Puppy 4.3 Final -- bug reports and suggestions

[ARAN]

Puppy doesn't usually come with git, try this one
http://www.puppylinux.ca/tpp/ttuuxxx/pr ... 3-i386.pet
username:puppy
password:linux
you'll have to type it twice
ttuuxxx

Hello ttuuxxx.

Thanks a lot for the Package.

Have just booted puppy in ram with "puppy pfix=ram" and installed the Git Pet Package that you provided.

For some reason this package is also broken.
It report the same error like the git in the devx sfs.

# git pull
/usr/libexec/git-core//git-pull: line 184: exec: git-merge: not found

Will now download a early devx file that has worked for me and look what for files differ in the git-core folder.

Thanks a lot for the Help.
Greetings ARAN

[technosaurus]

I bashed together a fix for the broken man pages that restores 99% functionality

Code: Select all

#! /bin/bash
defaulthtmlviewer "http://www.google.com/search?&q=man+"$1"+site:linux.die.net&btnI=Search"

[Michalis]

@ARAN that's strange. I don't whether git is included in puppy430 or in the devx (which I have installed) but for me git is working just fine out of the box. Actually I'm using "git clone" but I just tested "git pull" and works well without any message of something missing.

[ARAN]

@ARAN that's strange. I don't whether git is included in puppy430 or in the devx (which I have installed) but for me git is working just fine out of the box. Actually I'm using "git clone" but I just tested "git pull" and works well without any message of something missing.

Hello Michalis.

Thank you a lot for your clarification !
I can confirm that git is not in puppy and also not in devx.
The strange thing is that before i have upgraded Puppy i had a full working git. Dont know anymore where this package i have downloaded.

It would be very helpful if you could try to execute also git-merge
http://linux.die.net/man/1/git-merge

I get this error message here if i do that inside /usr/libexec/git-core

# ./git-merge
bash: git-merge: command not found

Normally this command should exist like all other programms too.
As a example "git commit -a" works great. changes can be commited to the tree. For some reason "git-merge" however is missing.

The place where git-merge get in action is in the File
/usr/libexec/git-core/git-pull

merge_name=$(git fmt-merge-msg $log_arg <"$GIT_DIR/FETCH_HEAD") || exit
test true = "$rebase" &&
exec git-rebase $strategy_args --onto $merge_head \
${oldremoteref:-$merge_head}
exec git-merge $no_stat $no_commit $squash $no_ff $log_arg $strategy_args \
"$merge_name" HEAD $merge_head

I assume the reason why you dont get a merge error is maybe that git has nothing to merge !

This Files or command should somehow work inside /usr/libexec/git-core

Main porcelain commands

git-add(1)
Add file contents to the index.
git-am(1)
Apply a series of patches from a mailbox.
git-archive(1)
Create an archive of files from a named tree.
git-bisect(1)
Find the change that introduced a bug by binary search.
git-branch(1)
List, create, or delete branches.
git-bundle(1)
Move objects and refs by archive.
git-checkout(1)
Checkout and switch to a branch.
git-cherry-pick(1)
Apply the change introduced by an existing commit.
git-citool(1)
Graphical alternative to git-commit.
git-clean(1)
Remove untracked files from the working tree.
git-clone(1)
Clone a repository into a new directory.
git-commit(1)
Record changes to the repository.
git-describe(1)
Show the most recent tag that is reachable from a commit.
git-diff(1)
Show changes between commits, commit and working tree, etc.
git-fetch(1)
Download objects and refs from another repository.
git-format-patch(1)
Prepare patches for e-mail submission.
git-gc(1)
Cleanup unnecessary files and optimize the local repository.
git-grep(1)
Print lines matching a pattern.
git-gui(1)
A portable graphical interface to Git.
git-init(1)
Create an empty git repository or reinitialize an existing one.
git-log(1)
Show commit logs.

git-merge(1)

