Recently, hplip-2.8.10 has been released (http://sourceforge.net/forum/forum.php?forum_id=882969), latest hplip in portage is 2.8.7. 2.8.10 should be included in portage.
Created attachment 171082 [details] hplip-2.8.10.ebuild
Created attachment 171083 [details, diff] hplip-2.8.7-2.8.10.ebuild.diff
However the attached ebuild doesn't work for me... i686-pc-linux-gnu-gcc "-DPACKAGE_NAME=\"HP Linux Imaging and Printing\"" -DPACKAGE_TARNAME=\"hplip\" -DPACKAGE_VERSION=\"2.8.10\" "-DPACKAGE_STRING=\"HP Linux Imaging and Printing 2.8.10\"" -DPACKAGE_BUGREPORT=\"2.8.10.33\" -DPACKAGE=\"hplip\" -DVERSION=\"2.8.10\" -DSTDC_HEADERS=1 -DHAVE_SYS_TYPES_H=1 -DHAVE_SYS_STAT_H=1 -DHAVE_STDLIB_H=1 -DHAVE_STRING_H=1 -DHAVE_MEMORY_H=1 -DHAVE_STRINGS_H=1 -DHAVE_INTTYPES_H=1 -DHAVE_STDINT_H=1 -DHAVE_UNISTD_H=1 -DHAVE_DLFCN_H=1 -DHAVE_LIBDL=1 -DHAVE_PTHREAD_H=1 -DHAVE_JPEGLIB_H=1 -DHAVE_UINT32_T=1 -DHAVE_LIBHPIP=1 -DHAVE_DBUS=1 -DHAVE_CUPS_CUPS_H=1 -DHAVE_USB_H=1 -DHAVE_PYTHON2_5_PYTHON_H=1 -I. -Iip -Iio/hpmud -Iscan/sane -I/usr/include/python2.5 -O2 -march=pentium4 -fomit-frame-pointer -pipe -c scan/scanext/scanext.c -fPIC -DPIC -o .libs/scanext_la-scanext.o make: *** No rule to make target `data/rules/70-hpmud.rules', needed by `all-am'. Stop. make: *** Waiting for unfinished jobs.... * * ERROR: net-print/hplip-2.8.10 failed. * Call stack: * ebuild.sh, line 49: Called src_compile * environment, line 3355: Called die * The specific snippet of code: * emake || die "Compilation failed" * The die message: * Compilation failed * * If you need support, post the topmost build error, and the call stack if relevant. * A complete build log is located at '/var/log/portage/net-print:hplip-2.8.10:20081111-113905.log'. * The ebuild environment file is located at '/var/tmp/portage/net-print/hplip-2.8.10/temp/environment'. * This ebuild is from an overlay: '/usr/local/portage/' * >>> Failed to emerge net-print/hplip-2.8.10, Log file:
Created attachment 171449 [details] /var/log/portage/net-print:hplip-2.8.10:20081111-113905.log
Created attachment 171450 [details] /var/tmp/portage/net-print/hplip-2.8.10/temp/environment please let's solve this troubles, before we accept it in portage...
(In reply to comment #5) > Created an attachment (id=171450) [edit] > /var/tmp/portage/net-print/hplip-2.8.10/temp/environment > > please let's solve this troubles, before we accept it in portage... > Your miss one file from hplip/files Copy from /usr/portage/net-print/hplip/files/70-hpmud.rules to /usr/local/portage/net-print/hplip/files/70-hpmud.rules It fix your problem.
Created attachment 171517 [details] 70-hpmud.rules
Thanx Yar for your hint, this solved my troubles.
This worked "out of the box" for me. Have used it in a fresh install on AMD64, with a brand new networked HP Office Jet Pro L7590. Needed to enable SNMP in the settings on the printer itself, plus use the "snmp" USE flag with this ebuild, and the printer was then detected properly and functions fine. (Well, for printing anyway.) Have not yet tested the faxing/scanning component yet, so that's still an unknown. An additional thought is that I needed to manually create a soft link from "foomatic-rip" to the cups directory before printing would actually complete: $ sudo ln -s /usr/bin/foomatic-rip /usr/libexec/cups/filter But that is likely a localised configuration requirement, rather than something this ebuild should be delayed for.
