Bug#372067: #356558: xserver-xorg: Matrox 550 DVI port doesn't work
i will try this and report back on my g550 non-dual dvi, if that'll help. but first, does anybody know: o which versions of x (xorg in etch/lenny/sid) will it work with? o does one simply upgrade to that xorg, install this driver, and change xorg.conf? or is there more to do, like tweak xrandr, to get it working? very exciting. thanks. -- Webmaster: do you believe that people will (a) switch browsers to view your best viewed with page or (b) go to your competitor? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#356558: #356558: xserver-xorg: Matrox 550 DVI port doesn't work
oops, my first question was unnecessary; Now that David pushed X.org 7.3 to unstable implies that we'll need to upgrade to unstable to get the dvi to work. my second question nevertheless applies, as it's worth recording here whether other stuff needs to happen, like installing xrandr or anything else. upgrading from xfree 4.2 + proprietary blob to floss xorg will be nice. thanks. -- Webmaster: do you believe that people will (a) switch browsers to view your best viewed with page or (b) go to your competitor? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#435472: closed by maximilian attems [EMAIL PROTECTED] (Re: Bug#435472: linux-source-2.6.22: make booting more pleasant to debug)
thanks for your reply. it is not fair to call this a support request; i was not asking for support. boot parameters won't work, because they should be implemented by default, and because afaik they are not all implemented. are you claiming that the items are implemented? i did file it as wishlist. for example, rebooting immediately or after a delay does not solve the problem. do you really need a command line option to let the user decide interactively when to reboot and not panic? in general, i think the kernel by default can make the booting experience better for users. it sounds like you disagree. On 7/31/07, Debian Bug Tracking System [EMAIL PROTECTED] wrote: This is an automatic notification regarding your Bug report #435472: linux-source-2.6.22: make booting more pleasant to debug, which was filed against the linux-source-2.6.22 package. It has been closed by maximilian attems [EMAIL PROTECTED]. Their explanation is attached below. If this explanation is unsatisfactory and you have not received a better one in a separate message then please contact maximilian attems [EMAIL PROTECTED] by replying to this email. Debian bug tracking system administrator (administrator, Debian Bugs database) -- Webmaster: do you believe that people will (a) switch browsers to view your best viewed with page or (b) go to your competitor? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#435469: closed by maximilian attems [EMAIL PROTECTED] (Re: Bug#435469: linux-source-2.6.22: udev: common user action causes undecipherable hang)
thanks for your reply. i guess that was not what you meant.. mirroring an active root to your new drive fails miserably. why? because /dev is hidden. but why *must* it fail miserably? i did propose two solutions. the error occurs before, not after, the first INIT message, but i see now that it is userspace. so why not reassign to the appropriate userspace package? it is not a bad thing to make the boot process easier for users without harming anything else. *i* don't need this fixed, because i figured out the workaround (copying .static/dev as a special case after copying everything else). so as you can see this is not a support request. but it is appropriate to save the lives of the people on the street below *other people* from having computers fall on them. mirroring root to a backup partition with rsync is common whether or not it's a good idea to do so without lvm snapshots and the like (i don't use lvm because it's an extra layer and i don't know if it would unshadow /dev or not; i assume not). thanks. -- Webmaster: do you believe that people will (a) switch browsers to view your best viewed with page or (b) go to your competitor? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#432150: [Pkg-cryptsetup-devel] Bug#432150: /sbin/cryptsetup: repair tools needed
i wish that i could help more. i do still have the corrupted partition (and the identical working one) and will try suggestions, but will have to overwrite it soon. the dm-crypt discussion was brief and afaik to the effect that the header format is well documented, so if you have a similar header on another partition, you repair the corrupted one by using dd for the homologous parts. e.g. 2007-06-08 07:41:52 GMT on the same url. a web search reveals data corruption stories. one person found that strace enabled opening (didn't work for me). there might be dm issues in addition to dm-crypt or luks issues (brief scan of dm mailing list subjects). i am physically limited in computer use else i would give you urls. thanks. -- Webmaster: do you believe that people will (a) switch browsers to view your best viewed with page or (b) go to your competitor? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#372067: xserver-xorg-video-mga: Please explicitly say in the package description and man page what the driver is capable of
nikolaus, or any others, have you found anything that works for you yet? (etch, lenny, sid, ubuntu, tuxx?) thanks. -- Webmaster: do you believe that people will (a) switch browsers to view your best viewed with page or (b) go to your competitor? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#372067: xserver-xorg-video-mga: Please explicitly say in the package description and man page what the driver is capable of
as of recently, it was not clear to a non-x-expert even with nontrivial google searching: whether the open source driver is sufficient for simple 2d dvi use with magic settings -- or not whether the tuxx driver is nec or sufficient -- or has issues how to best vet the tuxx driver for security (the author seems like a great guy, btw) whether the ubuntu .deb will work on etch (how to check its signature from etch is probably easy but requires work to find out how) whether feisty would work how likely any of the above are to break upon an upgrade of x (or libraries) -- Webmaster: do you believe that people will (a) switch browsers to view your best viewed with page or (b) go to your competitor? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#372079: DVI on G550, some bad news, some good news
