Menu
▾
▴
morphix-devel — Development issues, wishes et al.
You can subscribe to this list here.
| 2003 |
Jan
|
Feb
|
Mar
|
Apr
|
May
(5) |
Jun
(13) |
Jul
(6) |
Aug
|
Sep
(16) |
Oct
(14) |
Nov
(10) |
Dec
(23) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2004 |
Jan
(38) |
Feb
(49) |
Mar
(21) |
Apr
(41) |
May
(15) |
Jun
(38) |
Jul
(43) |
Aug
(13) |
Sep
(37) |
Oct
(61) |
Nov
(17) |
Dec
(2) |
| 2005 |
Jan
(68) |
Feb
(11) |
Mar
(28) |
Apr
(25) |
May
(11) |
Jun
(11) |
Jul
(21) |
Aug
(6) |
Sep
(117) |
Oct
(38) |
Nov
(4) |
Dec
(12) |
| 2006 |
Jan
(21) |
Feb
(34) |
Mar
(72) |
Apr
(21) |
May
(78) |
Jun
(61) |
Jul
(45) |
Aug
(46) |
Sep
(72) |
Oct
(18) |
Nov
(29) |
Dec
(31) |
| 2007 |
Jan
(9) |
Feb
(8) |
Mar
(9) |
Apr
(30) |
May
(8) |
Jun
(16) |
Jul
(3) |
Aug
(8) |
Sep
(10) |
Oct
(17) |
Nov
(4) |
Dec
(22) |
| 2008 |
Jan
(16) |
Feb
(9) |
Mar
|
Apr
(2) |
May
|
Jun
|
Jul
(2) |
Aug
(2) |
Sep
(1) |
Oct
|
Nov
|
Dec
|
| 2009 |
Jan
(7) |
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
| 2010 |
Jan
|
Feb
(2) |
Mar
(8) |
Apr
|
May
(3) |
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
|
From: Alex de L. <al...@de...> - 2007-04-11 21:32:04
|
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi Basanta, Basanta shrestha wrote: > Hi all, > I recently tried > /morphix.svn.sourceforge.net/svnroot/morphix/trunk/morphixinstaller2. I > couln't install into SATA hdd. Does this support SATA at all? Nope. A problem is that I don't yet have a machine with SATA, so I can't do this myself. Also, morphixinstaller2 is still in development, not yet ready for end-users. > mdxinstaller > from medianix.net was able to install into sata, I could have used this but > the final version is not available yet and I wonder if mdxinstaller will > have a gui version? I'm not aware of mdxinstaller or medianix. nextime from #morphix/irc.freenode.net is working on morphixinstaller2.5 (or similar, he's doing some major rewrites). Once he's finished the whole SATA-issue should be fixed. Feel free to pester him online about its state :) cheers, Alex -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.3 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFGHVPDQeuQA5TF/UsRAh4+AJ0T+sGIl5vTPb1d79O2H/nuL0TCcQCgg9Kd MJZLapUyqa0DxfEDxM+TgQk= =gQS2 -----END PGP SIGNATURE----- |
|
From: Basanta s. <bas...@gm...> - 2007-04-11 05:27:50
|
Dear all, I am urgently in need of a basemod with kernel 2.6.17 with bootsplash enabled. I request the concerned developer to make one and upload if possible. Thanking you. Regards, Basanta shrestha |
|
From: Basanta s. <bas...@gm...> - 2007-04-11 03:29:26
|
Hi all, I recently tried /morphix.svn.sourceforge.net/svnroot/morphix/trunk/morphixinstaller2. I couln't install into SATA hdd. Does this support SATA at all? mdxinstaller from medianix.net was able to install into sata, I could have used this but the final version is not available yet and I wonder if mdxinstaller will have a gui version? Basanta shrestha |
|
From: Paul <pa...@tu...> - 2007-04-09 20:08:42
|
On Monday 09 April 2007 20:04, Brendan M. Sleight wrote: > Cool - etch MainModules. > > Does this mean the AutoBuild bases should be etch builds ? When I can pull my finger out, I don't see why not - Once I've got Sarge archived, Etch will take it's place.. Regards, Paul. |
|
From: Brendan M. S. <me...@ba...> - 2007-04-09 19:04:40
|
On 4/8/07, Paul wrote: > > Now that Debian has officially released Etch as "stable", Zathras will be > getting an upgrade - This in effect means the Sarge autobuilds will be > dropped unless there is an overwhelming demand for them. Cool - etch MainModules. Does this mean the AutoBuild bases should be etch builds ? Lenny here we come. Sorry just finished watching Toy Story 2 on the telly. Regards, Brendan |
|
From: Alex de L. <al...@de...> - 2007-04-09 14:29:13
|