(In reply to comment #9) > This worked "out of the box" for me. > > Have used it in a fresh install on AMD64, with a brand new networked HP Office > Jet Pro L7590. > > Needed to enable SNMP in the settings on the printer itself, plus use the > "snmp" USE flag with this ebuild, and the printer was then detected properly > and functions fine. (Well, for printing anyway.) > > Have not yet tested the faxing/scanning component yet, so that's still an > unknown. > > An additional thought is that I needed to manually create a soft link from > "foomatic-rip" to the cups directory before printing would actually complete: > > $ sudo ln -s /usr/bin/foomatic-rip /usr/libexec/cups/filter > > But that is likely a localised configuration requirement, rather than something > this ebuild should be delayed for. > 1. SNMP - it's not a ebuild problem, probably it's certain device install particularity. 2. foomatic link - it's not a ebuild problem too - you emerge net-print/foomatic-filters without cups use flag.
Yep, you're right on both counts. Tested and verified on a new laptop too. :) This ebuild is fine. :)
Please, don't submit a full ebuild but a diff against the current one instead, and explain why the changes. This helps us a lot. Anyway, my trusty officejet is still on a ship somewhere in the middle of the Atlantic, so this will have to wait. Denis.
*** Bug 249915 has been marked as a duplicate of this bug. ***
Note, some models need additionally the proprietary plugins from http://www.linuxprinting.org/download/printdriver/auxfiles/HP/plugins/. E.g. for the HP Color LaserJet CM1015 you have to installation also hplip-2.8.10-plugin.run to support scanning over the net. Otherwise scanning fails and you will find an entry in the logs of the server like: =========== %< ============== Dec 16 00:38:33 bobbel saned[32342]: scan/sane/soap.c 89: unable to load restricted library /usr/share/hplip/scan/plugins/bb_soap.so: /usr/share/hplip/scan/plugins/bb_soap.so: cannot open shared object file: No such file or directory =========== %< ============== Bascially those plugins are alway needed if it is mensioned as required for the specific model like here: http://hplipopensource.com/hplip-web/models/color_laserjet/hp_color_laserjet_cm1015.html
Version 2.8.12 available: http://bugs.gentoo.org/show_bug.cgi?id=251733
*** Bug 251733 has been marked as a duplicate of this bug. ***
hplip-2.8.12 was released...
*** Bug 238500 has been marked as a duplicate of this bug. ***
Created attachment 177677 [details] hplip-2.8.12.ebuild Version bump to 2.8.12 Added new use flag "plugins" for models need in propietary firmware loading (test it please, i haven't such devices)
Created attachment 177678 [details] metadata.xml for 2.8.12 for 2.8.12
(In reply to comment #19) > Added new use flag "plugins" for models need in propietary firmware loading > (test it please, i haven't such devices) Have you reviewed the legal implications of this ? I'm not sure we have the right to do it. I'm not saying we don't but this needs to be verified. I still won't commit this before some time (same old reason, my OfficeJet is in a container somewhere waiting for me to have a house), but it's nice to have that discussion now so that it's solved in time. Please attach a diff against the current ebuild and comment on each change. A full ebuild is annoying because we can't easily see what changed. A user submitted ebuild doesn't go in before it has been thouroughly checked, so the easier you make it to check the faster it will be in the tree. By the way, I'm in the printing team, so no need to CC me. Denis.