On 5/8/07, Nikolaus Schulz [EMAIL PROTECTED] wrote: I'll probably try Linux 2.6.18 with matroxfb after upgrading to Etch. can you let us on this bug (i seem to be on its forward list, which is good and i hope will continue to be the case after any mergings) know the results? i can try the same by setting the right things and installing etch on a new disk, if i know what to set and can do so without debug iterations. unfortunately, i am too ignorant about x stuff. i don't even know: which kernel option is relevant, what enabling in the console means, whether anything would be required in the x conf file, what a framebuffer is exactly, and even what dfp means. (i guess digital flat panel.) icih, the last time i tried fixing this was a while back, but it was probably a 2.6 kernel. here is my current kernel. (1) 08-Tue-14-16-11 ~# uname -a;gunzip -c /proc/config.gz | grep -i matrox Linux debian 2.6.18myver #1 SMP Sun Mar 11 14:33:06 MST 2007 i686 GNU/Linux CONFIG_FB_MATROX=m # CONFIG_FB_MATROX_MILLENIUM is not set # CONFIG_FB_MATROX_MYSTIQUE is not set CONFIG_FB_MATROX_G=y # CONFIG_FB_MATROX_MULTIHEAD is not set p.s. if millenium is the right option, is it a typo for millennium that needs to be fixed, or is that just the way it is, kind of like http referer? thanks. -- Webmaster: do you believe that people will (a) switch browsers to view your best viewed with page or (b) go to your competitor? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#372067: xserver-xorg-video-mga: Please explicitly say in the package description and man page what the driver is capable of
On 5/7/07, Brice Goglin [EMAIL PROTECTED] wrote: forcemerge 372079 372067 thank you hey, thanks for taking an interest. I have not been able to determine whether this package is *supposed* to work with DVI output on a Matrox G550 card. Does regular single-head DVI operation require special Matrox modules? Some problems have been reported [1,2] and should be fixed. I just pingued your bug #372079 to know its status. I don't think it is a documentation issue. not sure i agree with the last statement. if it is known not to work because of bugs, users need to know enough details (without wading through bug reports) so that they don't get bloody concussions from banging their heads against a brick wall. the ambiguous wording in the man page imho invites those concussions. please, save the foreheads. (parenthetical note: i speak from experience. my non-working dvi output on my pretty-common g550 video card has been a thorn in my side for a really long time (years) and has led to this situation: o i have to use xfree 4.2 with an outdated proprietary sourceless driver o i have to keep many other packages held back to be consistent with that o i therefore can't even use firefox, much less iceweasel, because an infuriating bug, apparently but not necessarily in a version of a gdk library that cannot be upgraded, decides to deliberately crash, randomly and frequently, instead of simply failing the operation apparently not supported by xfree (server does not implement operation or some such). i *don't* mean to complain about this, because everybody is a volunteer and matrox is apparently making it really difficult for people. i wish i could help more. my purpose in mentioning this is to inform you that it is *not trivial* for a user to debug and fix this sort of thing when his or her computer use is physically limited (not everybody can do the typing required for debugging) or technical expertise is limited. my main point is that it is worse when it's hard to get a straightforward answer on whether it is even possible. i don't know whether ubuntu's decision re drivers is relevant or not, but it is imho understandable that ubuntu starts to look attractive if the only alternative is to buy another video card. but other than this one thorn, debian works ok end parenthetical note.) perhaps you mean that it is not JUST a documentation issue? If the driver is slowly picking away at the closed-source, kind of working and kind of not but expected to improve and would users please report bugs -- that sort of thing -- that would be good to mention. There are bugs, sorry, and few people to debug them or fix them, so it takes time to get them fix. If DVI still does not work for you with latest upstream, you might want to discuss this upstream and maybe even help them debug/fix. are you saying i should try again? if so, i'll try a fresh etch install and the installer's x conf. i have been following the driver version numbers and i don't think i saw any changes that might be relevant. perhaps i missed something? it still seems like the man page is ambiguous on whether single dvi is expected to work or not. it still seems like it's not going to work. perhaps i should try installing the ubuntu .deb for the driver if etch doesn't work. please do not close these bugs yet! again, thanks for looking at this. -- Webmaster: do you believe that people will (a) switch browsers to view your best viewed with page or (b) go to your competitor? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#372079: xserver-xorg-video-mga: strange frequency behavior with dvi
hi, please do not close these bugs yet. i replied in detail to your comment on Bug#372067. thanks. -- Webmaster: do you believe that people will (a) switch browsers to view your best viewed with page or (b) go to your competitor? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#387857: firefox crashes when it could probably continue instead
i went back to firefox because iceweasel crashes but had more bugs in session saving (with or without extensions) and other things. i have 90 tabs so that's a dealbreaker. i am still using xfree 4.2 because no xorg driver supports dvi on matrox g550. at least not easily. is it possible that this is a bug in gtk, gdk, or a similar library? i ran across a hint that that *might* be the case but do not have the ref. if so, i cannot upgrade the library until i upgrade to xorg. then i would have to buy a new video card. thanks. -- Webmaster: do you believe that people will (a) switch browsers to view your best viewed with page or (b) go to your competitor? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#377763: bash: here document fails with read-only root fs?