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi Paul (and m-devel this time, sorry), Paul wrote: > Had a comment aimed my direction about a Morphix derivative for another > project.. www.xenomai.org - The thread in question can be seen at > https://mail.gna.org/public/xenomai-help/2007-04/msg00066.html > > Should we consider the xenomai patches for a regular Morphix kernel, or would > it be preferable to maintain a separate base/main build for these guys ? I think it would be wiser to leave patches outside of the regular base build kernel. As soon as you accept one, you have to accept them all :) You could set up a second autobuilt-base using this kernel instead; as you know it would only require cloop/squashfs and unionfs modules to work. If you run into problems, just send me a ping. cheers, Alex -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.3 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFGGk2sQeuQA5TF/UsRAjssAKCIFrwC/x1ROJgk+pz+KgKSas9JPQCeKSje HLHJoQcAV81XShKykdRuB/E= =qwvl -----END PGP SIGNATURE----- |
|
From: Paul <pa...@tu...> - 2007-04-08 17:07:22
|
Now that Debian has officially released Etch as "stable", Zathras will be getting an upgrade - This in effect means the Sarge autobuilds will be dropped unless there is an overwhelming demand for them. At present, no fixed timetable for the upgrade - I may take the opportunity to install a new MoBo plus hard drive.. Regards, Paul. |
|
From: <ne...@ne...> - 2007-04-08 08:49:57
|
> Should we consider the xenomai patches for a regular Morphix kernel, or w= ould=20 > it be preferable to maintain a separate base/main build for these guys ? >=20 > [I already have kernel & xenomai debs, but without unionfs or squashfs] >=20 Assuming that the xenomai patch isn't intrusive and can't break something in the kernel if booted on a "normal" machine, we can add it. But in general i think that instead of doing two or more separated base modules or to do a single base with patches in the "standard" base kernel that aren't related directly with morphix, we shuld think about=20 doing a base module where one can choose to boot differents kernel. So, maybe a base module with one "regular" kernel AND one "xenomai" kernel, and maybe others, and permit to choose which kernel to boot in the grub menu. Just my 2 eurocents --=20 Franco (nextime) Lanza Busto Arsizio - Italy SIP://ca...@ca... NO TCPA: http://www.no1984.org you can download my public key at: http://danex.nexlab.it/nextime.asc || Key Servers Key ID =3D D6132D50 Key fingerprint =3D 66ED 5211 9D59 DA53 1DF7 4189 DFED F580 D613 2D50 ----------------------------------- echo 16i[q]sa[ln0=3Daln100%Pln100/snlbx]sbA0D212153574F444E4957204553552045= 4D20454B414D204F54204847554F4E452059415020544F4E4E4143205345544147204C4C494= 2snlbxq | dc ----------------------------------- |
|
From: Paul <pa...@tu...> - 2007-04-07 23:37:23
|
Hi Alex Had a comment aimed my direction about a Morphix derivative for another project.. www.xenomai.org - The thread in question can be seen at https://mail.gna.org/public/xenomai-help/2007-04/msg00066.html Should we consider the xenomai patches for a regular Morphix kernel, or would it be preferable to maintain a separate base/main build for these guys ? [I already have kernel & xenomai debs, but without unionfs or squashfs] Regards, Paul. |
|
From: Brendan M. S. <me...@ba...> - 2007-04-06 18:28:05
|
Hi All,
I have had a re-think on the patch I submitted. It works for examples such as:-
<retrieve>
<from>http://foo.com/bar.deb</from>
<to>/tmp/</to>
</retrieve>
However, it will not work for examples such as:-
<retrieve>
<from>http://foo.com/bar.txt</from>
<to>/tmp/foo.txt</to>
</retrieve>
Below is another improved patch. However my c skills are very, very
rusty. I can not remember the last time I did some c. Please can
someone review my 1337 coding, before it is submitted to the svn
repository.