Created attachment 178169 [details, diff] hplip-2.8.7-2.8.12.ebuild.diff
Comments on changes: Lines 1-19 - Add flag-o-matic for filter out some flags - Add SRC_URI for plugins (firmware). - Add flag plugins to IUSE Lines 32-37 - Add app-arch/gzip to rdepend needed to extract package with plugins and extract ppds. Lines 54-59 - Add RESTRICT="plugins? ( strip )" to avoid warning on already compiled plugins and firmware. Lines 74-82 - Unpack archive with plugins and fix wrong place for hal rules file. Lines 96-102 - Fix handling of ppds to use system foomatic-rip wich now if gziped. Lines 139-151 - Avoid multiprocess compilation - Replace deprecated SYSFS with ATTRS at plugins udev rules - Install plugins - Install docs
Created attachment 178170 [details, diff] metadata.xml-2.8.12.diff Add <flag name='plugins'>
(In reply to comment #21) > (In reply to comment #19) > > Added new use flag "plugins" for models need in propietary firmware loading > > (test it please, i haven't such devices) > > Have you reviewed the legal implications of this ? I'm not sure we have the I don't reviewd the legal implications of this because i don't know how. I attached license shiped with plugin, but i'm not a lawer and can't understand how this license deals with gnu, sorry. But look at this at debian: http://packages.debian.org/sid/i386/hplip/filelist > Please attach a diff against the current ebuild and comment on each change. A > full ebuild is annoying because we can't easily see what changed. A user > submitted ebuild doesn't go in before it has been thouroughly checked, so the > easier you make it to check the faster it will be in the tree. done, thanx. > By the way, I'm in the printing team, so no need to CC me. done
Created attachment 178173 [details] license shipped with plugins
From license: =============================================================================== 3. Copies and Adaptations. You may only make copies or adaptations of the Software for archival purposes or when copying or adaptation is an essential step in the authorized Use of the Software. You must reproduce all copyright notices in the original Software on all copies or adaptations. You may not copy the Software onto any public network. =============================================================================== As I read the last sentence we should be unable to download this plugins at all but probably linuxprinting.org has some agreements with HP. In any case we are not allowed to redistribute this files thus probably the best solution will be to separate plugins into separate fetch-restricted package. Users will have manually download distfiles...
(In reply to comment #27) > In any case we are not allowed to redistribute this files thus probably the > best solution will be to separate plugins into separate fetch-restricted > package. Users will have manually download distfiles... I could never try this as I don't have the necessary equipment, but I was told the hp-setup program (part of hplip) allowed you to download plugins when necessary. If that's true (which I always assumed) we don't need to add a package at all. To Yar Odin: thanks for posting the diff and comments. I'll review that. Denis.
No, I had to download the plugin package separately. Isn't it possible to have this file fetch restricted only so that the plugin USE flag can be kept?
(In reply to comment #29) > No, I had to download the plugin package separately. Isn't it possible to have > this file fetch restricted only so that the plugin USE flag can be kept? > Here a little variation of the current ebuild to make the gui work with qt4.
Created attachment 178613 [details, diff] hplip.ebuild
I am not sure we can still disable foomatic-rip as there is no ebuild providing the functionality at the moment. Please check that and change the ebuild accordingly. I had the problem when installing a new HP printer that need the foomatic-rip-hplip file and it wasnt installed.
(In reply to comment #32) > I am not sure we can still disable foomatic-rip as there is no ebuild providing > the functionality at the moment. Please check that and change the ebuild > accordingly. Last time I checked our latest version of foomatic-rip was enough, but that was a few bumps of hplip ago, so it may have changed. In any case, using an internal (to the package) copy of foomatic-rip is a definite no-no. That's valid for any package, tool or library, under any circumstances. http://blog.flameeyes.eu/2008/01/21/what-to-do-with-shared-code The right way to go is to make sure the in-tree version of foomatic-rip provides enough features for hplip, i.e. that it is recent enough. So you know what to do now: make sure a corresponding stable/unstable version of foomatic-rip has what you need (use another bug please). As a side note, I'm still away and chances that my officejet gets unpacked soon are low. Denis.
hplip 3.9.2 was released this week. would be nice to have a bump to it; at least for my psc_1215 the hplip currently in portage fails on printing (but not on scanning).
(In reply to comment #32) > I am not sure we can still disable foomatic-rip as there is no ebuild providing > the functionality at the moment. Please check that and change the ebuild > accordingly. I had the problem when installing a new HP printer that need the > foomatic-rip-hplip file and it wasnt installed. > foomatic from portage is about one year old. i think it need to update.