On 1/8/07, Justin Pryzby [EMAIL PROTECTED] wrote: What do you mean, mount set? just the set of mounts i was using. it does not seem to respect that variable. Did you do eg: TMPDIR=/foo bash, or just export TMPDIR=/foo? most likely the former with bash -i. i would have wanted to avoid bash using the previous value, which i suspect is what you are referring to. i am not currently in a position to test this. thanks. -- Webmaster: do you believe that people will (a) switch browsers to view your best viewed with page or (b) go to your competitor? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#387857: firefox crashes when it could probably continue instead
Package: firefox Version: 1.5.dfsg+1.5.0.6-5 Severity: normal i use xfree86 4.2 because my video card does not work with xorg. probably because of this, firefox crashes randomly about once per day, seemingly upon opening a new tab. here is a sample: The program 'Gecko' received an X Window System error. This probably reflects a bug in the program. The error was 'BadImplementation (server does not implement operation)'. ^^^ (Details: serial 87751702 error_code 17 request_code 153 minor_code 22) it looks as if the server, which perhaps means the x server, does not implement an operation. then ff crashes all 18 windows and 130 tabs. on my connection, restarting ff takes hours to finish downloading all of the tabs. if this is what is going on, there is no need to debug the situation, because the server simply does not have the feature ff wants. could ff be made to simply pop up a dialog or replace the tab contents with an error page instead of crashing? of course, with the error message given, we don't actually know what server it is referring to. thanks. -- System Information: Debian Release: testing/unstable Architecture: i386 (i686) Shell: /bin/sh linked to /bin/bash Kernel: Linux 2.6.16 Locale: LANG=C, LC_CTYPE=C (charmap=ANSI_X3.4-1968) Versions of packages firefox depends on: ii debianutils 2.17 Miscellaneous utilities specific t ii fontconfig2.3.2-7generic font configuration library ii libatk1.0-0 1.12.2-1 The ATK accessibility toolkit ii libc6 2.3.6.ds1-4GNU C Library: Shared libraries ii libcairo2 1.2.4-1The Cairo 2D vector graphics libra ii libfontconfig12.3.2-7generic font configuration library ii libfreetype6 2.2.1-5FreeType 2 font engine, shared lib ii libgcc1 1:4.1.1-13 GCC support library ii libglib2.0-0 2.10.3-3 The GLib library of C routines ii libgtk2.0-0 2.8.20-1 The GTK+ graphical user interface ii libjpeg62 6b-13 The Independent JPEG Group's JPEG hi libpango1.0-0 1.12.3-1+b1Layout and rendering of internatio ii libpng12-01.2.8rel-5.2 PNG library - runtime ii libstdc++64.1.1-13 The GNU Standard C++ Library v3 ii libx11-6 6.9.0.dfsg.1-6 X Window System protocol client li ii libxft2 2.1.8.2-5.1FreeType-based font drawing librar ii libxinerama1 1:1.0.1-4.1X11 Xinerama extension library ii libxp61:1.0.0.xsf1-1 X Printing Extension (Xprint) clie ii libxt66.9.0.dfsg.1-6 X Toolkit Intrinsics ii psmisc22.3-1 Utilities that use the proc filesy ii zlib1g1:1.2.3-13 compression library - runtime firefox recommends no packages. -- no debconf information -- Webmaster: do you believe that people will (a) switch browsers to view your best viewed with page or (b) go to your competitor? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#315350: privoxy: doesn't go there anyway
it's a year later, and it works now. i didn't notice that it was fixed because i rarely go there anyway. ok to close. a far bigger problem i have is that i get 404 pages 9 times out of 10. but i'm not sure it's a privoxy problem. -- Webmaster: do you believe that people will (a) switch browsers to view your best viewed with page or (b) go to your competitor? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#205702: less: provide text preprocessors as an option for users
thanks for your reply. clearly you don't want text preprocessors for yourself. but why not provide the most common ones for those who want it? i don't think they will ask you to make it default. instead they will be grateful for saving them the debugging. man using less does not work because i want all of the man pages in less so that i can go back and forth among the man pages. instead it only lets you go to the next file and you have to press return. -- Webmaster: do you believe that people will (a) switch browsers to view your best viewed with page or (b) go to your competitor? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#368973: colordiff: colorization makes some output lines disappear
i tested it and the hypothesis is correct. if you can't query the existing colors, then there is little that can be done. it would be nice if ansi included that. i'd definitely update the man page to say that. it's kind of like data loss otherwise. otherwise, by all means close. thanks. -- Webmaster: do you believe that people will (a) switch browsers to view your best viewed with page or (b) go to your competitor? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#369177: dpkg: error message about version string is insufficient
On 5/28/06, Guillem Jover [EMAIL PROTECTED] wrote: The output just before that error may be important to know which package is at fault, please could you send it? it didn't give any hints which packages it was looking at. i can't search for it now, having had a long run of upgrades and not using screen(1). I'll add the version string to the output, so next time it may be a bit easier to track down. thanks. imho this goes for just about all error messages everywhere. This, we cannot fix as the string is coming from something like: $ dpkg --compare-versions 1 2 gt 1.4 it would be great if the caller that had the package name printed the package name before making that call. thanks. -- Webmaster: do you believe that people will (a) switch browsers to view your best viewed with page or (b) go to your competitor?
