Re: Prevent deletion of file when it is being copied
On 27/09/12 06:28, Emmet Hikory wrote: Nimit Shah wrote: While copying a file from my computer to external disk, I by mistake shift+deleted the file. But still the file transfer dialog showed that it was continuing. At the end of the transfer it failed. Hence i request you to add a check for file transfer before deleting the file. As much as this would be a lovely feature, I don't believe that it is something that we could implement in Ubuntu. If a process on linux has opened a file (but not necessaries read it all into memory), and that file is then unlinked by rm or anything else, then the file should still be readable on disk until the process closes it. see http://www.barricane.com/undelete-open-file-from-inode for an example of undeleting a file with this method. this makes me think that this should just work (for the case of copying a single file with nautilus). for copying multiple files, relying on this, would mean that you would have to open (but not read), all the files at the start. this could hit limits of the number of open files when copying many files also, could this be done through file locking? it looks like linux might have the low level stuff needed to implement this. http://en.wikipedia.org/wiki/File_locking#In_Unix-like_systems also -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: SSH and the Ubuntu Server
On 17/11/10 21:38, Dustin Kirkland wrote: This proposal requests that: 1) a new prompt be added to the Ubuntu Server installer 2) this prompt be dedicated to the boolean installation, or non-installation, of the SSH service, as an essential facet of a typical server 3) the cursor highlights the affirmative (yes, please install SSH), but awaits the user's conscious decision you could make the ssh server recommend denyhosts or fail2ban (both prevent brute force attacks by blocking hosts that make to many failed login attempts) sam -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: An Open Letter to Ubuntu Developers -libraries
On 06/09/10 17:05, Tony Atkinson wrote: 1) Applications, and their dependence on desktop environment libraries Currently certain application rely on specific desktop environment libraries to operate. As examples, if you install kate on stock Ubuntu 10.04, you pull in 108 packages, totalling 330Mb just had a look in a maverick live cd (testdrive is amazing :-) ) sudo apt-get install kate 108 new packages 76.7MB download 272MB disk space sudo apt-get install kate --no-install-recommends 85 new packages 65.2MB download 238MB disk space which is slightly better. i suspect if the KDE libs were split more finely it would mean lots more smaller packages, longer install times, longer upgrade times and more work for devs. so it would be a compromise. sam -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: Desktop CDs: Around 60 MB saved on the installed system, 28 MB in binary packages
On 11/08/10 18:50, Till Kamppeter wrote: So on the running system of the Desktop CD 31 MB get saved which can be used for other useful things. i suggest that some of the space is used for poppler-data[0] which is needed[1,2] to display PDFs with eastern fonts (this is latin characters in some PDFs). package size 3,984 kB install size12,488 kB [0] http://packages.ubuntu.com/lucid/poppler-data [1] https://bugs.launchpad.net/ubuntu/+source/evince/+bug/470157 [2]https://bugs.launchpad.net/ubuntu/+source/evince/+bug/616322 -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: Remove OO Draw from the default install
On 16/05/10 12:27, Shane Fagan wrote: Hey all, I forgot to mention this at the session for default app selection but can we remove Open Office Draw from the default ubuntu install? The reasons are quite obvious it just isnt any good and I dont think any of the regular users actually use it. --fagan i agree that oodraw is not much use for most people. but as others have said it probably wont save much space. the draw package seems to be 2.3MB, and install size 9.4MB. i assume that the ability of openoffice to handle documents with shapes in them is in the core. the best vector graphics app is inkscape. it is probably a bit big to add to the default package list. (though it could probably be cut down. inkscape does ship with 36MB of uncompressed tutorial svg files for example) sam -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: Including a system-wide pulseaudio equalizer