Created attachment 182736 [details, diff] hplip-2.8.7-3.9.2.ebuild.diff
(In reply to comment #36) > hplip-2.8.7-3.9.2.ebuild.diff > Here the ebuild compiles fine. I still have two errors: localhost liviu # hp-check [..] Checking for dependency: CUPS DDK - CUPS driver development kit... warning: NOT FOUND! This is an OPTIONAL/RUNTIME ONLY dependency. Some HPLIP functionality may not function properly. [..] Checking for dependency: ppdev - Parallel port support kernel module.... error: NOT FOUND! This is a REQUIRED/RUNTIME ONLY dependency. Please make sure that this dependency is installed before installing or running HPLIP. Would this be an HPLIP, or a Gentoo issue? What concerns CUPS DDK, I have it installed. localhost liviu # eix cupsddk [I] net-print/cupsddk Installed versions: 1.2.3(13:54:17 21/02/09)(fltk) Considering ppdev, is it really necessary if I have the -parport USE flag? Perhaps helpful information: localhost liviu # uname -r 2.6.22-gentoo-r5-0.2 localhost liviu # eix cups -I [I] net-print/cups Installed versions: 1.3.9-r1!t(13:51:39 21/02/09)(X acl dbus java jpeg ldap pam perl png ppds python samba ssl tiff -avahi -gnutls -kerberos -linguas_de -linguas_en -linguas_es -linguas_et -linguas_fr -linguas_he -linguas_id -linguas_it -linguas_ja -linguas_pl -linguas_sv -linguas_zh_TW -php -slp -static -xinetd -zeroconf) localhost liviu # eix hplip [I] net-print/hplip Installed versions: 3.9.2!s[?](16:20:21 21/02/09)(cupsddk dbus gtk plugins ppds qt3 qt4 scanner snmp -doc -fax -minimal -parport)
(In reply to comment #37) > (In reply to comment #36) > > hplip-2.8.7-3.9.2.ebuild.diff > > > Here the ebuild compiles fine. I still have two errors: > > localhost liviu # hp-check > [..] > Checking for dependency: CUPS DDK - CUPS driver development kit... > warning: NOT FOUND! This is an OPTIONAL/RUNTIME ONLY dependency. Some HPLIP > functionality may not function properly. fixed > Checking for dependency: ppdev - Parallel port support kernel module.... > error: NOT FOUND! This is a REQUIRED/RUNTIME ONLY dependency. Please make sure > that this dependency is installed before installing or running HPLIP. > Considering ppdev, is it really necessary if I have the -parport USE flag? Do you have ppdev loaded? Show your lsmod| grep ppdev .
Created attachment 182752 [details] bad - fix autostart for hplip systray - fix cupsddk detection - added changes from Jimmy.Jazz@gmx.net - some ebuild code/text cleaning
Created attachment 182754 [details, diff] hplip-2.8.7-3.9.2.ebuild.diff - second edition - fix autostart for hplip systray - fix cupsddk detection - added changes from Jimmy.Jazz@gmx.net - some ebuild code/text cleaning
(In reply to comment #38) > Do you have ppdev loaded? Show your lsmod| grep ppdev . > I have just reconfigured my kernel with ppdev, and loaded the module, and this hp-check error is gone. (In reply to comment #40) > Created an attachment (id=182754) [edit] > /hplip-2.8.7-3.9.2.ebuild.diff - second edition > I also tried the 2nd edition ebuild, and the `CUPS DDK' error is also gone. Now hp-check passes with `No errors or warnings.' (Thanks, Yar!) I would also report that with USE flags +qt4 +gtk -qt3 all hp dialogues are QT4 and they seem to function properly. But I'm still unable to print anything with my PSC, and I'm greated with various messages: - "Device communication error" - "error: Cannot print: Printer is stopped. Please START the printer to continue this print. Job will begin printing once printer is started." - "error: Print command failed with status code 1. lpr -P psc_1200_8 "/home/liviu/main.log" " - "error: Unable to communicate with device (code=31): hp:/usb/psc_1200_series?serial=MY51TG117GT0" - "error: Unable to open device hp:/usb/psc_1200_series?serial=MY51TG117GT0." - "error: Unable to print to printer. Please check device and try again." - "error: Unable to communicate with printer psc_1200. Please check the printer and try again." Otherwise, scanning works OK, and so does copying. With some <2.8.6b version everything worked fine.