Join two or more development histories together.
git-mv(1)
Move or rename a file, a directory, or a symlink.
git-pull(1)
Fetch from and merge with another repository or a local branch.
git-push(1)
Update remote refs along with associated objects.
git-rebase(1)
Forward-port local commits to the updated upstream head.
git-reset(1)
Reset current HEAD to the specified state.
git-revert(1)
Revert an existing commit.
git-rm(1)
Remove files from the working tree and from the index.
git-shortlog(1)
Summarize git log output.
git-show(1)
Show various types of objects.
git-stash(1)
Stash the changes in a dirty working directory away.
git-status(1)
Show the working tree status.
git-submodule(1)
Initialize, update or inspect submodules.
git-tag(1)
Create, list, delete or verify a tag object signed with GPG.
gitk(1)
The git repository browser.

Ancillary Commands

Manipulators:

git-config(1)
Get and set repository or global options.
git-fast-export(1)
Git data exporter.
git-fast-import(1)
Backend for fast Git data importers.
git-filter-branch(1)
Rewrite branches.
git-lost-found(1)
(deprecated) Recover lost refs that luckily have not yet been pruned.
git-mergetool(1)
Run merge conflict resolution tools to resolve merge conflicts.
git-pack-refs(1)
Pack heads and tags for efficient repository access.
git-prune(1)
Prune all unreachable objects from the object database.
git-reflog(1)
Manage reflog information.
git-relink(1)
Hardlink common objects in local repositories.
git-remote(1)
manage set of tracked repositories.
git-repack(1)
Pack unpacked objects in a repository.
git-repo-config(1)
(deprecated) Get and set repository or global options.
Interrogators:

git-annotate(1)
Annotate file lines with commit info.
git-blame(1)
Show what revision and author last modified each line of a file.
git-cherry(1)
Find commits not merged upstream.
git-count-objects(1)
Count unpacked number of objects and their disk consumption.
git-fsck(1)
Verifies the connectivity and validity of the objects in the database.
git-get-tar-commit-id(1)
Extract commit ID from an archive created using git-archive.
git-help(1)
display help information about git.
git-instaweb(1)
Instantly browse your working repository in gitweb.
git-merge-tree(1)
Show three-way merge without touching index.
git-rerere(1)
Reuse recorded resolution of conflicted merges.
git-rev-parse(1)
Pick out and massage parameters.
git-show-branch(1)
Show branches and their commits.
git-verify-tag(1)
Check the GPG signature of tags.
git-whatchanged(1)
Show logs with difference each commit introduces.

[ARAN]

Okay with the very smart Guys at the git IRC channell i could fix the Bug.

The Problem is really that the Git Packages for Puppy Linux are Broken.
Even the one that ttuxxx posted it is also broken.

The Bug is that the Pet packages have in the folder /usr/libexec/git-core
about 95 system links missing to the git binary.

It looks like that the pet create scripts dont pay attention to symlinks

Here is a file list of of a complete git-core folder
http://pastie.org/635088

As you can see "git-merge" but also "git-diff", "git-fetch", "git-add" ... that point to git binary are all missing !

I have solved my problem with this command here in the terminal

ln /usr/bin/git /usr/libexec/git-core/git-merge

The symlinks should be created while building git from sources with the following command. For me the Pet Create Packages are Broken as they dont include the symlinks that are needed.

make prefix=/usr INSTALLDIRS=vendor DESTDIR="${pkgdir}" install quick-install-man

[dogone]

Two observations:

1. cdparanoia segfaults in both 4.3 main and small kernel 2.6.30.5. It does not segfault in 4.3 main kernel 2.6.25.16.

2. Ripoff consistently dies in mid full-disk rip. It's cpu utilization then goes from ~1% to 90+% and the process must be killed. Observed in all tested 4.30 final releases.

[ttuuxxx]

Okay with the very smart Guys at the git IRC channell i could fix the Bug.

The Problem is really that the Git Packages for Puppy Linux are Broken.
Even the one that ttuxxx posted it is also broken.

