system poses a security issue is FUD: it is FUD by definition, and it is FUD in particular because ext is massively used in security-critical contexts including Google's servers and Google's Android operating system. Why, ext4's key developer is a Google employee IIRC!
and because it interfered with implementing file system renaming (which looks to be surprisingly tricky to do right).
Because they didn't want to implement the few lines of code supposed to invoke the already existing facilities that set the file system label. A thing that, for tricky that it may be, was done right by the Commodore 64's 1541 floppy drive OS, by MS-DOS, by all versions of Windows, by all Linux-based desktops, by AmigaDOS, by OSX, and probably most existing operating system.
In no time at all, objections were posted, some of them rather aggressive in tone.
One of the last comments before disallowing further comments was that they were looking into keeping extFS support, but throwing an error message if you try to rename an extFS volume, and possibly implementing extFS support in userspace for security reasons.
After the slashdot story was published, after my comment was written, when more and more people started stating, most of them politely, that removing ext support would make ChromeOS unsuitable for their work, and that they were upset because there was no credible explanation for the removal of the feature, only after that developers stopped ignoring their discontent and decided to leave ext support in for the time being, but still without writing the code required to alter the filesystem label.
All of this seems quite reasonable when considering what ChromeOS is and its usual usecase.
And when did I say otherwise? I even said the same thing in another comment.
We are each entitled to our own opinion, but no one is entitled to his own facts. -- Patrick Moynihan