(In reply to comment #41) > But I'm still unable to print anything with my PSC, and I'm greated with > various messages: > - "Device communication error" > - "error: Cannot print: Printer is stopped. Please START the printer to > continue this print. Job will begin printing once printer is started." > - "error: Print command failed with status code 1. lpr -P psc_1200_8 > "/home/liviu/main.log" " > - "error: Unable to communicate with device (code=31): > hp:/usb/psc_1200_series?serial=MY51TG117GT0" > - "error: Unable to open device hp:/usb/psc_1200_series?serial=MY51TG117GT0." > - "error: Unable to print to printer. Please check device and try again." > - "error: Unable to communicate with printer psc_1200. Please check the > printer and try again." > > Otherwise, scanning works OK, and so does copying. With some <2.8.6b version > everything worked fine. > Is you try hp-setup?
(In reply to comment #41) > (In reply to comment #38) > > Do you have ppdev loaded? Show your lsmod| grep ppdev . > > > I have just reconfigured my kernel with ppdev, and loaded the module, and this > hp-check error is gone. > > (In reply to comment #40) > > Created an attachment (id=182754) [edit] > > /hplip-2.8.7-3.9.2.ebuild.diff - second edition > > > I also tried the 2nd edition ebuild, and the `CUPS DDK' error is also gone. Now > hp-check passes with `No errors or warnings.' (Thanks, Yar!) I would also > report that with USE flags +qt4 +gtk -qt3 all hp dialogues are QT4 and they > seem to function properly. > > But I'm still unable to print anything with my PSC, and I'm greated with > various messages: > - "Device communication error" > - "error: Cannot print: Printer is stopped. Please START the printer to > continue this print. Job will begin printing once printer is started." > - "error: Print command failed with status code 1. lpr -P psc_1200_8 > "/home/liviu/main.log" " > - "error: Unable to communicate with device (code=31): > hp:/usb/psc_1200_series?serial=MY51TG117GT0" > - "error: Unable to open device hp:/usb/psc_1200_series?serial=MY51TG117GT0." > - "error: Unable to print to printer. Please check device and try again." > - "error: Unable to communicate with printer psc_1200. Please check the > printer and try again." > > Otherwise, scanning works OK, and so does copying. With some <2.8.6b version > everything worked fine. > or even try to reinstall you devices at CUPS.
(In reply to comment #42) > Is you try hp-setup? > (In reply to comment #43) > or even try to reinstall you devices at CUPS. > Yeah, I did. These couple of days I probably reinstalled my printer a dozen times, using hp-setup and localhost:631. Now I have removed the printer configuration altogether, reinitialised and reinstalled the printer via hp-setup. Here's what I see in `localhost:631': psc_1200_1 (Default Printer) Description: home Location: first Printer Driver: HP PSC 1200 Series hpijs, hpijs 3.9.2.49 Printer State: processing, accepting jobs, published. Device URI: hp:/usb/psc_1200_series?serial=MY51TG117GT0 I also configured a psc_1200_2 it via `localhost:631', but same story: should I send a testpage, or any page (say, from a gtk+ editor), CUPS receives the job but seems unable to correctly process and send it to the printer. (As usual, after all the reinstalls scanning works fine.) With the latest ebuild once I also got the following error in CUPS printers: "Unable to start filter "foomatic-rip-hplip" - No such file or directory." However, could this be a ghostscript problem? Several times I found that after sending a page to print my processor went full-throttle because of an unresponsive `gs' process (bizarrely, this happens after I modify---change the driver of---my printer in CUPS; and funny, my printer's configuration does not change so that to try a different driver I have to reconfigure from scratch). Currently I have: localhost liviu # eix ghostscript [U] app-text/ghostscript-gpl Installed versions: 8.62(19:19:40 13/09/08)(X cups gtk -bindist -cjk -djvu -jpeg2k) I will experiment with 8.63-r1 and 8.64, would it make sense to switch to app-text/ghostscript-esp or app-text/ghostscript-gnu? These are advertised as patched versions of gs-gpl.