Bug#368936: zsh: portable shell xargs too hard to write (please make a truly raw read -r)
i guess i am one of those opinionated people who thinks that a shell should also be a good programming language, and that certain things should be easy to do portably and correctly and with confidence that it is portable and correct. imho it's bad that xargs is hard to write portably and with full attention to correctness (quoting, signals, reentrance, exit codes, variable shadowing, etc.), and it's bad when the user isn't sure whether something will work or not. imho basic things like an xargs or a truly raw read should be obvious, especially in a shell. use a function that splits on ifs but set ifs first is not entirely obvious. many of us do exactly that, but it is a convention rather than an obvious solution. at some point most noninteractive shell users will want to write such a loop. a person who thinks that languages should be designed to make obvious things obvious will assume that read is the wrong tool for the job, go on a wild goose chase trying to get line(1) to work, and ultimately abandon that route because of about 10 other limitations in the shells. only then will he go back to using read and ifs. and even then he will doubt its correctness. i hit on a solution that more or less works, but if the solution that i came up with (see my script in the original post) doesn't modify any lines (i.e. it is a truly raw read), then at the very least it would be helpful to have the man page say that if you want a truly raw read, it is safe to use ifs='\n' to get it. more broadly, it shouldn't be so time-consuming to do something like xargs correctly. i doubt that my solution above is even correct, although it works fine in normal circumstances. there are several solutions: implement xargs as a builtin, fix the other limitations, do a truly raw read without ifs, or, at the very least, add a sentence to the man pages. preferably while communicating with any bash people doing the same so that scripts can be portable. i hope that made it clear. if none of the solutions appeals to you, feel free to close the report. thanks. -- Webmaster: do you believe that people will (a) switch browsers to view your best viewed with page or (b) go to your competitor?
Bug#369177: dpkg: error message about version string is insufficient
also, i really like error messages that say what program is delivering the error message. so if it's dpkg, it could say: dpkg: version string ... in package ... is bad. it should not have spaces. or whatever. then i will know what package to file the bug against. something tells me it's not aptitude, although it could be. thanks. -- Webmaster: do you believe that people will (a) switch browsers to view your best viewed with page or (b) go to your competitor?
Bug#368936: zsh: portable shell xargs too hard to write (please make a truly raw read -r)
On 5/25/06, Clint Adams [EMAIL PROTECTED] wrote: what i would propose, if i thought that both the bash and zsh development teams would be interested in them, would be about 10 improvements, including a portable built in xargs and xargs -0 from the get-go. Have you looked at zargs? zargs is completely different from xargs and xargs -0. it is also nonportable. i would like something that takes lines or NUL-separated lines as input and performs the command on them. zargs does not do that and does not exist in bash. the shell script i provided is portable and does that, but it's a serious issue to write it, it isn't clear that it works in all edge cases, and it probably has efficiency issues. zargs looks nice for what it does, though. but i will limit my ambitions and propose a truly raw read -r. this is because i think that both teams would want that. You're proposing this? IFS=$'\n' read -r $1 no. as i explained, i already use that solution. if you want to say that it's wrong to change the behavior of read -r, then i'll accept that. but in that case it wouldn't hurt to change the docs in all 3 shells to emphasize that leading spaces are dangerous and mention ifs as the solution. ideally, however, some of the limitations to the shells that made it into the comments in the script i provided would be solved, or xargs and xargs -0 would be worked out portably for both shells and included either as improvements to my script or as efficient portable builtins. take a look at the comments in the script. of course, it's all up to the devs. if they think i'm wrong, then kill the bug. but surely they can see the merit in enabling a common type of loop. thanks. -- Webmaster: do you believe that people will (a) switch browsers to view your best viewed with page or (b) go to your competitor?
Bug#368683: linux-source-2.6.16: menuconfig should say what the raw default is
even more relevantly, it would also be good to know: o what defaults linux-source-* changed thanks. -- Webmaster: do you believe that people will (a) switch browsers to view your best viewed with page or (b) go to your competitor?
Bug#368672: w3m: ergonomics for scrolling
please try the following command from the shell prompt. LESS='-Q -i -W -j5' less /etc/passwd press space bar just enough times to get to the end of the document, but no more. notice how the last line is in reverse video. that is what i am suggesting. press space bar again. notice how the reverse video goes away. man less: -w or --hilite-unread Temporarily highlights the first new line after a forward movement of a full page. The first new line is the line imme- diately following the line previously at the bottom of the screen. Also highlights the target line after a g or p command. The highlight is removed at the next command which causes move- ment. The entire line is highlighted, unless the -J option is in effect, in which case only the status column is highlighted. -W or --HILITE-UNREAD Like -w, but temporarily highlights the first new line after any forward movement command larger than one line. -- Webmaster: do you believe that people will (a) switch browsers to view your best viewed with page or (b) go to your competitor?