The Bug is that the Pet packages have in the folder /usr/libexec/git-core
about 95 system links missing to the git binary.

It looks like that the pet create scripts dont pay attention to symlinks

Here is a file list of of a complete git-core folder
http://pastie.org/635088

As you can see "git-merge" but also "git-diff", "git-fetch", "git-add" ... that point to git binary are all missing !

I have solved my problem with this command here in the terminal

ln /usr/bin/git /usr/libexec/git-core/git-merge

The symlinks should be created while building git from sources with the following command. For me the Pet Create Packages are Broken as they dont include the symlinks that are needed.

make prefix=/usr INSTALLDIRS=vendor DESTDIR="${pkgdir}" install quick-install-man

usually petget does an excellent job at making system links, actually this is the first time that I've read that it didn't work right. I would be blaming git developers. I've compiled just about everything like gcc,qt,xorg,wx,alsa, etc and not once did I have an issue like that, I've used another git I compiled on 2.14X 2 days ago to download some sources and it work, So its a strange thing then.
I guess I'll have to recompile a few others later on
Thanks for the how to
ttuuxxx

[rerwin]

The referenced thread provides much info on this sound problem and leads me to suspect a modem-driver conflict. The sound board is 8086:24c5, although no one has posted the modem ID. But I see in the modules.alias file three listings for a modem ID 8086:34c6, which supports the observations of an Intel, SmartLink and Conexant (HSF) driver. All three support that ID, which I would guess is what is in the laptops. (Waiting for confirmation.) That narrows down the problem and is manageable!

I suspect the SmartLink (slamr) is superfluous, since the Intel driver results in use of the SmartLink logic (daemon). But the HSF-driver's claim on the modem is surprising. Fortunately, this issue does not seem to involve the snd-hda-intel driver, which is what got modified for 4.3.

Since I have accepted responsibility for modem support, I will look into it. I would like to find a way to discriminate the HSF from the non-HSF modems with that ID. There are more IDs associated with the devices, that show up in the log file: /tmp/udevtrace.log. If each Gateway loptop user in the referenced thread could attach that file to a posting or PM to me, that would help. Thanks.
Richard

UPDATE: To get the best advice (and a lot more), the linmodem group provides a utility to analyze your modem situation, scanModem. I have put the latest version into a dotpet for ease of installation and removal. Click on the dotpet to install it, then click on the "console" icon and enter:
./scanModem
(that is dot slash. . . and upper-case M)

The results appear in (/root/) Modem/ModemData.txt. That file will probably recommend the correct driver to use. You will probably end up blacklisting slamr and hsfmc97ich, to let snd-intel8x0m grab the modem (assuming the IDs are as I guessed).
R

[paulh177]

I think there may be a problem with the network wizard.
I have just looked at the wpa conf file in /etc/network-wizard/wireless/wpa_profiles and it contains over 4000 lines -- it seems the psk gets copied to the wpa config file every time wpa_supplicant runs? (I'm guessing willdly)

this is on my acer aspire one

s

[aarf]

approximately 1 in five shut downs stop at a screen with small rectangular white/grey windows arranged in about 15-20 rows on a black background. requires a hard shutdown to complete.
had seen this only occasionally in previous puppys.
doesn't cause any problems. just the improper shut down screen displays on reboot.

[paulh177]

aarf i get that irregularly too, also seen it in earlier betas

[01micko]

@aarf and paulh177

That might be something to do with what 'gyro' was talking about earlier in this thread whare the pupsave is not cleanly being unmounted. Just a guess.
__________________

Darn absvolume is gone, I read of other people having this bug in the beta's but did not experience it, took a week to go.

Cheers

[lwill]

rerwin,
Attached file has the info you asked for from:
4.30 full (no sound)
4.30 small (sound)
4.03 (sound)
4.31.1 I got working

I have not tested connecting with the modem in any of them.
Let me know if there is any other info I can give.