(In reply to comment #44) > (In reply to comment #42) > > Is you try hp-setup? > > > > (In reply to comment #43) > > or even try to reinstall you devices at CUPS. > > > Yeah, I did. These couple of days I probably reinstalled my printer a dozen > times, using hp-setup and localhost:631. Now I have removed the printer > configuration altogether, reinitialised and reinstalled the printer via > hp-setup. Here's what I see in `localhost:631': > psc_1200_1 (Default Printer) > Description: home > Location: first > Printer Driver: HP PSC 1200 Series hpijs, hpijs 3.9.2.49 > Printer State: processing, accepting jobs, published. > Device URI: hp:/usb/psc_1200_series?serial=MY51TG117GT0 > > I also configured a psc_1200_2 it via `localhost:631', but same story: should I > send a testpage, or any page (say, from a gtk+ editor), CUPS receives the job > but seems unable to correctly process and send it to the printer. (As usual, > after all the reinstalls scanning works fine.) > > With the latest ebuild once I also got the following error in CUPS printers: > "Unable to start filter "foomatic-rip-hplip" - No such file or directory." This message is very very strange. Please unmerge hplip totally and look at /usr/share/ppd/HP/ dir. Is you see any files with psc_1200 at name? If you find such files - delete it and reinstall new hplip (3.9.2 third). If it doesn't help - try to put manually this ppd http://openprinting.org/ppd-o-matic.cgi?driver=hpijs&printer=HP-PSC_1200&show=0 to /usr/share/ppd/HP/ and reinstall your device. > However, could this be a ghostscript problem? Several times I found that after I think, no.
Created attachment 182804 [details, diff] hplip-2.8.7-3.9.2.ebuild - third edition
3.9.2 works for me, including HP Device Manager (~x86, kde4) net-print/hplip-3.9.2 USE="cupsddk dbus doc fax gtk plugins ppds qt4 scanner snmp -minimal -parport -qt3" Tested with printer HP Officejet Pro l7500, I/O bus: usb Thank you, good work
(In reply to comment #45) > > "Unable to start filter "foomatic-rip-hplip" - No such file or directory." > This message is very very strange. > I got it only once with the 2nd edition ebuild. Dunno how to recreate the situation, and didn't get it with the 3rd edition. > Please unmerge hplip totally and look at /usr/share/ppd/HP/ dir. > Is you see any files with psc_1200 at name? > If you find such files - delete it and reinstall new hplip (3.9.2 third). > If it doesn't help - try to put manually this ppd > http://openprinting.org/ppd-o-matic.cgi?driver=hpijs&printer=HP-PSC_1200&show=0 > to /usr/share/ppd/HP/ and reinstall your device. > Still no luck. I unmerged hplip, and there was no sign of the file: liviu@localhost ~ $ ls /usr/share/pp* ls: cannot access /usr/share/pp*: No such file or directory Remerged 3rd edition ebuild, installed, reinstalled, specified the manual .ppd file, all to no avail. I seem to get less diverse error messages, but printing still don't work. With the qt4 dialogues, one thing that does not work here even with the 3rd ebuild is the hp-toolbox Settings dialogue. Here I suspect the problem lies in hplip: Traceback (most recent call last): File "//usr/share/hplip/ui4/devmgr5.py", line 1944, in PreferencesAction_activated dlg = SettingsDialog(self) File "//usr/share/hplip/ui4/settingsdialog.py", line 49, in __init__ self.updateControls() File "//usr/share/hplip/ui4/settingsdialog.py", line 58, in updateControls RefreshAllRadioButton.setChecked(True) NameError: global name 'RefreshAllRadioButton' is not defined > > However, could this be a ghostscript problem? Several times I found that after > I think, no. > I switched to gs-gpl-8.63-r1, but gs is still failing upon printer modification.