On 11/05/10 09:20, Luke Yelavich wrote: On Tue, May 11, 2010 at 08:15:30AM CEST, Chandru wrote: The default media players in Ubuntu, though quite capable do not have graphical equalizers. Rather than including an equalizer in every application, having a system wide equalizer can be very handy especially when playing online videos. I personally think that users will get confused with an EQ. If they find it, adjust something, and find sound is not as good, they will file bugs regarding sound problems that they have caused. I personally think we need to think very very carefully about how we implement EQ, and whether it is needed by the vast majority of people. Luke The solution to bugs like this is to make apport upload the equalizer settings (i think it already does with the volume settings). sam -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: Including a system-wide pulseaudio equalizer
On 12/05/10 10:07, Daniel Chen wrote: On Wed, May 12, 2010 at 11:03 AM, sam tygiersamtyg...@yahoo.co.uk wrote: The solution to bugs like this is to make apport upload the equalizer settings (i think it already does with the volume settings). It uploads the *alsa* mixer perspective, which may be desynced from pulse's idea of those settings (which shouldn't be the case, but...). Best, -Dan is it technically possible/easy to upload pulse's state as well? sam -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
kernel-ppa/mainline missing versions and packages
Hi the mainline ppa seems to be a bit neglected at the moment. there have been no stable updates to the 2.6.33 series (should be up to 2.6.33.2) the source and headers packages are missing for some kenrels, eg v2.6.34-rc4-lucid meaning i can't build extra modules (openafs). thanks sam -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: Remove F-Spot from the LiveCD
Dmitrijs Ledkovs wrote: 20/26 MB of Gimp is Documentation as pointed out on Ubuntu Planet. I think there is also thunderbird documentation(?) in thunderbird-locale-en-gb included on the live cd, even though thunderbird is not included. it seems to be ~1MB. sam -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: Thoughts for assisting those with limited bandwidth
Mackenzie Morgan wrote: I would like to know how they handle situations where the person hasn't updated in 3 weeks and the package has been updated in the meantime. Say, for example: -0ubuntu1 is currently installed -0ubuntu3 is available to install Do they need to install -0ubuntu2 and THEN -0ubuntu3? if you have a simple name scheme then apt could look for foo-0ubuntu3_upfrom-0ubuntu1 if its not found then apt falls back to using foo-0ubuntu3 then it would be up to the repository admin to decide what packages were kept. it might make sense to have a diff against the previous version, and diff against the original release version (for people who just installed from CD) looking at opensuse's repo http://download.opensuse.org/pub/opensuse/update/11.0/rpm/x86_64/ it looks like they only have the delta to the previous version. mac os system updates come it 2 versions. one that updates from the previous minor version (10.4.3 to 10.4.4) and one that will update from any previous (from any 10.4.x) sam -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: Command to flush file system cache to prepare for performance testing?
Martin Olsson wrote: Hi, If grep through all files in a directory twice, then the first time takes a lot longer because the second run will have most of the files available in the in memory file system cache. I want to do some basic performance testing for a scenario involving disk I/O. I'm wondering if there is a command to purge the file system cache, either for a specific directory or file (which would be great) or otherwise purge the whole thing. Basically, to create repeatable perf tests and get consistent results I would like to pin down all the system state so that any changes in numbers will be directly attributable to my code changes. Does anyone know such a command to flush the file system cache? Martin i think this is what you are after # echo 1 /proc/sys/vm/drop_caches see http://linux-mm.org/Drop_Caches for more info/options sam tygier -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: Are file permissions in files on external devices silly?
Marcin 'Qrczak' Kowalczyk wrote: This should not be done unconditionally. The mac os solution is to have a 'enforce permissions on this device' option (in the info/properties of the device). maybe this could be implemented in a similar way to the .is_audio_player file [0]. A safer way to backup to potentially old filesystems is to make a tar achieve. this protects ownership and permissions, and also prevents filename truncation, and case sensitivity problems. sam [0] http://live.gnome.org/Rhythmbox/FAQ -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: Cruft Remover (system-cleaner): testing help?