Regards,
Brendan
Index: makemodule.c
===================================================================
--- makemodule.c (revision 2471)
+++ makemodule.c (working copy)
@@ -550,7 +550,17 @@
gchar *http_s = "http://";
if (strncmp(r->from,http_s,strlen(http_s)) == 0) {
- gchar *cmdline = g_strdup_printf("wget
-P %s/%s %s",module->dirname,r->to,r->from);
+ #If the last charactor of r->from is a
/ then we need wget -P else wget -O
+ char * pch;
+ char wgetopt;
+ pch=strrchr(r->from,'/');
+ if (pch-r->from == strlen(r->from)-1) {
+ wgetopt='P';
+ }
+ else {
+ wgetopt='O';
+ }
+ gchar *cmdline = g_strdup_printf("wget
-%c %s/%s %s",wgetopt,module->dirname,r->to,r->from);
if (!ExecuteCommand(cmdline)) {
fprintf(stderr,"Warning:
unable to execute retrieval: %s\n",r->from);
}
|
|
From: Gabriel <sn...@gm...> - 2007-04-06 04:37:33
|
Hi there again ! I found the reason for the strange panic messages when using ccfs. It's related to the pipe between mkisofs and ccfs. I've noticed that everytime ccfs uses stdin/stdou to process the files it gives this warning: "Warning, number as file string found. Assuming old command syntax and choosing compatible parameters (-m -B 0). See the usage info (-h) for better/correct parameters." Even using "module-builder" the warning is there. I tried to make ccfs work with files (create the file.iso and then compress it into file.mod) and it worked. I think it's something related to the block size "65536" argument passed on the command line. This cause ccfs to use the -m option which kills my machine :P . It stops and leaves an incomplete file.mod inside the directory. I am using ccfs without pipe and it's working fine. I just wanted to report this issus as someone else could be in the same situation. Regards, Gabriel. |
|
From: Gabriel <sn...@gm...> - 2007-04-04 18:46:27
|
Thank you for you help :) I decided to give it another try. I added the option "-allow-limited-size" to the make-iso script I have here and it worked. I tested it with vmware and is working fine so far. I'll try on Qemu as well. Regards, Gabriel. Em Qua, 2007-04-04 às 20:34 +0200, Jure Cuhalev escreveu: > On Apr 4, 2007, at 8:19 PM, Gabriel wrote: > > > When I tried to use "module-builder -b source_dir base/mainmod/ > > mmod.mod" > > it starts normaly and suddenly gives my kernel a crash! Lots of kernel > > panic messages for my delight ;-P . After knocking my head a few times > > and reading the man pages for create_compressed_fs I decided to use a > > real file instead of stdin/stdout. So I create an iso from the > > directory > > and then compress this iso with ccfs. Now it works. Next issue ... > > > > Try running a full night full memory test suite memtest86 I believe > it's called. Userland application shouldn't be able to crash the kernel. > > > I tried to use "make-iso base_dir myimage.iso", but I think that > > make-iso doesn't like to deal with large files ;-P . I won't do > > it's job > > and keeps complainning about something related to "file bigger than > > 2Gib" and "no -allow-limited-size option given" . > > > > You can't really go over this with a single compressed file. See for > example wikipedia: http://en.wikipedia.org/wiki/ > ISO_9660#The_2_GB_file_size_limit > > What you could do is modify morphix base to overlay two 2Gb > filesystems. I'm not sure what tax would this have on memory. Your > other option would be to put the big media files on the CD itself and > somehow link to them from your system. > > regards, > Jure Cuhalev > > ------------------------------------------------------------------------- > Take Surveys. Earn Cash. Influence the Future of IT > Join SourceForge.net's Techsay panel and you'll get the chance to share your > opinions on IT & business topics through brief surveys-and earn cash > http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV > _______________________________________________ > Morphix-devel mailing list > Mor...@li... > https://lists.sourceforge.net/lists/listinfo/morphix-devel |
|
From: Jure C. <ga...@ow...> - 2007-04-04 18:34:34
|