3rd patch edition compiled and installed without problems. Working with configuration (printer is behind router which forwards 9100 port to the printer): Description: HP LaserJet 2300 Duplex Printer Driver: HP Color LJ cm2320fxi MFP hpijs, hpijs 3.9.2.49 Device URI: socket://xxx.xxx.xxx.xxx:9100
Created attachment 184692 [details] hplip 3.9.2 full ebuild (from Yar Odin patch) This is the ebuild you get from Yar Odin's patch applied to the old ebuild. I'm simply posting this up because I think its easier just to download the ebuild (instead of patch it as well)
*** Bug 262206 has been marked as a duplicate of this bug. ***
(In reply to comment #46) > Created an attachment (id=182804) [edit] > hplip-2.8.7-3.9.2.ebuild - third edition > I have a few dubts about the new ebuild: - You set app-arch/gzip as a RDEPEND, are you sure it's a RDEP instead of a DEP? Also, are you sure it's needed even being in system profile? - Seems that you are restricting strip when plugins USE flag is being set. This is a suboptimal solution as it will cause all hplip to be unstripped when users install plugins. Maybe the warnings are an upstream bug :-/ - Maybe "unpack_myself" function (from eutils.eclass) could be used for unpacking plugins .run file - Maybe I am wrong but, seems that you are sedding Makefile for installing hal file under /etc instead of /usr... it's not the proper way. Hal files provided by packages that are installed using emerge must be installed under /usr. If a user wants to modify it or use a custom hal file, he/she should put it under /etc - Why do you enable dependency-tracking? - Proper autostart dir would be /etc/xdg/autostart/ instead of /usr/share/autostart - Why do you use replace-flags for running emake with -j1 instead of simply run "emake -j1" ? - About replacing deprecated SYSFS by ATTRS... would be nice send it to upstream also :-) - About "arch_var" stuff, maybe ppc64 should also use arch_var="64" ;-) Seems that plugins are a bit hard to handle, Have you think about creating a separate ebuild for simply installing plugins? This could also workaround stripping problem in a better way (from my point of view) Anyway, I am a simple gentoo user and I am not a member of gentoo printing team, then, maybe you should wait for their opinion on this :-) Thanks and best regards
(In reply to comment #52) > (In reply to comment #46) > > Created an attachment (id=182804) [edit] > > hplip-2.8.7-3.9.2.ebuild - third edition > > > > I have a few dubts about the new ebuild: > - You set app-arch/gzip as a RDEPEND, are you sure it's a RDEP instead of a > DEP? Also, are you sure it's needed even being in system profile? My mistake. Will be fixed. > - Seems that you are restricting strip when plugins USE flag is being set. This > is a suboptimal solution as it will cause all hplip to be unstripped when users > install plugins. Maybe the warnings are an upstream bug :-/ > - Maybe "unpack_myself" function (from eutils.eclass) could be used for > unpacking plugins .run file > Seems that plugins are a bit hard to handle, Have you think about creating a > separate ebuild for simply installing plugins? This could also workaround > stripping problem in a better way (from my point of view) > This propietary plugins are the hard theme. Now i think that best solution is throught out it from ebuild and write bug report to upstream about SYSFS/ATTRS in udev rules. > - Maybe I am wrong but, seems that you are sedding Makefile for installing hal > file under /etc instead of /usr... it's not the proper way. Hal files provided > by packages that are installed using emerge must be installed under /usr. If a > user wants to modify it or use a custom hal file, he/she should put it under > /etc Hmm. For example sys-fs/ntfs3g install hal rules to /etc/hal/. And > - Why do you enable dependency-tracking? > - Proper autostart dir would be /etc/xdg/autostart/ instead of > /usr/share/autostart It's true only for WMs which supports xdg specs. For example for KDE3 it /usr/kde/3.5/share/autostart. > - Why do you use replace-flags for running emake with -j1 instead of simply run > "emake -j1" ? Agree. > - About replacing deprecated SYSFS by ATTRS... would be nice send it to > upstream also :-) You can write it if you want ;) > - About "arch_var" stuff, maybe ppc64 should also use arch_var="64" ;-) Agree. > Anyway, I am a simple gentoo user and I am not a member of gentoo printing > team, then, maybe you should wait for their opinion on this :-) Thanks for useful advices. I want to rewrite ebuild close to end of next week (may be at EAPI=2), when i finish other projects.