Lars Wirzenius wrote: http://code.liw.fi/ubuntu/pool/main/s/system-cleaner/ I'm asking for help with testing to verify that these bugs are, indeed, fixed, and that there aren't any new problems introduced. Any help with this would be appreciated. it offers to remove 2 packages for me apt-zeroconf, which i installed by downloading and using gdebi openafs-modules-2.6.26-7-generic, which i installed with module-assistant sam -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: Ubuntu 8.10 significantly slower than previous versions
Markus Hitter wrote: Am 06.11.2008 um 20:21 schrieb Dan Colish: They're using very different gcc versions between the os's. Well, newer gcc's are meant to produce faster code, aren't they? Quite a few GCC optimisations are for specific CPUs. 32bit ubuntu uses very conservative options so that it works on everything back to i586 (original pentium). The core duo used in the test could potentially do better if it took advantage of things like SSE instructions. I remember the ubuntu devs saying the past that they were unconvinced of the advantage of building packages optimised for newer CPUs, but if someone could show good benchmarks they might consider it. If GCC is now better at things like automatic vectorisation it might be a good time to make some new bench marks. https://wiki.ubuntu.com/GenBunToo is probably a good resource to start at if you want to take on this task. Sam Tygier -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: python-numpy
richard wrote: Hi All can you tell me which version python-numpy is being shipped with intrepid please ? the best places to find which versions are in ubuntu are https://launchpad.net/ubuntu http://packages.ubuntu.com/ which show https://launchpad.net/ubuntu/+source/python-numpy http://packages.ubuntu.com/search?keywords=python-numpysearchon=namessuite=allsection=all intrepid currently has 1:1.1.1-1 (which means upstreams 1.1.1 with 1 update from debian). its probably a bit late for 1.2 to get in. you can also look at http://packages.debian.org/ which shows you if debian have anything newer. another good source of information are the package change feeds. http://feeds.ubuntu-nl.org/HardyChanges http://feeds.ubuntu-nl.org/IntrepidChanges etc Sam Tygier -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: Intrepid compatibility with C3 CPUs #5212
Colin Watson wrote: On Thu, Aug 21, 2008 at 01:49:27AM -0400, M G wrote: Subject: Intrepid compatibility with C3 CPUs #5212 Are you referring to a particular bug here? https://launchpad.net/bugs/5212 is something quite different. MG's message got detached from the start of the thread. if you do a title search for C3 through the devel-discuss you will fine the start, from me on 20080818 the bug is https://bugs.launchpad.net/ubuntu/+source/linux/+bug/254453 Thanks Sam Tygier -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Intrepid compatibility with C3 CPUs
Hi The current intrepid test releases will not boot on machines with Via C3 CPUs. At boot the message: This kernel requires the following features not present on the CPU: cmov Unable to boot - please use a kernel appropriate for your CPU. is given. It appears that the kernel now requires the 'cmov' instruction. If i understand right this is a i686 instruction. Has it been officially decided to drop support for older CPUs? I guess there would be benefits for optimising for newer CPUs, but I have not seen anything announced. This has been reported as bug https://bugs.launchpad.net/ubuntu/+source/linux/+bug/254453 , but I think it warrants community discussion. Maybe it is time to split i386 and i686 into separate ubuntu versions. Maybe it is time for us folk with old CPUs to move to a more specialized distro. What should be considered obsolete (mini itx boards with C3s are still for sale http://www.mini-itx.com/store/?c=2 so if Via have stopped building them it can't have been long ago). Sam Tygier -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: Making apt-get powercut-proof
Evan wrote: Reading this, I just thought I'd mention that previous upgrades have caused issues with scrollkeeper-update, in which the only way to continue was to kill the process. Having a 'skip package' option would have been particularly useful in that situation. updating the docs database can take a very long time (i dont know why). you need to be patient. sam -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: Hardy+1 Idea: GoboLinux Filesystem Hierarchy?
what i think is very useful about the gobo system is that it trivially handles multiple versions of a package. /Programs/Firefox/2.0 /Programs/Firefox/3.0 with debian systems there are sometimes packages with multiple versions, eg apache2, but i think it is extra work for the packagers, so it does not happen often Maybe this would solve problems with backports breaking other installed software. Firefox does not get backported because so many packages are built on it. But if it was trivial to have several firefox versions then everything is easier. Sam -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: Hardy+1 Idea: GoboLinux Filesystem Hierarchy?