Lyle

[rhadon]

approximately 1 in five shut downs stop at a screen with small rectangular white/grey windows...

I also can confirm this.I can't say one of five, but it's more often than in earlier Puppies or upup. It seems to appear randomly or I didn't found out any system when it happens.

I don't think it has something to do with pupsave is not cleanly being unmounted because the system hangs before something is saved to the pupsave file.

So if I install something or change settings I save it with the save button on the desktop. After that it doesn't matter whether the system hangs at shutdown or not.

~ Rolf

[Iguleder]

2. Ripoff consistently dies in mid full-disk rip.

Ripoff died here too, but no CPU spikes.

[rerwin]

lwill,
Thanks for all of your log files. They confirm my suspicion. The simplest fix might be to tell puppy to always use the intel8x0m driver for that hardware. If it works, the HSF modems of the same ID would not be supported. For now, that is good enough. Please try inserting a line into /etc/rc.d/MODULESCONFIG, just before the line with only the single-quote ('), line 23, making a new line 23:

Code: Select all

snd_intel8x0m 0x00008086 0x000024c6

(and the single-quote line as 24). Do that in your full installation or a frugal that includes zp430305.sfs.

I need to see only the /tmp/udevtrace.log file, but would like you to also check "lsmod" to see whether module "slamr" actually gets loaded, too. If the sound is restored and slamr is not loaded, please try the modem if you can. If you don't have a dialup ISP, just call yourself or maybe a fax number. If the connection gets a "busy" signal or "no carrier", that's good enough.

EDIT: After editing MODULESCONFIG, be sure to also erase the HSF modem setup. Go into pupdial and probe > ERASE, click OK and "X" to return to pupdial, then exit it and reboot. /EDIT
Richard

[Jim1911]

Barry,

See this http://www.murga-linux.com/puppy/viewtopic.php?t=46902 for the changes needed for 430 to boot with a frugal installation on an ext4 partition.

This should take care of the final ext4 problem. Please also consider adding the capability of saving the pupsave as ext2, ext3, and ext4 depending on what type of partition that is being used, to take full advantage of their capabilities.

Thanks,
Jim

[Michalis]

@ARAN
I'm happy that you managed to solve your problem, but in my installation git is working out of the box with puppy-430 small and devx. To tell you the truth I'm not familiar with git, I know just the very basic and only in order to download something. So because I needed to download from sf the latest ayttm I just typed the git clone blah blah and worked out of the box. I wasn't expecting it but it did. Latter because of your problems I used git pull to test and didn't had any error messages.

Anyway since it works for you now, that's what matters.

[Iguleder]

I got this:

daemon.notice modprobe: WARNING: Could not open '/lib/modules/2.6.30.5/kernel/drivers/parport/parport.ko': No such file or directory

And because of that, also this:

daemon.notice modprobe: FATAL: Error inserting lp (/lib/modules/2.6.30.5/kernel/drivers/char/lp.ko): Unknown symbol in module, or unknown parameter (see dmesg)

Fix:
ln -s /lib/modules/2.6.30.5/initrd/kernel/drivers/parport/parport.ko /lib/modules/2.6.30.5/kernel/drivers/parport

[lwill]

rerwin,
Did as you asked. No sound.
Tried modem, did not dial*, still shows ttySHSF0.

*EDIT: sorry my bad, modem does work, they changed to 10 digit dialing and I forgot to dial 1.Too use to using the cell.

Here is udevtrace and lsmod for you to see.
I will reboot into the version I have working and see if the modem works there and report back here. Edit: cannot find a modem.

Lyle

Update: I tried a bit more investigation on my own. I commented out the line that disables the snd_intel8x0m in modem_hsfmodem.conf Reboot. Sound worked and modem stopped. hsf modules got still loaded.
So while it may conflict with the hsf stuff, I seem to need snd_intel8x0m for sound to work, which is more important to me at the moment.
If there is anything else you would like me to try, let me know.