Bug#368673: w3m: ergonomics for searching
that is very nice. but that keeps the search home in the middle of the page. also, could it be made to highlight ALL of the matches with the current match highlighted differently? emacs does this. -- Webmaster: do you believe that people will (a) switch browsers to view your best viewed with page or (b) go to your competitor?
Bug#364848: general: please have package short descriptions say what the package REALLY does
imo users who don't know what a short description refers to will ignore it or look it up, but they will not likely go to the trouble of filing a bug report. so i think that guidelines for maintainers are likely the best place, rather than individual bug reports. -- Webmaster: do you believe that people will (a) switch browsers to view your best viewed with page or (b) go to your competitor?
Bug#307108: Workrave obscures?
info that you want to use while away. -- Webmaster: do you believe that people will (a) switch browsers to view your best viewed with page or (b) go to your competitor?
Bug#366763: httrack: traps signals and attempts to be fancy, causing wrong behavior
is it definitely the case that you can't use a different, less common keystroke for that purpose, like ctrl-\ or something? ^Z is reversible, but the backgrounding thing does not seem to be. -- Webmaster: do you believe that people will (a) switch browsers to view your best viewed with page or (b) go to your competitor?
Bug#366763: httrack: traps signals and attempts to be fancy, causing wrong behavior
since the docs do not seem to be in the man page, i suggest documenting the nonstandard ^Z behavior thoroughly in the man page, including how to reconnect the terminal to the process, how to communicate with it, where the nohup file gets dropped (not always in the directory given in the command line, it seems), etc. also, i think it would make sense to document it in its output, perhaps on a single line. i doubt i'm the only surprised user, even if i am the most vocal. :-) thanks. -- Webmaster: do you believe that people will (a) switch browsers to view your best viewed with page or (b) go to your competitor?
Bug#364848: general: please have package short descriptions say what the package REALLY does
are you saying that guidelines for maintainers already remind the maintainer sufficiently? if not, i think they should; if so, i will do as you suggest when i can. (perhaps the guidelines didn't explain it clearly enough for some maintainers.) thanks. -- Webmaster: do you believe that people will (a) switch browsers to view your best viewed with page or (b) go to your competitor?
Bug#306693: acknowledged by developer (Bug#306693: fixed in cpio 2.6-6)
- copyin.c: Separate out path sanitizing to safer_name_suffix(): Apart from leading slashes, filter out .. components from output file names if --no-absolute-filenames is given, to avoid path traversal. [CAN-2005-1229] closes: #306693. bug submitter here. thanks for working on this. does this mean cpio -i --no-absolute-filenames? i have not tried the new version yet (i will as soon as possible but it will be a while so i wanted to respond now). does it address symlinks with .. also (i.e. my second note)? -- Webmaster: do you believe that people will switch browsers to view your page instead of going to your competitor?
Bug#319295: mozilla-firefox: keyboard frequently focuses previous tab
On 7/20/05, Eric Dorland [EMAIL PROTECTED] wrote: I can't reproduce this. What do you mean it might be related to c-w? my kb actions include ctrl-pgup/down and ctrl-w. that is all i meant. might also have to do with having a slow connection. Do you have any extensions installed? happens with no extensions. -- Webmaster: do you believe that people will switch browsers to view your page instead of going to your competitor?
Bug#319286:
ok, i found the culprit. please reassign to gtk-theme-switch, and add the following problem in addition to the problem of not naming the program in the file. switch2 is far too generic a name for the program. lots of things switch things. i don't mean this rudely at all, but i'd really like to know what people are thinking when they don't make error messages or other text show the name of the program, and name programs really generically. i mean that. it is not a rhetorical question or meant rudely. what are they thinking? thanks. -- Webmaster: do you believe that people will switch browsers to view your page instead of going to your competitor?
Bug#319286: libgtk2.0-common: please say AUTO-WRITTEN by whom
not to mention the fact that the file is overwritten. thanks. -- Webmaster: do you believe that people will switch browsers to view your page instead of going to your competitor?
Bug#315191: mozilla-firefox: fails to increase unreadably small text
On 6/25/05, t takahashi [EMAIL PROTECTED] wrote: one fix might be additive, not multiplicative scaling. if you double 1mm and 10mm, then you should get 2mm and 11mm instead of 2mm and 80mm. i meant 2mm and 20mm.