On Apr 4, 2007, at 8:19 PM, Gabriel wrote: > When I tried to use "module-builder -b source_dir base/mainmod/ > mmod.mod" > it starts normaly and suddenly gives my kernel a crash! Lots of kernel > panic messages for my delight ;-P . After knocking my head a few times > and reading the man pages for create_compressed_fs I decided to use a > real file instead of stdin/stdout. So I create an iso from the > directory > and then compress this iso with ccfs. Now it works. Next issue ... > Try running a full night full memory test suite memtest86 I believe it's called. Userland application shouldn't be able to crash the kernel. > I tried to use "make-iso base_dir myimage.iso", but I think that > make-iso doesn't like to deal with large files ;-P . I won't do > it's job > and keeps complainning about something related to "file bigger than > 2Gib" and "no -allow-limited-size option given" . > You can't really go over this with a single compressed file. See for example wikipedia: http://en.wikipedia.org/wiki/ ISO_9660#The_2_GB_file_size_limit What you could do is modify morphix base to overlay two 2Gb filesystems. I'm not sure what tax would this have on memory. Your other option would be to put the big media files on the CD itself and somehow link to them from your system. regards, Jure Cuhalev |
|
From: Gabriel <sn...@gm...> - 2007-04-04 18:19:15
|
Hi there friends, I've setup a system full of features and applications. I'm sure this one won't fit on a single CD even after compressing with ccfs. So that's why I'd like to pack it on a DVD. A few interesting things: When I tried to use "module-builder -b source_dir base/mainmod/mmod.mod" it starts normaly and suddenly gives my kernel a crash! Lots of kernel panic messages for my delight ;-P . After knocking my head a few times and reading the man pages for create_compressed_fs I decided to use a real file instead of stdin/stdout. So I create an iso from the directory and then compress this iso with ccfs. Now it works. Next issue ... I tried to use "make-iso base_dir myimage.iso", but I think that make-iso doesn't like to deal with large files ;-P . I won't do it's job and keeps complainning about something related to "file bigger than 2Gib" and "no -allow-limited-size option given" . I think it's possible to build a liveDVD, I just haven't found the way yet :) . My machine is not that great and it takes too much time just to wait for compression and genisoimage to work. Could anyone give me a hint ? Or the exact command to type ? :) Thank you Gabriel. |
|
From: Alex de L. <al...@de...> - 2007-04-03 20:38:29
|
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Brendan M. Sleight wrote: > I am not getting many Autobuilds to feed my AutoTesting machine. The > machine wants to eat - its hungry :) Heh, I've been getting more complaints about the lack of autobuilds. Seems like people _are_ using them after all! I'm considering moving the sid autobuilds to my sarge box, as my current sid box won't be up all that often, but I'll see what I can do over the next few days. If anyone out there has a spare server, we'd love to use it as a build box! (Paul and Brendan need not apply) -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.3 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFGErs0QeuQA5TF/UsRAvM2AJ9K0brsdb5j+z3Lxcx9nph0DYMIMACfbG58 cvZ4h3jjU8SlRqn5+FfKGHQ= =/yL7 -----END PGP SIGNATURE----- |
|
From: Brendan M. S. <me...@ba...> - 2007-04-03 20:13:09
|
Hi Alex, (et al), I am not getting many Autobuilds to feed my AutoTesting machine. The machine wants to eat - its hungry :) Regards, Brendan |
|
From: Paul <pa...@tu...> - 2007-04-01 23:05:14
|
Hi Paras No need to CC me, I read the list... On Sunday 01 April 2007 23:40, Paras pradhan wrote: > Do i add ndiswrapper into the basmod or build myself a basemod using > isomorph? Kernel modules in basemod... > I edited the basemod xml for 2.6.17-morphix and added > > <packagereq>ndiswrapper-utils</packagereq> User space utils in mainmod (or a minimod) as long as the init scripts in base don't use them. > just need few basic tips. That's the way I would go about it... But as Alex pointed out in another thread, there are "issues" with using ndiswrapper - Not the package it's self, but the nature of the M$ binary drivers. These are generally closed, proprietary binaries, and quite likely with restrictive licenses.. Regards, Paul. |
