LibreOffice + language packs (latest version: 6.1.4)

Word processors, spreadsheets, presentations, translation, etc.
Message
Author
User avatar
Puppus Dogfellow
Posts: 1667
Joined: Tue 08 Jan 2013, 01:39
Location: nyc

#46 Post by Puppus Dogfellow »

machine finally decided it was in need of an update, and i couldn't resist, so:

LO414pet, us-en, 195mb
LO414sfs, us-en, 163mb

414 is arguably a nicer number...

User avatar
boxR
Posts: 338
Joined: Sat 13 Aug 2011, 21:58
Location: France

#47 Post by boxR »

LibreOffice 4.1.4 fr (PET) is now available here :

User avatar
vicmz
Posts: 1262
Joined: Sun 15 Jan 2012, 22:47

#48 Post by vicmz »

Thank you all! Everything added to the first post.
[url=http://murga-linux.com/puppy/viewtopic.php?t=76948]Puppy Linux en español[/url]

ac2011
Posts: 134
Joined: Wed 09 Feb 2011, 08:22

v. 4.1.x won't work in older Puppies

#49 Post by ac2011 »

I've read the note on the first page that versions past 4.1.x won't work on older Puppies. I can confirm that's true. I had 4.0.3 working fine on Fluppy, but 4.1.2 and 4.1.4 won't run, even after recompressing the SFS to use gzip rather than xz compression.

If I run /opt/libreoffice4.1/program/soffice.bin I get either "Segmentation fault" or "Trace/breakpoint trap" or occasionally "Illegal instruction".

Is there no fix for this? It's not a big deal, I'm just curious to know what's changed with the v. 4.1.x releases and whether I can hack it to work...

ac2011
Posts: 134
Joined: Wed 09 Feb 2011, 08:22

Re: v. 4.1.x won't work in older Puppies

#50 Post by ac2011 »

ac2011 wrote:I've read the note on the first page that versions past 4.1.x won't work on older Puppies. I can confirm that's true. I had 4.0.3 working fine on Fluppy, but 4.1.2 and 4.1.4 won't run, even after recompressing the SFS to use gzip rather than xz compression.

If I run /opt/libreoffice4.1/program/soffice.bin I get either "Segmentation fault" or "Trace/breakpoint trap" or occasionally "Illegal instruction".

Is there no fix for this? It's not a big deal, I'm just curious to know what's changed with the v. 4.1.x releases and whether I can hack it to work...
Actually even 4.0.6 doesn't work - same errors, even with pfix=ram. So 4.0.3 is the last one that'll work on this machine. I'd love to know why.

User avatar
Puppus Dogfellow
Posts: 1667
Joined: Tue 08 Jan 2013, 01:39
Location: nyc

Re: v. 4.1.x won't work in older Puppies

#51 Post by Puppus Dogfellow »

ac2011 wrote:
ac2011 wrote:I've read the note on the first page that versions past 4.1.x won't work on older Puppies. I can confirm that's true. I had 4.0.3 working fine on Fluppy, but 4.1.2 and 4.1.4 won't run, even after recompressing the SFS to use gzip rather than xz compression.

If I run /opt/libreoffice4.1/program/soffice.bin I get either "Segmentation fault" or "Trace/breakpoint trap" or occasionally "Illegal instruction".

Is there no fix for this? It's not a big deal, I'm just curious to know what's changed with the v. 4.1.x releases and whether I can hack it to work...
Actually even 4.0.6 doesn't work - same errors, even with pfix=ram. So 4.0.3 is the last one that'll work on this machine. I'd love to know why.
have you tried the portable apps version? they're apparently up to 4.1.4 in that format. i've never personally actually ever used this format, but i can remember coming across various favorable things about it. in fact, i think the person who made the dropbox pet i've been using for the last two years or so derived it from that platform...

anyway, here's the link:

http://portableapps.com/apps/office/lib ... e_portable


have you tried the pet version? i've had machines that work better one way than the other. I'm using precise 5.5 right now on this machine. around the time i upgraded LO from 3 to 4 on it, i lost the ability to cleanly install/buglessly use libreoffice sfs files--too many remnants of old installs or something. i got rid of what i could, went the pet route, and no problems since.

think i'll try to sfs route next time, just to see if the problem's healed...