Bug#315191: mozilla-firefox: fails to increase unreadably small text
On 6/24/05, Eric Dorland [EMAIL PROTECTED] wrote: Can you give me a page where this is the case? To my eye, all the text is increased in size, proportional to each other. reasonable request. a thought experiment applies to every page with varying font size. imagine: o the smallest font you can read is 4mm high o the largest font is 4cm high o the smallest font is 1mm high now double it. you get 2mm and 8cm fonts. the 2mm is not readable. the 8cm is less readable. double it again. you get 4mm and 16cm. 4mm is just readable. 16cm can be half the page. see what i mean? even if increase is proportional, this is highly troublesome. i erred in not pointing that out preemptively. one fix might be additive, not multiplicative scaling. if you double 1mm and 10mm, then you should get 2mm and 11mm instead of 2mm and 80mm. this is what i was referring to. the comment about small fonts sometimes not increasing at all is true sometimes, but not most of the time and was a red herring causing you to notreproducible. the large fonts currently increase the most. 1mm and 1cm becomes 2mm and 2cm, which is an increase of 1mm and 1cm. sometimes it is even worse if there are only a few large sizes available. the small font not increasing problem is not what i want to emphasize. it might be due to a minimum font size or a hardcoded element. hih
Bug#315331: zip warning: No such device or address does not specify pathname
i'm sorry, is that error message not in the source? i'd think it would be possible to change the printf string, or whatever, to ...: %s, pathname) or whatever. i have many thousands of pathnames piped to it, and i do not know which one is causing it. that is the point of requesting more information. it will take time to do the binary search to find the one that is causing the error. for 64000 files, i will have to run zip about 16 times. please tell me if you want me to do that, or if a grep on the source will find the error message instead. thanks.
Bug#315331: zip warning: No such device or address does not specify pathname
to reproduce, create a socket and pipe its pathname to zip -0qy - -@ somefile. i expected to see what file caused the error. the message did not contain it. that what you wanted?
Bug#311882: ipkungfu: script output not verbose enough/too verbose
i'd recommend debconf to ask the user modify firewall now? before doing anything upon installation. some people want to inspect the script first. it's a firewall.
Bug#311880:
see bug 311868. warning: this package will activate a firewall upon installation. i still do not know what ipkungfu did to my system. did it install anything that was not purged?
Bug#311880:
i got your i'm splitting message, but not the split bugs. why? and why am i not being cc'ed on updates to the bugs? my own bts cluelessness, no doubt. just didn't find docs and thought i had set reportbug to do that. thanks.
Bug#311880:
On 6/3/05, Nigel Jones [EMAIL PROTECTED] wrote: you have your reply-to's set to invalid addresses, as for what files: do you mean [EMAIL PROTECTED] if so, they are valid and get to me. gmail ignores the + and everything after it. they are useful for filtering. if bts ever leaks to spammers i am covered. any way to tell bts to understand this and send anyway? -rwxr-xr-x root/root 1868 2005-06-03 22:38:14 ./etc/init.d/ipkungfu am i right in assuming that is the only one that could cause problems if not purged? I have attached the new init.d file, so if the package is still installed, it will remove correctly. no, i purged it and the old init.d is not there. is that a prob? BTW: sorry i'm a little unorganized, i'll tell control off list to make sure your real email address is the submitter address (that way you will/should get CC'ed) (they are all real.) now how to do it for other bugs :-( thanks.
Bug#310805: gnupg: fully exportable armored homedir is completely impossible now!
here's another point that is closely related and along the way explains more about the general topic. in order to encrypt something, such as a regular backup, in such a way that the human need not be present, it is necessary to use asymmetric encryption, so that passphrases do not need to be supplied. because the secret key is necessary for decryption, however, it *must* be included along with the encrypted file. if you do not do this, you might lose your homedir and only have access to the encrypted file, which would be useless. when you have to keep track of 2 files, there is always the risk that one gets lost or changed. naturally, including the secret key is less secure than having the secret key separate, because an attacker can attempt brute forcing the pass phrase, but it is a reasonable tradeoff in limited circumstances such as certain types of regular backups. now there are 2 choices: either include the entire homedir with the encrypted file, or only include the secret key. technically, only the secret key is necessary -- now -- so it would seem more secure to only include the secret key. i do not know if this is true from a strict crypto perspective -- does having the other information in the homedir, such as the public key, help an attacker? but somebody might not want the other information there, for whatever reason, such as because there is revealing information in the public key ring. on the other hand, if you only include the secret key, there is a risk that a new version of gnupg will not be able to decrypt with only the secret key present. this is a risk from the user's perspective -- the user does not know what kinds of homedir assumptions gnupg will make in the future. it is a future-proofing issue. gpg developers are probably pretty confident one way or the other, but if gpg is potentially developed with the assumption that nobody will ever want to include only the secret key, then the user doesn't know whether some other part of the homedir will become necessary for decryption in a future version of gpg. so not only is there an issue of needing to be sure that gpg will properly serialize a homedir in such a way that it contains everything necessary for decryption, but there is also an issue of whether the secret key by itself, (whether it is serialized or merely copied from ~/.gnupg -- both must be possible), will remain sufficient for decryption. the general issue is more important, of course, because without it, it is impossible, while if you have to include the entire homedir, you are merely potentially incurring privacy leaks and you can work around them by creating a clean homedir if gpg lets you. i hope this helps. let me know if anything was not clear enough.
Bug#310805: gnupg: fully exportable armored homedir is completely impossible now!
severity: normal OK, with some kludging, we can finally get the old kludges to export and import the homedir. However, the following output still occurs, presumably because the key is no longer ultimately trusted, and -q fails to suppress that output, and it is not clear how to do --edit-key in batch mode. so we have to redirect stderr to /dev/null :-(. gpg: 3 marginal(s) needed, 1 complete(s) needed, PGP trust model gpg: depth: 0 valid: 1 signed: 0 trust: 0-, 0q, 0n, 0m, 0f, 1u trying to change severity to normal because it semi-works. can't find an explanation of the above output in /usr/share/doc or the man page. is it there? thanks.
Bug#310805: gnupg: fully exportable armored homedir is completely impossible now!
another example of unnecessary accessing of homedir is when you want to compress a file. when you do gpg --store, homedir access is unnecessary but gpg tries to access it (perhaps to find preferences?). however, there is no way to turn off its errors or have it simply default.
Bug#310805: gnupg: fully exportable armored homedir is completely impossible now!
first, thanks for replying. and thanks for monitoring the debian list. it gives me a warm feeling to know that this critical program is being maintained. On 5/26/05, Werner Koch [EMAIL PROTECTED] wrote: GnuPG requires is home directory for internal purposes and you should not fiddle around with it. We can't use /var/lib/gnupg because these i absolutely understand and agree. that is why i need an --export-homedir option to be certain that i do not unintentionally fiddle with the contents of that directory with my homegrown --export/--export-secret-key/... attempts. i am especially concerned that those commands are, or will be in the future, insufficient to export the homedir exactly. you see, i need the script to contain everything gpg might need on the other end. if a user of the script has a corrupted .gnupg, it will not work. but if i include it in the script, everything will be ok. Using this data is on your own risk. There is a documented interface on how to access it (--export/--import etc.). You are on your own if does that documented interface export the homedir without leaving anything out or changing anything? will it after 1.9? my problem is that currently i have to use 3 export commands (pub, sec, trust) to even attempt this. and i fear 1.9 now. will i have to use 4 or more commands to export it? will it be exportable? the rest of this message is a reply to your thoughtful comments. the above is my biggest issue. BTW, never ever look at any data printed for human reading - scripts need to parse the --status-fd and --with-colon output. For thanks for the pointer. i am not currently having the script look at human-readable output -- it just passes it to the human or log. however, it should bail at problems. i assume (and hope) that you consider the output of --armor --export, etc. to be machine-readable. controlling gpg aside from this, the --command-fd/--status-fd interface is the way to go. Please don't use canned commands but parse the requests seen on --status-fd before sending commands to gpg - if there is something you can't cope with: bail out. Note that in most cases using the default response (which is sending an empty string) leads to proper results. i will use those if i ever need to parse output. now all i do is verify the resulting encrypted file manually occasionally. btw, i found no gpg command that will verify that a file was properly encrypted after it has been created. that would be nice after a pipeline. i suppose the solution is to try to decrypt, then exit successfully when it wants a pass phrase. i notice that --with-key-data is documented to print the public key data. i assume it just means the meta data -- not the key itself. is that right? for documentation i am using the gnupg package and the gnupg-doc package. if it is the key itself, there appears to be no equivalent for secret keys, trust db, etc. Salam-Shalom, indeed.
Bug#310514: reportbug: bts vs. upstream issues
On 5/23/05, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: 2. the bts should make it easy to cc: upstream o single interface o fewer debian bugs sent to upstream o fewer upstream bugs not sent to upstream
Bug#306693: cpio: allows extracting insecure pathnames (leading slash = / and dotdot = ..)
Tags: security Severity: grave
Bug#307419: mozilla-firefox: firefox allows paragraphs wider than the screen, requiring horizontal scrolling for each line
On 5/4/05, Eric Dorland [EMAIL PROTECTED] wrote: It's UNIX tradition that straight ascii files do their own line wrapping. I'm not powerful enough yet to defy that tradition. Thanks for replying. Wow! Little did I know that you couldn't fill paragraphs easily in emacs or vi without affecting other documents, unless you had power. Is it UNIX tradition to have a browser that can't use pipes easily? Is Firefox UNIX-only? Does the UNIX world require power to let the user do something?
Bug#307107: workrave: forces breaks shortly after inactivity
On 5/2/05, Michael Piefel [EMAIL PROTECTED] wrote: Your breaks do count. Look closely at the timers. As soon as you don't type, they start filling up in green. If your non-activity is long enough, it counts as a break for workrave. seems to work when i watch it, at least when suspend timer when inactive is checked, but there seem to be times when i am gone for a long time then 30s later get a microbreak. i suppose this isn't very helpful to you. i tried simulating spontaneous mouse movements due to desk vibration, but couldn't get the timers to misbehave. still, that might be the cause. or interaction with the screensaver? btw, tooltips for checkboxes and timers might help users interpret the ui.
Bug#307419: mozilla-firefox: firefox allows paragraphs wider than the screen, requiring horizontal scrolling for each line
Please allow user to specify paragraph widths. Thanks.
Bug#306904:
Clarification: scriptability wants to call a command without having to create the environment for the command, such as creating dot files or requiring user input. Thus, the first line wishlist item. Wishlist because can kludge the dot file.
Bug#306904: msmtp: new configuration can't script password
Severity: wishlist It is possible to do using config file kludge, so it is merely a wishlist.
Bug#306904: Severity: wishlist
It is possible to do using config file kludge, so it is severity wishlist.
Bug#306697: coreutils: inconsistent behavior from ls and ls -l
This is hilarious -- a program whose basic form has existed since around the first moon landing has either you or me very confused. On 4/27/05, Bob Proulx [EMAIL PROTECTED] wrote: #ls as af af is the contents of the directory. #ls -l as lrwxrwxrwx 1 root root 2 Apr 27 19:38 as - ad as is not the contents of the directory. Atari! It is supposed to behave that way. But your statement confuses me. What other way would you expect it to behave? What documentation and for which program of mkdir, touch, ln or ls would you change? I would change ls as to list the symlink, and document ls as/. Or I would change ls -l as to list the directory and get rid of the slash trick. It's miai. Another option is to document that ls and ls -l are nonorthogonal for reasons that are mysterious. :-) This is a serious question. The above is in exactly the order of your example. What did you expect? What was unclear? Lemme throw that back at you! Is this a ko? Bob Proulx [EMAIL PROTECTED] Nice to meet you. P.S. to others: atari, miai, and ko are just jokes for Bob; please ignore them.
Bug#306697: coreutils: inconsistent behavior from ls and ls -l
On 4/28/05, Bob Proulx [EMAIL PROTECTED] wrote: #ls as af #ls -l as lrwxrwxrwx 1 root root 2 Apr 27 19:38 as - ad Note that af (a file) does not equal as (a symlink). Let's look at the man page: -l use a long listing format Note that it says nothing about not dereferencing symlinks. IMHO it should, or the standard should be changed. ls as af ls -F as as@ Wow, that is wrong also :-). -F is documented: -F, --classify append indicator (one of */=@|) to entries Again it says nothing about not dereferencing symlinks. That would break standards conformance and existing practice. That You're right. I am not going to change the standard, so please document -l and -F more completely. :-) i.e., please say that they do not dereference symlinks. I did not know that the standard was (IMHO of course) broken. Thanks. around the first moon landing has either you or me very confused. But symlinks were added to the system much later. They were not Not surprising, although I did not know that. I meant it hyperbolically, not literally, actually. Misleading on my part. This is not an either this or that situation. It is not miai. It is to somebody who thinks that long listings and changing whether dereferencing is done should be orthogonal options. Although my view is nonstandard, having ls and ls -l both dereference or both not dereference are orthogonal solutions. My honorable opponent (the standard) had one changed (in my view, from what it should be) but not the other (which one it is depends on whether you favor dereferencing both or not dereferencing both). I must now go in the place where my opponent did not, by fixing the standard in either direction. We call that miai. Yes, one direction is probably better than the other, but it was just a joke (as I documented :-)). Is miai always exactly symmetrical, anyway? I do know about the other related options. This is not a Debian bug. I suggest this bug be closed. Even if you end up agreeing with me that ls -l should mention that it doesn't dereference? (Not a rhetorical question.)
Bug#306693: cpio: allows extracting insecure pathnames (leading slash = / and dotdot = ..)
severity: important After looking at the severities of other cpio bug reports that have been around for hundreds of days, I concluded that this should be important instead of normal.
Bug#306693: cpio: allows extracting insecure pathnames (leading slash = / and dotdot = ..)
tags: security The docs suggest grave or critical for security bugs, but I'm not sure whether that is appropriate.
Bug#306693: cpio: allows extracting insecure pathnames (leading slash = / and dotdot = ..)
P.P.S. I found a more subtle security hole. It is even more dangerous. /tmp/aaa$ mkdir ../b /tmp/aaa$ ln -s ../b b /tmp/aaa$ touch ../b/trojan /tmp/aaa$ ls b trojan /tmp/aaa$ find b b/trojan b b/trojan /tmp/aaa$ find b b/trojan | cpio -o dangerous cpio: b: truncating inode number cpio: b/trojan: truncating inode number 1 block /tmp/aaa$ /bin/rm -v b/trojan b removed `b/trojan' removed `b' /tmp/aaa$ ls dangerous /tmp/aaa$ cpio -tdangerous b b/trojan 1 block /tmp/aaa$ cpio -vtdangerous lrwxrwxrwx 1 kpc kpc 4 Apr 27 19:46 b - ../b -rw--- 1 kpc kpc 0 Apr 27 19:46 b/trojan 1 block Notice that grep '\.\.' on the output of cpio -t would not find the relative pathname. You have to use cpio -vt. Now watch this: /tmp/aaa$ cpio -idangerous 1 block /tmp/aaa$ ls b dangerous /tmp/aaa$ ls ../b trojan IMHO cpio should disallow this by default. Imagine ../../../../../../../etc/cron.daily again. cpio should check for extracting in directories that are not below pwd, even if it is via indirect means such as a symlink. Wow!