Created attachment 185645 [details, diff] hplip-3.9.2.diff to version 2.8.7 Attached a more conservative approach on the update to hplip. Also I left those plugins out as I agree this should be done in a separate ebuild. So finally there may be just a plugins USE flag within the hplip ebuild that pulls in the plugins ebuild if set. Some changes are re-used from the previous ebuilds posted here where I have chosen the ones making most sense to me plus adding some additional stuff. re-used stuff: - seding gzipped ppd files - seding installer/core_install.py - fix for qt4 gui build - update-desktop-database additional stuff: - Converted ebuild to EAPI 2 - Removed dev-python/ctypes dependency as the homepage states it is included in python-2.5 which is already stable on gentoo - placed setting of QT_VER="4" after QT_VER="3" so qt version 4 is actually used for the following seds if the qt4 flag is activated - econf dies on its own
(In reply to comment #53) > Hmm. For example sys-fs/ntfs3g install hal rules to /etc/hal/. And In specs/hal-spec-fdi-files.xml (in hal-0.5.11 sources tarball) says that /usr is for files provided by packages and /etc is for files provided by the system administrator / user , maybe ntfs3g is not doing the right thing also :-/ > > > - Why do you enable dependency-tracking? > > - Proper autostart dir would be /etc/xdg/autostart/ instead of > > /usr/share/autostart > It's true only for WMs which supports xdg specs. > For example for KDE3 it /usr/kde/3.5/share/autostart. > This should change with kde4 :-/, doesn't kde3 read /etc/xdg/autostart also? > > - About replacing deprecated SYSFS by ATTRS... would be nice send it to > > upstream also :-) > You can write it if you want ;) > Well, I have no idea about udev stuff :-/, this is why I suggested you to do that but, if you aren't able to do so, I can report it also > Thanks for useful advices. > I want to rewrite ebuild close to end of next week (may be at EAPI=2), when i > finish other projects. > Thanks to you for your work :-)
(In reply to comment #55) > (In reply to comment #53) > > Hmm. For example sys-fs/ntfs3g install hal rules to /etc/hal/. And > > In specs/hal-spec-fdi-files.xml (in hal-0.5.11 sources tarball) says that /usr > is for files provided by packages and /etc is for files provided by the system > administrator / user , maybe ntfs3g is not doing the right thing also :-/ I can't imagine situation when user install hplip or ntfs3g and doesn't want it automatic support at hal rules. But i still haven't answer to this question from developers. > > > > > - Why do you enable dependency-tracking? It's my mistake too. I forget to disable this after debug. > > > - Proper autostart dir would be /etc/xdg/autostart/ instead of > > > /usr/share/autostart > > It's true only for WMs which supports xdg specs. > > For example for KDE3 it /usr/kde/3.5/share/autostart. > > > > This should change with kde4 :-/, doesn't kde3 read /etc/xdg/autostart also? Nope, kde3 doesn't read /etc/xdg/autostart and all other xdg places(redifined at start-kde) > > > - About replacing deprecated SYSFS by ATTRS... would be nice send it to > > > upstream also :-) > > You can write it if you want ;) > > > Well, I have no idea about udev stuff :-/, this is why I suggested you to do > that but, if you aren't able to do so, I can report it also Done. https://bugs.launchpad.net/hplip/+bug/346390
Created attachment 185778 [details, diff] hplip-3.9.2.diff to version 2.8.7 A few additional modifications to the ebuilds with the help from yarodin on IRC. - removed versioned dependencies for packages having no lower stable version as the required one - only sed base/device.py and hplip-systray.desktop.in if one ore both of qt3/4 flags are set - use lower case for local variables in every case - fix permissions on /usr/libexec/cups/backend/hp - also remove files installed under /etc/xdg if qt3/4 are unset both - add autostart desktop file for kde3 to correct location
For backward compatibility with older distros using udev we use SYSFS instead of ATTRS. -dave
http://sourceforge.net/forum/forum.php?forum_id=928083 >>The hplip-3.9.2-high_cpu_utilization_logout.patch file fixes the following issue.: >>1. Qt4 version 3.9.2 doesn't close from a switch user and uses high CPU. Please, add the patch in hplip-3.9.2.ebuild
Created attachment 186530 [details, diff] hplip-3.9.2.diff (In reply to comment #59) > http://sourceforge.net/forum/forum.php?forum_id=928083 > > >>The hplip-3.9.2-high_cpu_utilization_logout.patch file fixes the following issue.: > >>1. Qt4 version 3.9.2 doesn't close from a switch user and uses high CPU. > > > Please, add the patch in hplip-3.9.2.ebuild > Et voilà, added patch to the diff. The patch itself follows.
Created attachment 186531 [details, diff] hplip-3.9.2-high_cpu_utilization_logout.patch
I just tried this and it worked wonderfully. I'm running ~arch amd64 in a gnome environment. I installed with the qt4 USE flag and then removed it and used the cli and both worked great. Thanks to the creators!
Added hplip-3.9.2 to the main tree