|
From: Paras p. <pra...@gm...> - 2007-04-01 22:40:27
|
Hi Paul, Thanks for your reply. I have few questions. Do i add ndiswrapper into the basmod or build myself a basemod using isomorph? I edited the basemod xml for 2.6.17-morphix and added <packagereq>ndiswrapper-utils</packagereq> I can see ndiswrapper binary when running basemod alone , but when i added my mainmod i don;t see it. I am bit confused on this.. So if possible, please tell me how do i make use of ndiswrapper to run in live CD with mainmod. just need few basic tips. Thanks in advance Paras. On 3/30/07, Paul <pa...@tu...> wrote: > > > Hi Paras > > On Friday 30 March 2007 08:28, Paras pradhan wrote: > > Lots of ppl are demaning wifi support on it. Hence i am seeking help in > > this regard. > > Don't you just love users ;-) > > > I want to use ndiswrapper. > > 1) Which ndiswrapper should i use with kernel 2.6.17 downloaded from > > autobuilds? > > I'd recommend the latest available version. > > > 2) ndiswrapper should be installed in basemod or main mod? > > Installing in basemod would keep all the kernel modules in one place at > the > expense of a slight increase in size. > > <snip> > > > I complied ndiswrapper after booting the livecd with the help of kernel > > headers 2.6.17-morphix.. and while doing modprobe ndiswrapper it throws > me > > something like: Fatal moudle format ndiswrapper.ko.. > > When ever you get a message of this type, look at the output from dmesg. > At a > guess, I suspect it is complaining about a mismatch in the "magic number" > as > a result of using a different compiler version for ndiswrapper - If the > kernel was compiled with gcc-3.4, and ndiswrapper with gcc-4.0, that > message > would be generated. > > > Regards, Paul. > |
|
From: Alex de L. <al...@de...> - 2007-03-31 15:56:25
|
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi all, On the topic of including madwifi and ndiswrapper into Morphix: - - ndiswrapper: I'm all for it. It's free, it's in Debian, and I'll likely include it into the next base. - - madwifi: this is in Debian non-free, and the licensing is rather murky (part-GPL, part-closedsource). It won't be included into the base (at least not if I can help it). Like the nvidia lkm, it should instead be put into a separate minimodule so people can easily make their base 'pure'. Two "buts": - - ndiswrapper wouldn't work without non-free firmware. It might be an idea to include these into a separate minimodule, as to not to have to taint the base. - - to pre-empt the 'Where the hell is the next base?' questions: the unionfs developers are currently working very hard to get unionfs into the mainstream kernel (they're already in the -mm tree). However, their 2.0 version does have a number of limitations (like not being able to add branches). I'm hoping that when 2.6.21 goes gold the unionfs devs will have a solid release for usage by Morphix. To be continued... cheers, Alex Basanta shrestha wrote: > Hi all, > I would like to propose to the morphix developers to include "madwifi" > wireless driver along with the morphix kernel. The very popular "atheros" > chipset works great with this driver. I believe with this driver installed, > atheros wifi cards will work directly out of the box in live cd. Here are > some relevent links -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.3 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFGDnZ+QeuQA5TF/UsRAq81AJ9FI5Kn6JkfOZt9sJ1PA/UHpd9/UgCeKOyT acyRpw4LWIjBd9jD4iOwBWw= =0Rld -----END PGP SIGNATURE----- |
|
From: Brendan M. S. <me...@ba...> - 2007-03-30 12:24:39
|
HI Basanta, > Please check with the licencing matters before considering this. > With respect. If might you useful for the person requesting this extra work to outline the licencing before we consider wifi drivers. For example. Hi My name is Y, Please do X which would have a great deal of benefit to Y people. I hope X is not illegal, do as it is a benefit to me - please can you check to see if X is not illegal. Regards, Y |
|
From: Basanta s. <bas...@gm...> - 2007-03-30 10:49:14
|
Hi all, I would like to propose to the morphix developers to include "madwifi" wireless driver along with the morphix kernel. The very popular "atheros" chipset works great with this driver. I believe with this driver installed, atheros wifi cards will work directly out of the box in live cd. Here are some relevent links http://madwifi.org/ http://www.hpl.hp.com/personal/Jean_Tourrilhes/Linux/Linux.Wireless.drivers.802.11ag.html#AtherosMADWiFi Please check with the licencing matters before considering this. Regards Basanta Shrestha |
|
From: Paul <pa...@tu...> - 2007-03-30 10:05:42
|
Hi Paras On Friday 30 March 2007 08:28, Paras pradhan wrote: > Lots of ppl are demaning wifi support on it. Hence i am seeking help in > this regard. Don't you just love users ;-) > I want to use ndiswrapper. > 1) Which ndiswrapper should i use with kernel 2.6.17 downloaded from > autobuilds? I'd recommend the latest available version. > 2) ndiswrapper should be installed in basemod or main mod? Installing in basemod would keep all the kernel modules in one place at the expense of a slight increase in size. <snip> > I complied ndiswrapper after booting the livecd with the help of kernel > headers 2.6.17-morphix.. and while doing modprobe ndiswrapper it throws me > something like: Fatal moudle format ndiswrapper.ko.. When ever you get a message of this type, look at the output from dmesg. At a guess, I suspect it is complaining about a mismatch in the "magic number" as a result of using a different compiler version for ndiswrapper - If the kernel was compiled with gcc-3.4, and ndiswrapper with gcc-4.0, that message would be generated. Regards, Paul. |
|
From: Paras p. <pra...@gm...> - 2007-03-30 07:28:16
|
HI all, I hope to get help from this mailing list since i have not been able to search for any resources in the internet. We have been working on a localized Linux distro named NepaLinux ( http://www.nepalinux.org) . We have released 2 versions of it and it is going on very well in my country Nepal. It is based on morphix and debian. Lots of ppl are demaning wifi support on it. Hence i am seeking help in this regard. I want to use ndiswrapper. I need help on using ndiswrapper in Live CD mode. Few questions: 1) Which ndiswrapper should i use with kernel 2.6.17 downloaded from autobuilds? 2) ndiswrapper should be installed in basemod or main mod? 3) If any one has tips for me to use ndiswrapper in live CD. please help me out. What i have tried is: I complied ndiswrapper after booting the livecd with the help of kernel headers 2.6.17-morphix.. and while doing modprobe ndiswrapper it throws me something like: Fatal moudle format ndiswrapper.ko.. Help Thanks Paras, |
|
From: <lm...@de...> - 2007-03-12 07:06:51
|
Hello My name is Lara Thynne and I am a PhD candidate at Deakin University Australia. I am currently researching the boundary between work and leisure activities directly related to the open source community and open source program development. As part of this I am running a survey at the following address. https://dcarf.deakin.edu.au/surveys/oss/ The survey is completely confidential and looks at your views and motivations to use Open Source software and to participate in the community. It will only take a five to ten minutes to complete and your contact details will not be recorded. You can withdraw your participation at any stage. I sincerely apologize for the spammish nature of this e-mail - I don't mean to abuse this list. I am trying to collect responses from as many open source developers and users as possible and a mailing list like can be the only way to reach many developers. Thanks again Lara P.S The program that I am using is open source, of course (www.phpsurveyor.org)! |
|
From: Alex de L. <al...@de...> - 2007-03-10 17:26:04
|
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hi Nextime, (sorry for the delay until posting, mailman complained about
your message's subjectline)
The reason for the '--after 0' parameter was because minimod unionfs
branches are read-only. By forcing precedence of the mainmod branch the
whole filesystem remains read/writable.
Note that I haven't thoroughly tested this, so I might be wrong.
Naturally a minimodule would ideally 'overwrite' mainmodule files, not
doing so is needlessly confusing, but afaik it isn't possible without
also blanking out the rw-layer.
Having an option to set the minimod layer position seems reasonable.
Doing so would mean an un-writable filesystem if the above is correct,
but it would at least allow flexibility in this department.
Re-reading your post, I think that if we changed '--after 0' to '--after
1', the result would allow minimod precedence over the mainmod, and
still the filesystem would be writable. Do you think this is correct?
cheers & thanks for looking into this,
Alex
ne...@ne... wrote:
> First of all, sorry for the wrong post on the "cvs" mailing list!
>
> The message was:
>
> Ciao all.
>
> I think that MiniModule unionfs overlaying is conceptually wrong.
>
> in the base module we have 2 init script that mount main and mini
> modules by using unionfs.
>
> While Mainmodule is overlayed in morphix-start in this way:
>
> mount -t unionfs -o dirs=/mnt/main${UNIONFS_MOUNT_OPT},debug=$DEBUGOVERLAY none /mnt/main
> unionctl /mnt/main --add --mode ro /mnt/main_ro
> unionctl /mnt/main --add --mode rw $DIR2/root
>
> MiniModules are overlayed in load-minimodules.sh with:
>
> if [ -d /mnt/main/mnt/mini/mod$count/root ]; then
> unionctl /mnt/main --add --after 0 --mode ro /mnt/main/mnt/mini/mod$count/root
> fi
>
> This make the root file system on the resulting live cd booted like
> this:
>
> root@Medianix:~# unionctl / --list
> /mnt/.hidden_base/ramdisk/var/tmp/overlay/root (rw-)
> /mnt/mini/mod2/root (r--)
> /mnt/.hidden_base/mnt/main_ro (r--)
> root@Medianix:~#
>
>
> as the "0" branch is the rw mounted ramdisk, so, when add to the unionfs
> the MiniModule, it will go after this and before the Main that is
> /mnt/.hidden_base/mnt/main_ro.
>
> In this way if we have 2 files (let me say for an example, /etc/passwd)
> both in main *and* minimodule, the main one take precedence over the
> minimodule one.
>
> While having multiple minimodule with the same file can open some
> issues, i think that it will be better to permit to minimodules
> to take precedence over the mainmodule, so, i think that in
> load-minimodules.sh the unionctl line shuld be
> unionctl /mnt/main --add --mode ro /mnt/main/mnt/mini/mod$count/root
> without the --after 0 switch.
>
> Another idea is to use a file inside the minimodule, something like
> "unionfsposition" that contain the position to insert the minimodule
> branch, letting us configure this from the minimodule itself. Something
> like:
>
> if [ -d /mnt/main/mnt/mini/mod$count/root ]; then
> if [ -f /mnt/main/mnt/mini/mod$count/unionfspos ]; then
> ufsafter="--after `cat /mnt/main/mnt/mini/mod$count/unionfspos`"
> fi
> unionctl /mnt/main --add $ufsafter --mode ro /mnt/main/mnt/mini/mod$count/root
> fi
>
>
>
>
> ------------------------------------------------------------------------
>
> -------------------------------------------------------------------------
> Take Surveys. Earn Cash. Influence the Future of IT
> Join SourceForge.net's Techsay panel and you'll get the chance to share your
> opinions on IT & business topics through brief surveys-and earn cash
> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
>
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> Morphix-devel mailing list
> Mor...@li...
> https://lists.sourceforge.net/lists/listinfo/morphix-devel
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.3 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
iD8DBQFF8uoRQeuQA5TF/UsRAowDAJ9lHKjB7TO2qECjgBZMLz/uQlpkXACfRouI
tV9b5T9HQb09dvonfGaFeGE=
=WHhz
-----END PGP SIGNATURE-----
|
12 messages has been excluded from this view by a project administrator.