Mackenzie Morgan wrote: There are two versions of Firefox in Gutsy, and nothing broke from it. Yes, but i think it takes the packagers some effort to modify the firefox3 package to not interfere with the older version. As far as apt is concerned they are separate programs in separate packages. This just reminds me of the FSH on this darned Mac I'm stuck using. I don't like it. It makes paths longer and more annoying to type (remember, our fs is case-sensitive). as long as $PATH is set in a clever way you don't need to type the full path for an executable. I don't know how gobo handle this, but I assume they do. sam tygier -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: apt-cacher in main + apt-zeroconf
Fabian Rodriguez wrote: -BEGIN PGP SIGNED MESSAGE- Hash: RIPEMD160 Sam Tygier wrote: it looks like they have got the security side covered. Now, one might think this could potentially pose a security threat as everyone can offer and distribute debs without any authentication whatsoever. This is not the case as we are not yet caching the package lists or pdiffs, which are PGP-signed and contain MD5, SHA1 and SHA256 checksums of the packages. But due to the trusted PGP signatures, caching package lists shouldn't be an issue. Is there any reason this would not be sufficient? I see many ways to trick someone into installing newer versions of existing common packages that include malicious files, using apt-zeroconf. You'd be surprised how many people will click through any amount of security warnings if approached with authority by a neighbor. An Internet cafe comes to mind, but many other public places would also serve this purpose. You'd guess I love being paranoid about this. As I understand all the computers still get the package list from the ubuntu repo. it is only the packages that they get from local peers. there is no way the local peers can tamper with the package list. the package list contains the MD5 sum of packages. so if a local peer claims to have a package, and gives you something that has been tampered, then apt will reject it. sam -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: apt-cacher in main + apt-zeroconf
Kevin Fries wrote: I am not sure it needs to be moved. But, what would be totally cool is if the installer scanned the local network on install and configured apt-cacher in sources.list instead of the normal repos by default when if finds a server. That would be a terrific usability upgrade. could apt-zeroconf[0] be installed and enabled by default. distributed apt-cacher for local networks implemented in Python. It's called apt-zeroconf since we use avahi for automatically finding other apt-zeroconf instances on the LAN, similar to Apple's Rendezvous/Bonjour/Zeroconf technology. sam [0] http://trac.phidev.info/trac/wiki/AptZeroconf -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: apt-cacher in main + apt-zeroconf
Fabian Rodriguez wrote: apt-zeroconf is actually a replacement for apt-cacher, not a complement to it, according to its site. I think we already know the answer to enabled by default autodiscovery / other networking services. I would have some trust issues using apt-zeroconf, but that's just me :) F. it looks like they have got the security side covered. Now, one might think this could potentially pose a security threat as everyone can offer and distribute debs without any authentication whatsoever. This is not the case as we are not yet caching the package lists or pdiffs, which are PGP-signed and contain MD5, SHA1 and SHA256 checksums of the packages. But due to the trusted PGP signatures, caching package lists shouldn't be an issue. Is there any reason this would not be sufficient? The only thing I can imagine is some sort of DOS attack by sending a large number of requests to one machine. Maybe checking for shared packages on the network could be enabled by default, but sharing disabled. The option to enable sharing could be in System - Administration - Software Sources Sam -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: regular fsck runs are too disturbing
Waldemar Kornewald wrote: Are there any alternatives? Here are two examples: one alternative is fsck at shut down. https://wiki.ubuntu.com/AutoFsck sam tygier -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss