As of yesterday, Google pushed Chrome stable v28 into their google-chrome repository and effectively killed RHEL6 support:
--> Running transaction check
---> Package google-chrome-stable.x86_64 0:27.0.1453.110-202711 will be updated
---> Package google-chrome-stable.x86_64 0:28.0.1500.45-205727 will be an update
--> Processing Dependency: libstdc++.so.6(GLIBCXX_3.4.15)(64bit) for package: google-chrome-stable-28.0.1500.45-205727.x86_64
--> Finished Dependency Resolution
Error: Package: google-chrome-stable-28.0.1500.45-205727.x86_64 (google)
When using pax -s to rewrite archived files on the fly, it also rewrites the symlink reference even if these symlinks are relative. This is highly unexpected, not documented and incorrect behavior.
Imagine you the following two files:
bin/test sbin/test -> ../bin/test
If you would archive this using:
find . | pax -d -w -x ustar -s ',^,foo-1.2/,' >foo-1.2.tar
You end up with the following files in the archive:
foo-1.2/bin/test foo-1.2/sbin/test -> foo-1.2/../bin/test
Today I learned that Facebook support is non-existing. I had a problem joining a corporate network, and it requires you to add you work email-address and Facebook will send you a link through work email to confirm.
Very simple, right ? Not so for many users, including me. "Sorry, something went wrong. We're working on getting this fixed as soon as we can." The FAQ does not really answer to anything, the forums are full with people reporting this issue, cursing at Facebook, mocking Zuckerbergs billions and the shameful state of support.
I am still alive, yes. My blog cannot be considered a good heartbeat :-) Last blog post must have been at least 3 months ago, but I have been busy. Really busy.
Good thing is that I will be presenting at the LibreOffice conference in Paris tomorrow about Unoconv. I am excited to pour some more energy into this tool and already the first day I have met some interesting people. Thanks to Cor Nouws from Nou&Off who was so kind to introduce me everywhere.
Today Red Hat Enterprise Linux 5.7 was released with the following improvements:
- The new certificate-based CDN infrastructure to get faster access to security updates and bugfixes (which was first introduced with RHEL 6.1)
- Various KVM and Xen improvements, including:
- Xen 32bit performance improvement
- Xen boot time reduced
- KVM CD-ROM emulation improvements
- KVM live migration speed improvements
- Updates to network, storage and video drivers (too many to list)
- XFS is now fully supported with RH HA/Clustering
- Many SSSD improvements
So this looks more like a regular hardware improvement update, with some necessary improvements introduced with RHEL5.6. The most exciting part is that there is not a lot to be excited about, which is what Enterprise Linux is mostly about ;-)
It happened to me today, not sure if it was because I closed my main window and firefox had some other window in the back ? But the next run all my tabs were gone, and they were nowhere to be found in my 'Recently Closed Tabs' list. My sessionstore.js was huge, but re-opening Firefox only brought up the default homepage.
If this happens to you on re-opening Firefox, don't blink a second, make a copy of your sessionstore.bak and sessionstore.js so that Firefox doesn't replace them and you lose everything.
Almost 6 months after the first Red Hat Enterprise Linux 6 release and 2 months after RHEL 6.1 Beta, Red Hat has announced its RHEL 6.1 update release and everyone testing and using RHEL6 have been waiting for this.
Not only because the second release is usually more important, but also to measure the pace of development and fixes going into its update releases. Ever since I have been running RHEL6, I have had a few kernel crashes and a good share of application crashes. RHEL6.1 Beta was a disaster and I regret having moved to it for my personal laptop (but hey, I like to know what to expect and need something to write about).
So while I do not expect a lot of new features just yet, a whole lot of fixes and improvements should have been coming from vendors, customers and users.
I could not find official release-notes just yet, RHN has not been updated either and no mail yet to rhelv6-announce, but do expect an update on this blog when that happens. And I expect to summarize the items I find the most interesting from the release notes and experience using it.
I just received word that the FrOSCon call for papers is due in one week, and since I did not even know there was a FrOSCon CfP out yet, here's hoping more people read this blog that did not know about it :-) Thanks to ScottyTM for the heads up.
Today I resigned from the centos-devel mailinglist, after yet another long thread where the core CentOS devs basically ignore the issues raised and use smoke-and-mirror techniques and personal attacks to obfuscate the discussion.
I have been through this recently with the CentOS 5.6 release delays, today it was regarding confusing version changes that make it hard to compare or relate CentOS packages with RHEL packages. After a whole bunch of hypothetical cases and even hard evidence of the past, a mistake I made was sufficient to start the name-calling, I don't know why I even bother. I don't think I deserve this anyway.
So I am out. No need for me to be on the CentOS-devel mailinglist if everything I say is being twisted, mischaracterized or ignored. I can use my free time for more interesting things.
So long CentOS, and thanks for all the fish.
Update: Since there seems to be some confusion about my involvement in CentOS. I had left the CentOS core team two years before announcing that I left the centos-devel mailinglist.
Way back in July I ordered a Pixel Qi screen hoping I could make it work with an EeePC 1000HE, but while being shipped I started to doubt. I could be ruining both in the process.
Now 8 months later, I bought a Lenovo IdeaPad S10-2 for the purpose of putting that idle (and expensive) Pixel Qi screen in. I had the most difficulties trying to remove the bezel undamaged (I failed slightly). But everything else was a breeze. If I have to do this again I will have to be more brutal instead (ironically).
Sometimes it's good to have someone basically forcing you to make time to improve a lingering situation. Subversion was a long time a bottleneck (literally in I/O, and figuratively in collaboration) in RPMforge. I also had been using our public Subversion infrastructure for some unrelated Open Source projects.
So thanks to Yury V. Zaytsev (!) we are finally undertaking the migration to Github infrastructure. This will also mark the beginning of the project's name translation from RPMforge to RepoForge.
From the very start I have been using OpenOffice (now LibreOffice) for creating my own invoices. The first thing you have to decide is whether to create your invoice with Calc or with Writer. I decided to use Writer, as it is harder to create a proper layout with a spreadsheet document (especially when using window envelopes) that comes out well on different printers, but that's up to one's preference.
Yes, we have heard it a few times before, but this time it is true. CentOS 5.6 is being seeded to mirrors and work has started to bring the Release Notes up to speed. Already 82 days after RHEL 5.6.
Next up is CentOS 6.0, hopefully this one is released before RHEL 6.1, since the RHEL 6.1 Beta is already two weeks out. The fact that CentOS 6.0 is already 145 days behind RHEL 6.0 is something the team will have to think about. Leveraging the community by opening up the QA process is a no-brainer to me.
I haven't blogged as much as I used to, here's hoping I find some time to keep this up :-)
Something I always have to look up in older code is how to test if an argument is numerical. This can be useful for a function that optionally accepts something like a return-code, but also accepts a long string (eg. an error message). Putting the return-code at the end of a long string is far from developer-friendly and the error-message doesn't have to be single string.