[gentoo-dev] xml2 use flag deprecation - ping (Bug 116346)
This is just a friendly reminder that Bug 116346 doesn't seem to be moving much. :P The two months old list seems still almost fully valid. http://bugs.gentoo.org/attachment.cgi?id=79178action=view Do we manage to kill the flag? Always good to have one redundant flag less... ;) Thanks. -- jakub signature.asc Description: OpenPGP digital signature
[gentoo-dev] gtk2 use flag must die
This is a (not-so happy) reminder that the agony of gtk2 use flag will have been lasting for half a year soon. It *really* needs to die. For affected ebuilds, please see the attached list and Bug 106560. Thanks. -- jakub app-crypt/pinentry-0.7.2 app-editors/mp-3.3.12 app-editors/mp-3.3.14 app-emulation/fuse-0.6.2.1 app-emulation/fuse-0.7.0 app-i18n/poedit-1.2.3 app-i18n/poedit-1.2.4 app-i18n/poedit-1.2.5 app-i18n/poedit-1.3.0 app-i18n/poedit-1.3.1 app-i18n/poedit-1.3.2 app-pda/jpilot-0.99.7-r1 app-pda/jpilot-0.99.8 app-pda/jpilot-0.99.8_pre9 app-pda/jpilot-backup-0.50 app-pda/jpilot-plucker-0.01 app-pda/jpilot-syncmal-0.72.1 dev-python/wxpython-2.4.2.4 dev-python/wxpython-2.4.2.4-r3 dev-python/wxpython-2.6.0.0-r1 dev-python/wxpython-2.6.1.0 dev-ruby/wxruby-0.6-r1 dev-scheme/bigloo-lib-0.17 media-gfx/xsane-0.97 media-gfx/zphoto-1.2-r1 media-libs/blib-1.1.7 media-sound/audacity-1.2.3-r1 media-sound/aumix-2.8-r2 media-sound/gamix-1.99_p14-r1 media-sound/timidity++-2.13.2-r2 media-video/blinkensim-2.7 media-video/ogle-gui-0.9.2 media-video/vlc-0.8.1-r1 net-ftp/gftp-2.0.18-r1 net-p2p/xmule-1.10.0 net-p2p/xmule-1.10.1 net-p2p/xmule-1.8.4-r1 sci-chemistry/chemtool-1.6.4 sci-chemistry/chemtool-1.6.6 sci-chemistry/chemtool-1.6.7 x11-libs/gtk-server-2.0.5 x11-libs/wxGTK-2.4.2-r2 x11-libs/wxGTK-2.4.2-r3 x11-libs/wxGTK-2.4.2-r4 x11-libs/wxGTK-2.5.3 x11-libs/wxGTK-2.6.0-r1 x11-libs/wxGTK-2.6.1 x11-libs/wxGTK-2.6.1-r1 x11-misc/linuxwacom-0.6.7 x11-misc/linuxwacom-0.6.8 x11-misc/linuxwacom-0.6.9 x11-misc/linuxwacom-0.7.2 x11-misc/macopix-1.0.4 x11-misc/macopix-1.2.1 x11-plugins/i8krellm-2.3 x11-plugins/i8krellm-2.5 x11-themes/gtk-engines-qtpixmap-0.28-r1 x11-wm/aewm-1.2.3 x11-wm/fvwm-2.5.12 x11-wm/fvwm-2.5.13-r1 signature.asc Description: OpenPGP digital signature
Re: [gentoo-dev] gtk2 use flag must die
On Sunday 02 April 2006 12:05, Jakub Moc wrote: This is a (not-so happy) reminder that the agony of gtk2 use flag will have been lasting for half a year soon. It *really* needs to die. too bad it doesnt address packages which still legitimately utilize gtk/gtk2 i for one wont be fixing these packages anytime soon -mike -- gentoo-dev@gentoo.org mailing list
[gentoo-dev] virtual/acl
We have a fair number of packages in the tree (57 someone said, but a non-trivial number) which depend upon sys-apps/acl for ACL support. Since the packages needed for this differ between platforms (sys-apps/acl is for linux only), if noone has any reasonable objections I will be adding a (new-style) virtual/acl for these packages sometime in the next week or so. This email brought to you on behalf of Flameeyes. -- gentoo-dev@gentoo.org mailing list
Re: [gentoo-dev] virtual/acl
On Sun, 2 Apr 2006 18:26:27 +0100 Stephen Bennett [EMAIL PROTECTED] wrote: | We have a fair number of packages in the tree (57 someone said, but a | non-trivial number) which depend upon sys-apps/acl for ACL support. | Since the packages needed for this differ between platforms | (sys-apps/acl is for linux only), if noone has any reasonable | objections I will be adding a (new-style) virtual/acl for these | packages sometime in the next week or so. Good. This'll move the icky deps to one place rather than all over the tree, at least. As an aside, and not directed at spb... Would be nice if in the future this kind of discussion were done *before* packages start to be changed in weird and wonderful ways. It's generally considered bad manners to start screwing around with other people's packages without asking... Plus there's the whole learn from past mistakes issue. -- Ciaran McCreesh : Gentoo Developer (Wearer of the shiny hat) Mail: ciaranm at gentoo.org Web : http://dev.gentoo.org/~ciaranm signature.asc Description: PGP signature
Re: [gentoo-dev] virtual/acl
Ciaran McCreesh wrote: On Sun, 2 Apr 2006 18:26:27 +0100 Stephen Bennett [EMAIL PROTECTED] wrote: | We have a fair number of packages in the tree (57 someone said, but a | non-trivial number) which depend upon sys-apps/acl for ACL support. | Since the packages needed for this differ between platforms | (sys-apps/acl is for linux only), if noone has any reasonable | objections I will be adding a (new-style) virtual/acl for these | packages sometime in the next week or so. Good. This'll move the icky deps to one place rather than all over the tree, at least. As an aside, and not directed at spb... Would be nice if in the future this kind of discussion were done *before* packages start to be changed in weird and wonderful ways. It's generally considered bad manners to start screwing around with other people's packages without asking... Plus there's the whole learn from past mistakes issue. Drop it. -Antarus -- gentoo-dev@gentoo.org mailing list
Re: [gentoo-dev] virtual/acl
On Sun, 02 Apr 2006 13:41:25 -0400 Ned Ludd [EMAIL PROTECTED] wrote: In your own words what benefit does this have over kernel_linux? ( acl? ( sys-apps/acl )) It moves all of the platform-conditional voodoo into one place, which helps maintainability and will greatly reduce the work involved in adding a new port that may use a different package for ACL support. We had a lengthy discussion on this topic between the *BSD and portage teams some time last year, and the consensus was that the correct solution was to move all the deps that change between platforms into one place (the new-style virtual category). I haven't seen anything change that would make for a different conclusion were the same discussion to happen now. Especially when more packages appear which are widely depended upon and change between platforms, and more platforms are added, this approach will greatly reduce the maintenance burden that Gentoo/Alt projects place upon the main tree. -- gentoo-dev@gentoo.org mailing list
Re: [gentoo-dev] virtual/acl
On Sunday 02 April 2006 14:04, Diego 'Flameeyes' Pettenò wrote: Thus my request to zmedico to have a per-package use.mask so that we can mask the flag for the packages that can use only the sys-apps/acl interface. or you can be the hero and fix the packages ;) -mike -- gentoo-dev@gentoo.org mailing list
Re: [gentoo-dev] gtk2 use flag must die
On Sun, 2006-02-04 at 13:08 -0400, Mike Frysinger wrote: On Sunday 02 April 2006 12:05, Jakub Moc wrote: This is a (not-so happy) reminder that the agony of gtk2 use flag will have been lasting for half a year soon. It *really* needs to die. too bad it doesnt address packages which still legitimately utilize gtk/gtk2 Is there a legitimate reason to use gtk1 if the gtk2 support is useable? Either way, there shouldnt be a gtk2 flag.. -- Olivier Crête [EMAIL PROTECTED] signature.asc Description: This is a digitally signed message part
Re: [gentoo-dev] virtual/acl
On Sunday 02 April 2006 20:16, Mike Frysinger wrote: or you can be the hero and fix the packages ehehe yeah I'll do when i'll have more time, like i did for binutils ;) By the way, that patch? :P -- Diego Flameeyes Pettenò - http://dev.gentoo.org/~flameeyes/ Gentoo/Alt lead, Gentoo/FreeBSD, Video, AMD64, Sound, PAM, KDE pgpTtfMvCqDE6.pgp Description: PGP signature
Re: [gentoo-dev] gtk2 use flag must die
On Sunday 02 April 2006 14:22, Olivier Crête wrote: On Sun, 2006-02-04 at 13:08 -0400, Mike Frysinger wrote: On Sunday 02 April 2006 12:05, Jakub Moc wrote: This is a (not-so happy) reminder that the agony of gtk2 use flag will have been lasting for half a year soon. It *really* needs to die. too bad it doesnt address packages which still legitimately utilize gtk/gtk2 Is there a legitimate reason to use gtk1 if the gtk2 support is useable? nothing personal, but who are you to say whether it's legit ? if a package provides the interfaces and users like to use them, where's the bug ? ive seen peeps from time to time who bring the hate on gtk2 because of its fattiness compared to gtk1 ... they like to build packages that still can be against gtk1 -mike -- gentoo-dev@gentoo.org mailing list
Re: [gentoo-dev] [last rites] media-gfx/sodipodi
On Saturday 01 April 2006 22:52, Mark Loeser wrote: Yes, there is. It's slowing down the process, getting into the flow. Waiting 30 days is a lot of time. A regular user does not necessarily follow the dev-gentoo mailing list and it doesn't matter for him, if the package is masked or removed. First of all I'm not a dev but I do read the dev mailing list exactly because I want to know whats going on because changes to the tree normally really hit me without warning. The apache config layout change hit me when I needed to upgrade and had not the time to mess with configuration issues. Thats because I am not the typical gentoo-i-have-to-sync-every-30-seconds user. The last time I synced was about 3 weeks ago. On my server its even more than that. I generally sync when I need a new version of something because I know of a bugfix/feature I need or when I see a GLSA (or Bugtraq posting by some other distro/vuln researcher and the gentoo package is already fixed without a GLSA being out). That being said, changes in package naming/categorization or configuration layout (think apache mess), is by far worse than an old package being removed. For a dev 30 days may be a long time indeed. Because everyone is sitting in anticipation of the package being removed? Mask the package, and go about your life as if it was gone. Then in a month when you remember about it again, remove it. What does the month give you? Nothing. If someone was trying to update his sodipodi manually he would mostly get nothing because upstream was dead. Masking helps during that 30 days, after that you don't know whats going on either. And considering that upstream is dead for about a year I think most people will not try to update that package every 2 days or something like that. By your logic, we should do away with the entire masking process and just remove stuff when we like, and that would just lead to users filing lots of bugs asking where their package went. How about removing the package (its dead anyway) BUT keeping a message for all those trying to update. I don't know if that is even possible but it would be good to keep that message for more than 30 days. Everyone used to wait the month, but lately it seems like no one can ignore the package for that long after putting it in p.mask. I totally agree with you on that. If 30 days are the documented period, then devs have to stick to it. If it is too long for them: Change the documented time period first! At least that would be professinal. Alex -- gentoo-dev@gentoo.org mailing list
Re: [gentoo-dev] [last rites] media-gfx/sodipodi
Alexander Gretencord wrote: And considering that upstream is dead for about a year I think most people will not try to update that package every 2 days or something like that. Most people upgrade the whole system at once - those would see a warning about masked package. Cheers, -jkt -- cd /local/pub more beer /dev/mouth signature.asc Description: OpenPGP digital signature
Re: [gentoo-dev] gtk2 use flag must die
On Sunday 02 April 2006 20:41, Mike Frysinger wrote: nothing personal, but who are you to say whether it's legit ? It's really not a question what's legit (heck, you started using this term, so blaming Olivier for using it is a bit odd), but what we (can and want to) support. Wouldn't it have been time for you to speak up, when the Gnome herd announced to deprecate Gtk1 support for applications that build again Gtk2!? Instead playing the road block for the very few people who may still favor Gtk1, it should be more the question, if there's anyone supporting Gtk1 upstream with regards to security issues etc.. I for one favor it to eliminate toolkits that are superseeded by their successor as fast as possible. Carsten pgpbG5Bik6Ctb.pgp Description: PGP signature
Re: [gentoo-dev] gtk2 use flag must die
On Sunday 02 April 2006 15:02, Jakub Moc wrote: Mike Frysinger wrote: On Sunday 02 April 2006 14:22, Olivier Crête wrote: On Sun, 2006-02-04 at 13:08 -0400, Mike Frysinger wrote: On Sunday 02 April 2006 12:05, Jakub Moc wrote: This is a (not-so happy) reminder that the agony of gtk2 use flag will have been lasting for half a year soon. It *really* needs to die. too bad it doesnt address packages which still legitimately utilize gtk/gtk2 Is there a legitimate reason to use gtk1 if the gtk2 support is useable? nothing personal, but who are you to say whether it's legit ? Oh please... no flames :/ it isnt a flame, hence the nothing personal expecting Olivier to know every single gtk1/gtk2 package inside and out is asinine if a package provides the interfaces and users like to use them, where's the bug ? ive seen peeps from time to time who bring the hate on gtk2 because of its fattiness compared to gtk1 ... they like to build packages that still can be against gtk1 Except that you won't get any support whatsoever for gtk-1 upstream, and except that it's been agreed upon quite some time ago that gtk vs. gtk2 use flags is a bad way to select between gtk-1/gtk-2/no gtk at all and confuses users, except that... oh well, been discussed so many times, not going to beat a dead horse. and if there are no bugs filed ? this sort of stance is like the lets remove packages from portage because upstream is dead ... it benefits no one Please, remove the gtk2 flag from ebuilds you maintain. Thanks. no -mike -- gentoo-dev@gentoo.org mailing list
Re: [gentoo-dev] [last rites] media-gfx/sodipodi
On Sunday 02 April 2006 04:48, Daniel Goller wrote: exactly, what's the point of removing it so fast? give people a chance to miss it, it does not matter if it's removed or masked only as far as going woah, what? and if masked it is a matter of unmasking rather than recommitting We haven't had a single issue with the usual seven day period as far as I can remember, so please come up with a valid argument against it, instead assuming turning my argument would be one. in short, if it's slowing down the process, why do you need it to be quick in the first place? Getting the junk out of tree and mind as fast as possible is a value in itself. Carsten pgpM3Qs9oescc.pgp Description: PGP signature
Re: [gentoo-dev] gtk2 use flag must die
On Sunday 02 April 2006 15:12, Carsten Lohrke wrote: On Sunday 02 April 2006 20:41, Mike Frysinger wrote: nothing personal, but who are you to say whether it's legit ? It's really not a question what's legit (heck, you started using this term, so blaming Olivier for using it is a bit odd) i didnt blame Olivier for anything ... doesnt anyone get the nothing personal part ? but what we (can and want to) support. i dont recall asking you to support my packages Wouldn't it have been time for you to speak up, when the Gnome herd announced to deprecate Gtk1 support for applications that build again Gtk2!? Instead playing the road block for the very few people who may still favor Gtk1, last time i recall following the gtk/gtk2 stuff, the idea was that in the future to move to a gtk/gtk1 situation ... but this was back when Spider was The Man, so i guess people forgot about that it should be more the question, if there's anyone supporting Gtk1 upstream with regards to security issues etc.. and when such a situation arises, the solution may to simply drop the optional support. such a situation has not arose, so using such hypothetical examples is meaningless. -mike -- gentoo-dev@gentoo.org mailing list
Re: [gentoo-dev] [last rites] media-gfx/sodipodi
On Sun, 2 Apr 2006 21:20:21 +0200 Carsten Lohrke [EMAIL PROTECTED] wrote: | We haven't had a single issue with the usual seven day period The usual period is thirty days. | in short, if it's slowing down the process, why do you need it to be | quick in the first place? | | Getting the junk out of tree and mind as fast as possible is a value | in itself. Once it's in p.mask it's effectively gone, to the extent that ignoring it for a month is fine. -- Ciaran McCreesh : Gentoo Developer (Wearer of the shiny hat) Mail: ciaranm at gentoo.org Web : http://dev.gentoo.org/~ciaranm signature.asc Description: PGP signature
Re: [gentoo-dev] gtk2 use flag must die
Mike Frysinger wrote: and if there are no bugs filed ? this sort of stance is like the lets remove packages from portage because upstream is dead ... it benefits no one No bugs filed? Well, just search the archives of this ML, and search bugzilla for all those bugs about portage pulling in gtk2 when users had USE=-gtk2 set, or for all bugs about the confusion what USE=gtk -gtk2 or USE=-gtk gtk2 actually means... Please, remove the gtk2 flag from ebuilds you maintain. Thanks. no -mike Shrug... Maybe speak up sooner next time, this debate has been over for a long time and the decision was clearly to deprecate gtk2 use flag. Not going to do that? Shrug, oh well, perhaps QA will then, or not - I don't care if you are going to confuse users. That bug is set as a blocker for Gnome 2.14 release. My email was intended to speed up closing that bug, not to start such debate again. So - that's all from me. -- jakub signature.asc Description: OpenPGP digital signature
Re: [gentoo-dev] gtk2 use flag must die
On Sun, Apr 02, 2006 at 09:12:28PM +0200, Carsten Lohrke wrote: On Sunday 02 April 2006 20:41, Mike Frysinger wrote: nothing personal, but who are you to say whether it's legit ? It's really not a question what's legit (heck, you started using this term, so blaming Olivier for using it is a bit odd), but what we (can and want to) support. Wouldn't it have been time for you to speak up, when the Gnome herd announced to deprecate Gtk1 support for applications that build again Gtk2!? Others did speak up at that time. The result: http://article.gmane.org/gmane.linux.gentoo.devel/31641 -- gentoo-dev@gentoo.org mailing list
Re: [gentoo-dev] gtk2 use flag must die
On Sunday 02 April 2006 21:51, Harald van Dijk wrote: Others did speak up at that time. The result: http://article.gmane.org/gmane.linux.gentoo.devel/31641 Yeah, that was the one and only single voice. Carsten pgplFkefqq6Ma.pgp Description: PGP signature
Re: [gentoo-dev] gtk2 use flag must die
On Sunday 02 April 2006 15:34, Jakub Moc wrote: Mike Frysinger wrote: and if there are no bugs filed ? this sort of stance is like the lets remove packages from portage because upstream is dead ... it benefits no one No bugs filed? Well, just search the archives of this ML, and search bugzilla for all those bugs about portage pulling in gtk2 when users had USE=-gtk2 set, or for all bugs about the confusion what USE=gtk -gtk2 or USE=-gtk gtk2 actually means... unrelated i'm talking about buts in the packages themselves, not end user confusion -mike -- gentoo-dev@gentoo.org mailing list
Re: [gentoo-dev] gtk2 use flag must die
On Sunday 02 April 2006 21:28, Mike Frysinger wrote: last time i recall following the gtk/gtk2 stuff, the idea was that in the future to move to a gtk/gtk1 situation ... but this was back when Spider was The Man, so i guess people forgot about that No, see the whole thread Harald references in his email. it should be more the question, if there's anyone supporting Gtk1 upstream with regards to security issues etc.. and when such a situation arises, the solution may to simply drop the optional support. such a situation has not arose, so using such hypothetical examples is meaningless. The problem is that no one is working on the code, so the chances for black hats to find something they can abuse for a long time are to consider. The situation is always given. It's just the question, when and if the good guys get to know about it. Carsten pgpc1PEkIjwrp.pgp Description: PGP signature
Re: [gentoo-dev] gtk2 use flag must die
On Sun, Apr 02, 2006 at 10:00:25PM +0200, Carsten Lohrke wrote: On Sunday 02 April 2006 21:51, Harald van D??k wrote: Others did speak up at that time. The result: http://article.gmane.org/gmane.linux.gentoo.devel/31641 Yeah, that was the one and only single voice. On gentoo-dev. I'd try to show more if I had IRC logs that far back. -- gentoo-dev@gentoo.org mailing list
Re: [gentoo-dev] [last rites] media-gfx/sodipodi
On Sunday 02 April 2006 21:31, Ciaran McCreesh wrote: The usual period is thirty days. Grep this mailing list, most often a one week period was used. Once it's in p.mask it's effectively gone, to the extent that ignoring it for a month is fine. Who said a package gets masked before it gets removed? There is no such requirement in the ebuild policy. Carsten pgpFHPt6EHSxi.pgp Description: PGP signature
Re: [gentoo-dev] [last rites] media-gfx/sodipodi
Carsten Lohrke wrote: Who said a package gets masked before it gets removed? There is no such requirement in the ebuild policy. Come on. Is this a 'policy doesn't say I have to be sane' war? It's absolutely reasonable to p.mask a package that is pending for removal. That way you give the users a timeframe which they can search for alternative tools in. I don't know whether policy does state this or not, I don't care. It's not like you would get any bugs for a masked package. It's not like you would gain a lot of space because you freed up 3 ebuilds and a few digests. It's not like you would gain anything from removing it immediately. But those who use the package do gain a lot from you giving them a hint to search for alternatives. -- Kind Regards, Simon Stelling Gentoo/AMD64 Developer -- gentoo-dev@gentoo.org mailing list
Re: [gentoo-dev] [last rites] media-gfx/sodipodi
On Sun, 2 Apr 2006 22:20:49 +0200 Carsten Lohrke [EMAIL PROTECTED] wrote: | On Sunday 02 April 2006 21:31, Ciaran McCreesh wrote: | The usual period is thirty days. | | Grep this mailing list, most often a one week period was used. This is a recent change, and usually someone replies with why not a month?. | Once it's in p.mask it's effectively gone, to the extent that | ignoring it for a month is fine. | | Who said a package gets masked before it gets removed? There is no | such requirement in the ebuild policy. It's not a requirement. It's a courtesy to your users and fellow developers, at least some of whom would very much appreciate it if you left things for ~four weeks rather than ~one. -- Ciaran McCreesh : Gentoo Developer (Wearer of the shiny hat) Mail: ciaranm at gentoo.org Web : http://dev.gentoo.org/~ciaranm signature.asc Description: PGP signature
Re: [gentoo-dev] gtk2 use flag must die
On Sunday 02 April 2006 16:09, Carsten Lohrke wrote: On Sunday 02 April 2006 21:28, Mike Frysinger wrote: it should be more the question, if there's anyone supporting Gtk1 upstream with regards to security issues etc.. and when such a situation arises, the solution may to simply drop the optional support. such a situation has not arose, so using such hypothetical examples is meaningless. The problem is that no one is working on the code, so the chances for black hats to find something they can abuse for a long time are to consider. The situation is always given. It's just the question, when and if the good guys get to know about it. lets apply the same logic to all things unmaintained ! besides, you're talking about removing GTK1 completely ... this thread is talking about deprecating the gtk2 USE flag the GTK1 lib is hopefully going to live on much much longer than the gtk2 USE flag -mike -- gentoo-dev@gentoo.org mailing list
Re: [gentoo-dev] [last rites] media-gfx/sodipodi
On Sunday 02 April 2006 22:33, Ciaran McCreesh wrote: This is a recent change, and usually someone replies with why not a month?. This is simply not true or we have very different ideas of the meaning of recent. The vast majority of last rites emails from 2005 had slated removals of one week or less. It's not a requirement. It's a courtesy to your users and fellow developers, at least some of whom would very much appreciate it if you left things for ~four weeks rather than ~one. I don't see the necessity for devs and users would have to look at the package.mask file regularly to get the information that a package is masked. If Portage would be that smart to spit out the relevant information on emerge --sync, a longer period would probably make sense. Carsten pgp0vBYtK5V7j.pgp Description: PGP signature
Re: [gentoo-dev] gtk2 use flag must die
Mike Frysinger wrote: On Sunday 02 April 2006 15:34, Jakub Moc wrote: Mike Frysinger wrote: and if there are no bugs filed ? this sort of stance is like the lets remove packages from portage because upstream is dead ... it benefits no one No bugs filed? Well, just search the archives of this ML, and search bugzilla for all those bugs about portage pulling in gtk2 when users had USE=-gtk2 set, or for all bugs about the confusion what USE=gtk -gtk2 or USE=-gtk gtk2 actually means... unrelated i'm talking about buts in the packages themselves, not end user confusion -mike Not really, since the retarded handling of those use flags combos in ebuilds has been one of the key reasons to deprecate this. Also, this is rather funny why xml2 use flags needs to be removed (it was you who suggested that) - yet you resist to get rid of gtk2 use flag, which is *way* more confusing (heck, look at that old wxGTK stuff, it's nasty and confusing like hell). I'd like to note that all bugs complaining about gtk2 getting pulled in with USE=-gtk2 as being marked as dupe of that deprecation bug, not really sure how are you going to explain to the users that there are exceptions to this rule. -- jakub signature.asc Description: OpenPGP digital signature
Re: [gentoo-dev] [last rites] media-gfx/sodipodi
On Sunday 02 April 2006 22:29, Simon Stelling wrote: Come on. Is this a 'policy doesn't say I have to be sane' war? It's absolutely reasonable to p.mask a package that is pending for removal. That way you give the users a timeframe which they can search for alternative tools in. This is not the case. At least unless the user actively looks at package.mask. Since Portage doesn't provide the information, this point is void. And even if - four weeks are a too long, imho. Carsten pgpnCZy5so6UC.pgp Description: PGP signature
Re: [gentoo-dev] [last rites] media-gfx/sodipodi
Carsten Lohrke wrote: I don't see the necessity for devs and users would have to look at the package.mask file regularly to get the information that a package is masked. If Portage would be that smart to spit out the relevant information on emerge --sync, a longer period would probably make sense. Not that I'd care so much whether it's a week or a month (IMO individual depending on ebuild in question) - so just a technical note. Portage 2.1 *does* spit out the relevant info. # emerge -uDpv world These are the packages that would be merged, in order: Calculating world dependencies / !!! Packages for the following atoms are either all !!! masked or don't exist: net-ftp/glftpd ... done! # esearch glftpd [ Results for search key : glftpd ] [ Applications found : 1 ] * net-ftp/glftpd [ Masked ] Latest version available: 2.01 Latest version installed: 1.32 -- jakub signature.asc Description: OpenPGP digital signature
Re: [gentoo-dev] gtk2 use flag must die
On Sunday 02 April 2006 22:40, Mike Frysinger wrote: lets apply the same logic to all things unmaintained ! Yes, that's one reason I am so annoyed of the unmaintained parts of the tree. besides, you're talking about removing GTK1 completely ... this thread is talking about deprecating the gtk2 USE flag Well, from my POV - beep/. You could at least change your packages to use gtk2 by default to be consistent with the other packages and add a (local) gtk1 use flag, so we can get rid of the gtk2 flag. Carsten pgpEAUXX0myA2.pgp Description: PGP signature
Re: [gentoo-dev] [last rites] media-gfx/sodipodi
On 4/2/06, Carsten Lohrke [EMAIL PROTECTED] wrote: This is not the case. At least unless the user actively looks at package.mask. Since Portage doesn't provide the information, this point is void. And even if - four weeks are a too long, imho. I still do not understand what the rush is with removing a package. Readding a package if necessary will be much more troublesome than just keeping it masked for a month. I believe this is the general consensus on the subject. Marcelo -- Marcelo Góes [EMAIL PROTECTED] [EMAIL PROTECTED] -- gentoo-dev@gentoo.org mailing list
Re: [gentoo-dev] gtk2 use flag must die
On Sun, 2006-04-02 at 23:20 +0200, Jakub Moc wrote: Mike Frysinger wrote: On Sunday 02 April 2006 15:34, Jakub Moc wrote: Mike Frysinger wrote: and if there are no bugs filed ? this sort of stance is like the lets remove packages from portage because upstream is dead ... it benefits no one No bugs filed? Well, just search the archives of this ML, and search bugzilla for all those bugs about portage pulling in gtk2 when users had USE=-gtk2 set, or for all bugs about the confusion what USE=gtk -gtk2 or USE=-gtk gtk2 actually means... unrelated i'm talking about buts in the packages themselves, not end user confusion -mike Not really, since the retarded handling of those use flags combos in ebuilds has been one of the key reasons to deprecate this. Also, this is rather funny why xml2 use flags needs to be removed (it was you who suggested that) - yet you resist to get rid of gtk2 use flag, which is *way* more confusing (heck, look at that old wxGTK stuff, it's nasty and confusing like hell). wxGTK-2.6* is fully supported upstream as linked against gtk1 and gtk2. However, that only is for things that exist already in case of gtk1 - we can't implement everything for gtk1 due to it not being as feature-rich as gtk2, and there is no point in implementing new features for gtk1 too, if it's easy with gtk2. The official support for gtk1 will end with the development cycle of 2.7.x, and won't be considered a fully supported port with the stable release of 2.8.x either. gtk1 code has been separated into a different port, duplicating some of the common gtk code. wxGTK1 (wxGTK built against GTK+-1.2) is essentially at the same level as wxMotif, wxX11 and other similar community driven ports who don't get the full manpower of developers. wxGTK2 however lives a happy life, free from the gtk1 burden. So, for wxGTK-2.8.x whenever it's out (a year?) there is no USE flag confusion - one port builds only against one major version of gtk2 (until gtk3 comes out, see Project Ridley). There should be a different package for those that wish the wxGTK1 port, if that package is needed at all. This is also the way that I would suggest handling wxGTK today (2.4 and 2.6). Have two different ebuilds, nuke that crazy no_wxgtk1 USE flag (and perhaps have a USE flag for building non-unicode version together with unicode version instead), use gtk instead of gtk2 flag and call it a day. Name em wxGTK and wxGTK1 for example, wxGTK1 perhaps even package.mask'ed - not as feature-rich, more bugs than the gtk2 version, etc. That's then my view and suggestion on how to approach the wxGTK issue in the way to get rid of the gtk2 USE flag. As for the only valid reason to keep it that I've heard, which was people wishing to prefer gtk1 when available for leaner resource usage, I'd personally suggest a gtk1 USE flag instead of gtk2, as that is rarely needed for normal people, and therefore no extra USE flag to set in the common case. I'd also suggest said people to disable anti-aliasing, use the default gtk2 theme, and use gtk2.6 for not that bigger resource usage. Delaying GNOME-2.14 for non-GNOME packages using gtk2 USE flag is mildly funny to me, too. Some two weeks have passed from 2.14 release, I would have expected it to be in x86 at least a week ago... but I'm living in a utopian land. -- With regards, Mart Raudsepp Project manager of wxMUD - http://wxmud.sourceforge.net/ Developer of wxWidgets- http://www.wxwidgets.org/ GTK+ port maintainer of OMGUI - http://www.omgui.org/ -- gentoo-dev@gentoo.org mailing list
Re: [gentoo-dev] [last rites] media-gfx/sodipodi
On Sun, 2 Apr 2006 18:42:50 -0300 Marcelo Góes [EMAIL PROTECTED] wrote: On 4/2/06, Carsten Lohrke [EMAIL PROTECTED] wrote: This is not the case. At least unless the user actively looks at package.mask. Since Portage doesn't provide the information, this point is void. And even if - four weeks are a too long, imho. I still do not understand what the rush is with removing a package. Readding a package if necessary will be much more troublesome than just keeping it masked for a month. I believe this is the general consensus on the subject. +1 on this one. Give people (and developers) time, not everybody is lightning's younger brother like you seem to be. Contrary to popular belief, there actually *are* people who sync less often than once per week. It's not like package.mask-ed package hurts anyone. I can't see what's the rush here either. Kind regards, -- Andrej Ticho Kacian ticho at gentoo dot org Gentoo Linux Developer - net-mail, antivirus, sound, x86 signature.asc Description: PGP signature
Re: [gentoo-dev] gtk2 use flag must die
On Sun, 2006-04-02 at 15:28 -0400, Mike Frysinger wrote: last time i recall following the gtk/gtk2 stuff, the idea was that in the future to move to a gtk/gtk1 situation ... but this was back when Spider was The Man, so i guess people forgot about that That was never the case. We actually saw the gtk2 flag only as a transitional tool during the initial release of gnome 2, too bad it stuck around as long as it did. it should be more the question, if there's anyone supporting Gtk1 upstream with regards to security issues etc.. and when such a situation arises, the solution may to simply drop the optional support. such a situation has not arose, so using such hypothetical examples is meaningless. Already security related issues have been dropped by upstream for the simple reason that it hasn't been maintained since the day gtk went 2.0 . The only reason there has been some minimal support are the bling distros like RH and the fact that Debian was stuck in the stone age. Let's be realistic, if an application hasn't been ported to gtk+-2 yet it is not maintained or it is an internal to some commercial business. I don't think gtk 1 will leave the tree soon, but at least we can try to make it unneeded on most users systems. - foser signature.asc Description: This is a digitally signed message part
Re: [gentoo-dev] gtk2 use flag must die
On Sun, 2006-04-02 at 15:16 -0400, Mike Frysinger wrote: and if there are no bugs filed ? this sort of stance is like the lets remove packages from portage because upstream is dead ... it benefits no one Sure it does, in my experience unmaintained packages tend to depend on unmaintained libs, which depend on other libs in older slotted versions. Usually parts of such a dependency chain have open bugs, that have been open for years and that are not going to be solved by anyone, because frankly nobody cares about that old crap, but isn't bothered enough to try and remove it and all of its reverse deps and take the flak for that, because just one guy in this world is still a frantic user of said package and will let the world know within 3 months after it has been removed. If you find something that hasn't been updated in 2-3 years, you are bound to find a trail of bugs and tree garbage leading away from it. Get rid of it, keep it clean. - foser signature.asc Description: This is a digitally signed message part
Re: [gentoo-dev] gtk2 use flag must die
On Monday 03 April 2006 00:29, foser wrote: I don't think gtk 1 will leave the tree soon, but at least we can try to make it unneeded on most users systems. I would just give my 2 eurocents about this, although I originally wasn't so keen on having gtk2 useflag dropped entirely. gtk 1.2 has also quite a bit of unicode/utf-8 problems... on alsaplayer we were plenty of crashes due to that. We're in 2006 and many users with native languages that requires special characters uses UTF-8, as it was discussed a while ago, with time, also users having English as native language will start using UTF-8 to be able to exchange information with the others... Mac OS X already uses UTF-8 by default as system codepage. This means that always more users will find problems and crashes with GTK 1.2... -- Diego Flameeyes Pettenò - http://dev.gentoo.org/~flameeyes/ Gentoo/Alt lead, Gentoo/FreeBSD, Video, AMD64, Sound, PAM, KDE pgputVkGTic1B.pgp Description: PGP signature
Re: [gentoo-dev] gtk2 use flag must die
On Mon, 2006-04-03 at 00:43 +0300, Mart Raudsepp wrote: Delaying GNOME-2.14 for non-GNOME packages using gtk2 USE flag is mildly funny to me, too. These two things are not related, 2.14 is not delayed whatsoever. Jakub's call was just to get attention to the bugs and didn't originate from the gnome team at all. Some two weeks have passed from 2.14 release, I would have expected it to be in x86 at least a week ago... but I'm living in a utopian land. I don't know where these expectations come from, but we intend to iron out the major known issues before we put stuff in ~arch . 2 weeks is rather short for a volunteer team of 2-3 active people for something the size of gnome. It is the same sort of nonsense we got with earlier releases, where people expect things to be in stable the day upstream declares it release day. People seem to expect the impossible, if you come from Debian the Gentoo cycle seems perfect, but as soon people are used to Gentoo the complaining starts anew. Get a grip and try to help out in constructive ways. - foser signature.asc Description: This is a digitally signed message part
Re: [gentoo-dev] gtk2 use flag must die
On Monday 03 April 2006 00:29, foser wrote: Already security related issues have been dropped by upstream for the simple reason that it hasn't been maintained since the day gtk went 2.0 . Why didn't you file (Gentoo) security bugs? Perfect reason to drop Gtk1 support, if no one steps up to fix them. Carsten pgpX2fQbQLdxW.pgp Description: PGP signature
Re: [gentoo-dev] gtk2 use flag must die
On Mon, 2006-04-03 at 00:53 +0200, foser wrote: On Mon, 2006-04-03 at 00:43 +0300, Mart Raudsepp wrote: Delaying GNOME-2.14 for non-GNOME packages using gtk2 USE flag is mildly funny to me, too. These two things are not related, 2.14 is not delayed whatsoever. Jakub's call was just to get attention to the bugs and didn't originate from the gnome team at all. Some two weeks have passed from 2.14 release, I would have expected it to be in x86 at least a week ago... but I'm living in a utopian land. I don't know where these expectations come from, but we intend to iron out the major known issues before we put stuff in ~arch . 2 weeks is rather short for a volunteer team of 2-3 active people for something the size of gnome. It is the same sort of nonsense we got with earlier releases, where people expect things to be in stable the day upstream declares it release day. People seem to expect the impossible, if you come from Debian the Gentoo cycle seems perfect, but as soon people are used to Gentoo the complaining starts anew. Get a grip and try to help out in constructive ways. That's what I did, and that's exactly the major part of what you cut out from the reply quotes. It being a blocker is exactly what I read out from the mails, without having found a bug number, which I perhaps lost in all the long thread. It being a _personal_ expectation was written with the notion that it would be as such in an ideal world, with the context of it being possibly blocked due to a USE flag in mind, and it was explicitly expressed as such. I do not see why I am getting such unconstructive replies to my majorly constructive e-mails. Should I cease writing e-mails to gentoo-dev, at the rare times I have something constructive to say? Now someone that deals with wxGTK or poEdit feel free to put use the constructive things I said in the thread if it's a good suggestion, and I'll use my time on working on wxGTK instead and upgrading my ~x86 system that has GNOME-2.14 ;) Great work it being in ~x86 already, btw! -- Mart Raudsepp -- gentoo-dev@gentoo.org mailing list
Re: [gentoo-dev] [last rites] media-gfx/sodipodi
On Sunday 02 April 2006 15:12, Jan Kundrát wrote: And considering that upstream is dead for about a year I think most people will not try to update that package every 2 days or something like that. Most people upgrade the whole system at once - those would see a warning about masked package. Yeah I know most people do. But then all that emerge world will tell most people is something like no new package since you emerged world one hour ago :) But well, you mean that easily overlooked message at the top of the emerge -vp world listing, which is followed by about 3 screens of wonderfully coloured messages? I'm sure I wouldn't notice it if I was to merge world. Ok the normal gentoo user does probably not get so many updateable packages but then again, if you don't pretend before merging world you don't even get that message. And if I pretend without -v I don't get the message either. I'm on portage 2.0.54 if that makes any difference. If it is more noticeable in other portage versions then forget everything I said :) This is the same kind of issue as the but there was a message printed in pkg_postinst!!!11, that you can observe in the who renamed adsl-start...-thread. Alex -- gentoo-dev@gentoo.org mailing list
Re: [gentoo-dev] [last rites] media-gfx/sodipodi
On Sunday 02 April 2006 17:23, Carsten Lohrke wrote: This is not the case. At least unless the user actively looks at package.mask. Since Portage doesn't provide the information, this point is void. And even if - four weeks are a too long, imho. As Andrej Kacian already noted, there are quite some people that don't sync every week. My general feeling is that most gentoo devs are far too fast in their decision/actions. As I already stated elsewhere in this thread, I generally only sync when I need to upgrade for feature/bug-/security-fixes and I don't see why that would be a bad idea. That way I get the benefits of gentoo but don't spend all day merging stuff that will have a new version two hours later :) Regarding your argument that you have to be actively looking at p.mask, that is not entirely true because a verbose world/system merge will tell you about a masked package, although I do think that this is not enough and even the message in verbose mode is not really noticeable. But just because portage does not really alert the user anyway, does not mean that masking first is bad, does it? I think the reporting of missing/masked packages in portage has to be improved, instead of removing the masking process :) Just my 0.02EUR. Alex -- gentoo-dev@gentoo.org mailing list
Re: [gentoo-dev] gtk2 use flag must die
On Mon, 2006-04-03 at 01:17 +0200, Carsten Lohrke wrote: On Monday 03 April 2006 00:29, foser wrote: Already security related issues have been dropped by upstream for the simple reason that it hasn't been maintained since the day gtk went 2.0 . Why didn't you file (Gentoo) security bugs? Perfect reason to drop Gtk1 support, if no one steps up to fix them. you are really trying hard to get gtk(1) Carsten signature.asc Description: This is a digitally signed message part
Re: [gentoo-dev] gtk2 use flag must die
On Monday 03 April 2006 01:54, Daniel Goller wrote: you are really trying hard to get gtk(1) Everyone as s/he likes. I favor the deprecation of the gtk2 flag and start dancing on my chair, once we have a Portage version with slot/use depends in arch. But this is a completely different topic: Knowingly providing our userbase with software that is vulnerable is a very bad. I'd argue the same for any software. Carsten pgpYdw5G6PRUo.pgp Description: PGP signature
Re: [gentoo-dev] [last rites] media-gfx/sodipodi
On Sun, 2006-04-02 at 21:20 +0200, Carsten Lohrke wrote: On Sunday 02 April 2006 04:48, Daniel Goller wrote: exactly, what's the point of removing it so fast? give people a chance to miss it, it does not matter if it's removed or masked only as far as going woah, what? and if masked it is a matter of unmasking rather than recommitting We haven't had a single issue with the usual seven day period as far as I can remember, so please come up with a valid argument against it, instead assuming turning my argument would be one. in short, if it's slowing down the process, why do you need it to be quick in the first place? Getting the junk out of tree and mind as fast as possible is a value in itself. you should apply a finer granularity and not call them all junk, even a unmaintained package that only has 50% of its features working might be the only thing someone has, where does this hurt anyone?, or maybe it is unmaintained but has no single (uncovered flaw), where does this hurt anyone? or or or, point is, say you would like certain vulnerable packages removed quicker, without making the waiting the usual 30 days sound insane. with that kind of grace period you give people the chance to say oh hey, i have this patch in my patch overlay, let me give it to you just wait a little, it hurts noone usually, if it's a security issue, say it is and use a shorter time, noone is gonna have a problem, unless carlo suddenly goes under the cloak of security and yanks everything he wants under those pretences... :) my $1 Daniel signature.asc Description: This is a digitally signed message part
[gentoo-dev] aging ebuilds with unstable keywords
Hi, This is an automatically created email message. http://gentoo.tamperd.net/stable has just been updated with 15047 ebuilds. The page shows results from a number of tests that are run against the ebuilds. The tests are: * if a version has been masked for 30 days or more. * if an arch was in KEYWORDS in an older ebuild, but not in the newer ones. * if SRC_URI contains hosts specified in thirdpartymirrors. * if ebuild uses patch instead of epatch. * if ebuild sets S to ${WORKDIR}/${P}. * if ebuild redefines P, PV, PN or PF. * if ebuild doesn't inherit eutils when it uses functions from eutils. * if ebuild doesn't inherit flag-o-matic when it uses functions from flag-o-matic. * if ebuild has $HOMEPAGE in SRC_URI (cosmetic). * if ebuild has $PN in SRC_URI (cosmetic). * if ebuild forces -fPIC flag to CFLAGS. * if ebuild has deprecated WANT_AUTO(CONF|MAKE)_?_?. * if ebuild uses is-flag -fPIC, should be changed to has_fpic. * if ebuild appends $RDEPEND or $DEPEND to $RDEPEND or $DEPEND to $DEPEND. * if ebuild has arch keyword(s) in iuse. * if ebuild overrides MAKEOPTS. * if ebuild has automake, autoconf or libtool in RDEPEND. * if ebuild exists in ChangeLog. * if ebuild installs COPYING and/or INSTALL into doc. The database is updated once a day and this email is sent once a week. Questions and comments may be directed to [EMAIL PROTECTED] Script has been running for 159 minutes. -- gentoo-dev@gentoo.org mailing list
[gentoo-dev] Special request for maintainer / auditor of LiSt Package
Let me preface this by saying that I am aware of the large amounts of bad blood existing between various Gentoo dev members and 'dma147'. I was not poking around under the hood of Gentoo when things transpired, so I am unaware of what happened... and frankly don't care to know what happened. dma147 has started a new Linux project --- linux-stats.org (LiSt). It uses a client to poll files in proc and other such things to generate a statistical profile of the user's machine, and then submit this info (only if the user agrees to) to the LiSt servers, where it then compiles things into collective statistical breakdowns. I would like to see this client added to the official tree, but that requires a dev to be maintainer and do a security audit of the client. Given that upstream is dma147, this obviously isn't something most of the devs I talked to on Freenode... errr jumped at so to speak. So I'm asking here if anybody is willing to do this. If necessary I will serve as middle man between bugzilla and upstream if that's the only way to see this through. I think it a shame and unfair to the end user for this not to be included just because of some spat between devs. The details of LiSt: All information is submitted anonymously as far as I can tell. By default nothing is submitted without the user's approval. The user can create a public profile page at linux-stats.org, but by default that is disabled. As far as I can tell, almost everything is attained simply by polling appropriate files within /proc (meaning this client is severely crippled for people using user mode kernels). The valuable end result is: - An *accurate* database of hardware, generated from lspci polling by client, where users can rate how it performs under Linux. This would be of alot of use to me when building machines for myself and clients to assure the best Linux experience possible. - My inner geek loves truck loads of stats for how things are being deployed as a whole... window managers, file systems, etc. If adopted wide scale, it would be arguably the most accurate source of statistics for kernel/fs/wm adoption etc. If anybody is willing to become maintainer and perform an audit, please reply. If anybody has questions, I have been temporarily voiced in #gentoo-dev under the name GTswagger ... I will try to help answer your questions as best I can if I'm around. - Jeremy Sands CS Undergrad, University of South Carolina Columbia/NCAAbbs LUG -- gentoo-dev@gentoo.org mailing list
[gentoo-dev] New virtuals: virtual/findutils virtual/admin-users virtual/admin-processes
Okay three more virtuals that I think should be implemented. The first would be for generic findutils, that is find and xargs commands; it would be satisfied by sys-apps/findutils or one of the BSD -ubin packages. This will solve problems of packages depending on sys-apps/findutils and then using a generic find instead of gfind on non-GNU userland. The second would be an alias for shadow or the -usbin packages, this way one can actually depend on the right utilities when using enewuser for instance. The third would be an alias for psmisc+procps or the -ubin packages, providing things like ps, kill and killall. If nobody has problems with them (considering I'll fix the dependencies when i get around keywording things, although everybody is welcome to fix their packages right after adding the virtuals if they want ;) ), I'll commit them Tuesday night or Wednesday afternoon (depends when I find time for them, actually, and how much I can go on without having the need of these virtuals). Thanks, -- Diego Flameeyes Pettenò - http://dev.gentoo.org/~flameeyes/ Gentoo/Alt lead, Gentoo/FreeBSD, Video, AMD64, Sound, PAM, KDE pgpWSOyhpHpi4.pgp Description: PGP signature