User avatar
vicmz
Posts: 1262
Joined: Sun 15 Jan 2012, 22:47

#52 Post by vicmz »

ac2011,

I think it may have to do with glibc, a glibc upgrade lets you run new software on older Puppy but it can have side effects. I you want to give it a go, here's a thread with information on updating glibc:

http://murga-linux.com/puppy/viewtopic.php?t=66373
[url=http://murga-linux.com/puppy/viewtopic.php?t=76948]Puppy Linux en español[/url]

ac2011
Posts: 134
Joined: Wed 09 Feb 2011, 08:22

#53 Post by ac2011 »

vicmz wrote:ac2011,

I think it may have to do with glibc, a glibc upgrade lets you run new software on older Puppy but it can have side effects. I you want to give it a go, here's a thread with information on updating glibc:

http://murga-linux.com/puppy/viewtopic.php?t=66373
I don't think it's that. GlibC versions are the same in Fluppy and Carolina and it works fine in Carolina. I have tried tweaking GlibC for LibreOffice in Fluppy (as per my post halfway down this page: http://www.murga-linux.com/puppy/viewto ... start=1065) but I get the same result: segmentation fault.

I'd like to know the underlying reason. It might be possible to work around it.

ac2011
Posts: 134
Joined: Wed 09 Feb 2011, 08:22

Re: v. 4.1.x won't work in older Puppies

#54 Post by ac2011 »

Puppus Dogfellow wrote:
ac2011 wrote:
ac2011 wrote:I've read the note on the first page that versions past 4.1.x won't work on older Puppies. I can confirm that's true. I had 4.0.3 working fine on Fluppy, but 4.1.2 and 4.1.4 won't run, even after recompressing the SFS to use gzip rather than xz compression.

If I run /opt/libreoffice4.1/program/soffice.bin I get either "Segmentation fault" or "Trace/breakpoint trap" or occasionally "Illegal instruction".

Is there no fix for this? It's not a big deal, I'm just curious to know what's changed with the v. 4.1.x releases and whether I can hack it to work...
Actually even 4.0.6 doesn't work - same errors, even with pfix=ram. So 4.0.3 is the last one that'll work on this machine. I'd love to know why.
have you tried the portable apps version? they're apparently up to 4.1.4 in that format. i've never personally actually ever used this format, but i can remember coming across various favorable things about it. in fact, i think the person who made the dropbox pet i've been using for the last two years or so derived it from that platform...

anyway, here's the link:

http://portableapps.com/apps/office/lib ... e_portable


have you tried the pet version? i've had machines that work better one way than the other. I'm using precise 5.5 right now on this machine. around the time i upgraded LO from 3 to 4 on it, i lost the ability to cleanly install/buglessly use libreoffice sfs files--too many remnants of old installs or something. i got rid of what i could, went the pet route, and no problems since.

think i'll try to sfs route next time, just to see if the problem's healed...
Thanks, I might give that a try. There's definitely something fundamentally changed, though. Even with a clean boot (pfix=ram) the segmentation fault remains.

User avatar
mikeb
Posts: 11297
Joined: Thu 23 Nov 2006, 13:56

#55 Post by mikeb »

hi ac2011

ok you aroused my curiosity so I did some tests to add to your findings.

Testing on 4.12.... segmentation fault. Same machine with Lucid is fine so not a cpu issue.

tried glibc 2.10 ...same.

Tried gcc change...same

tried gtk 2.14 same... gtk2.18 ibiblio gave bad pet so could not test..I suspect this may be a likely cause of the problem.

Glib 2.18 ... same

did try using mozilla libs and lincurl from the pup rather than supplied... same

Sort of ran out of ideas and surprised as the build in general seems to be built to be compatible.

Tried with and without java and the included python 3 runs fine.

If anything else occurs to me I will post.

mike

ac2011
Posts: 134
Joined: Wed 09 Feb 2011, 08:22

#56 Post by ac2011 »

mikeb wrote:hi ac2011

ok you aroused my curiosity so I did some tests to add to your findings.

Testing on 4.12.... segmentation fault. Same machine with Lucid is fine so not a cpu issue.

tried glibc 2.10 ...same.

Tried gcc change...same

tried gtk 2.14 same... gtk2.18 ibiblio gave bad pet so could not test..I suspect this may be a likely cause of the problem.

Glib 2.18 ... same

did try using mozilla libs and lincurl from the pup rather than supplied... same

Sort of ran out of ideas and surprised as the build in general seems to be built to be compatible.

Tried with and without java and the included python 3 runs fine.

If anything else occurs to me I will post.

mike
Thanks for taking the time to do that. Nice to know it's not just me ;-)

According to Claws-mail I'm running GTK+ 2.22.1 / GLib 2.27.4 in Fluppy. Carolina runs GTK+ 2.24 from memory.

You think a later GTK+ may be required? That would make sense. If I can find a later PET I might try it on a test system (I'm aware this will probably bork everything else).

User avatar
mikeb
Posts: 11297
Joined: Thu 23 Nov 2006, 13:56

#57 Post by mikeb »

According to Claws-mail I'm running GTK+ 2.22.1 / GLib 2.27.4 in Fluppy. Carolina runs GTK+ 2.24 from memory.
hmm thats would tend to rule out that possibility. My gcc test was a bit rushed too since I was having problems with pets from ibiblio.

Google did not come up with anything that seemed relevant either.

I will sit on it for a while ...segmentation faults sort of narrow down what to look for but at the same time give no clues of specifically where to look.

I have stuck with 3.6.4 for now as i have yet to find any quirks but at some point would probably want to update.

My reapproach might involve something like swapping libs from Lucid one at a time....you might want to try a similar method with your two pups.

I did an ldd of soffice.bin to get a library list and focused on the shared ones rather than the ones that come withe libreoffice.

mike

ac2011
Posts: 134
Joined: Wed 09 Feb 2011, 08:22

#58 Post by ac2011 »

I installed gtk+-2.24.8-w5c PET and Claws now reports GTK+ 2.24.8 / GLib 2.27.4 so I think it 'took' OK.

Using the latest 4.1.4 SFS (expanded and then recompressed with GZIP so Fluppy can use it) I still get a segmentation fault. Same thing with my GlibC 2.11.3 tweak.

Running ldd /opt/libreoffice4.1/program/soffice.bin gives me this:

Code: Select all

	linux-gate.so.1 =>  (0xffffe000)
	libuno_sal.so.3 => /opt/libreoffice4.1/program/../ure-link/lib/libuno_sal.so.3 (0xb777a000)
	libsofficeapp.so => /opt/libreoffice4.1/program/libsofficeapp.so (0xb7720000)
	libc.so.6 => /lib/libc.so.6 (0xb75e0000)
	libpthread.so.0 => /lib/libpthread.so.0 (0xb75c9000)
	libdl.so.2 => /lib/libdl.so.2 (0xb75c4000)
	librt.so.1 => /lib/librt.so.1 (0xb75bc000)
	libstdc++.so.6 => /usr/lib/libstdc++.so.6 (0xb74d5000)
	libm.so.6 => /lib/libm.so.6 (0xb74ad000)
	libgcc_s.so.1 => /usr/lib/libgcc_s.so.1 (0xb74a2000)
	libcomphelper.so => /opt/libreoffice4.1/program/libcomphelper.so (0xb736a000)
	libuno_cppu.so.3 => /opt/libreoffice4.1/program/../ure-link/lib/libuno_cppu.so.3 (0xb7345000)
	libuno_cppuhelpergcc3.so.3 => /opt/libreoffice4.1/program/../ure-link/lib/libuno_cppuhelpergcc3.so.3 (0xb72a6000)
	libdeploymentmisclo.so => /opt/libreoffice4.1/program/libdeploymentmisclo.so (0xb727f000)
	libi18nlangtag.so => /opt/libreoffice4.1/program/libi18nlangtag.so (0xb7254000)
	libuno_salhelpergcc3.so.3 => /opt/libreoffice4.1/program/../ure-link/lib/libuno_salhelpergcc3.so.3 (0xb724c000)
	libsfxlo.so => /opt/libreoffice4.1/program/libsfxlo.so (0xb6e4d000)
	libsvllo.so => /opt/libreoffice4.1/program/libsvllo.so (0xb6d72000)
	libsvtlo.so => /opt/libreoffice4.1/program/libsvtlo.so (0xb6a73000)
	libtklo.so => /opt/libreoffice4.1/program/libtklo.so (0xb67ad000)
	libtllo.so => /opt/libreoffice4.1/program/libtllo.so (0xb6740000)
	libucbhelper.so => /opt/libreoffice4.1/program/libucbhelper.so (0xb66db000)
	libutllo.so => /opt/libreoffice4.1/program/libutllo.so (0xb65ed000)
	libvcllo.so => /opt/libreoffice4.1/program/libvcllo.so (0xb60a8000)
	/lib/ld-linux.so.2 (0xb77c8000)
	libicuuc.so.51 => /opt/libreoffice4.1/program/libicuuc.so.51 (0xb5f73000)
	libreglo.so => /opt/libreoffice4.1/program/../ure-link/lib/libreglo.so (0xb5f59000)
	libunoidllo.so => /opt/libreoffice4.1/program/../ure-link/lib/libunoidllo.so (0xb5f34000)
	libxmlreaderlo.so => /opt/libreoffice4.1/program/../ure-link/lib/libxmlreaderlo.so (0xb5f2b000)
	libxmlscriptlo.so => /opt/libreoffice4.1/program/libxmlscriptlo.so (0xb5eb8000)
	libxml2.so.2 => /opt/libreoffice4.1/program/../ure-link/lib/libxml2.so.2 (0xb5d56000)
	libbasegfxlo.so => /opt/libreoffice4.1/program/libbasegfxlo.so (0xb5ce4000)
	libdrawinglayerlo.so => /opt/libreoffice4.1/program/libdrawinglayerlo.so (0xb5bcc000)
	libfwelo.so => /opt/libreoffice4.1/program/libfwelo.so (0xb5b56000)
	libsaxlo.so => /opt/libreoffice4.1/program/libsaxlo.so (0xb5b3d000)
	libsblo.so => /opt/libreoffice4.1/program/libsblo.so (0xb59e0000)
	libsotlo.so => /opt/libreoffice4.1/program/libsotlo.so (0xb598f000)
	libi18nutil.so => /opt/libreoffice4.1/program/libi18nutil.so (0xb5979000)
	libjvmfwklo.so => /opt/libreoffice4.1/program/../ure-link/lib/libjvmfwklo.so (0xb5963000)
	libX11.so.6 => /usr/X11R7/lib/libX11.so.6 (0xb5865000)
	libz.so.1 => /lib/libz.so.1 (0xb5856000)
	libcups.so.2 => /usr/lib/libcups.so.2 (0xb582a000)
	libdbus-glib-1.so.2 => /usr/lib/libdbus-glib-1.so.2 (0xb580e000)
	libdbus-1.so.3 => /usr/lib/libdbus-1.so.3 (0xb57d7000)
	libgobject-2.0.so.0 => /usr/lib/libgobject-2.0.so.0 (0xb57a4000)
	libglib-2.0.so.0 => /usr/lib/libglib-2.0.so.0 (0xb56d9000)
	libfontconfig.so.1 => /usr/lib/libfontconfig.so.1 (0xb56ab000)
	libfreetype.so.6 => /usr/lib/libfreetype.so.6 (0xb5636000)
	libnspr4.so => /opt/libreoffice4.1/program/libnspr4.so (0xb55f6000)
	libnss3.so => /opt/libreoffice4.1/program/libnss3.so (0xb54a1000)
	libsmime3.so => /opt/libreoffice4.1/program/libsmime3.so (0xb5474000)
	libjvmaccesslo.so => /opt/libreoffice4.1/program/../ure-link/lib/libjvmaccesslo.so (0xb546d000)
	libicule.so.51 => /opt/libreoffice4.1/program/libicule.so.51 (0xb541e000)
	liblcms2.so.2 => /opt/libreoffice4.1/program/liblcms2.so.2 (0xb53d7000)
	libicudata.so.51 => /opt/libreoffice4.1/program/libicudata.so.51 (0xb3e8d000)
	libstorelo.so => /opt/libreoffice4.1/program/../ure-link/lib/libstorelo.so (0xb3e77000)
	libcanvastoolslo.so => /opt/libreoffice4.1/program/libcanvastoolslo.so (0xb3e45000)
	libcppcanvaslo.so => /opt/libreoffice4.1/program/libcppcanvaslo.so (0xb3e03000)
	libfwilo.so => /opt/libreoffice4.1/program/libfwilo.so (0xb3dd6000)
	libxcb.so.1 => /usr/X11R7/lib/libxcb.so.1 (0xb3dc1000)
	libssl.so.0.9.8 => /usr/lib/libssl.so.0.9.8 (0xb3d8d000)
	libcrypto.so.0.9.8 => /usr/lib/libcrypto.so.0.9.8 (0xb3c89000)
	libcrypt.so.1 => /lib/libcrypt.so.1 (0xb3c56000)
	libpcre.so.0 => /usr/lib/libpcre.so.0 (0xb3c27000)
	libgthread-2.0.so.0 => /usr/lib/libgthread-2.0.so.0 (0xb3c23000)
	libexpat.so.1 => /usr/lib/libexpat.so.1 (0xb3c07000)
	libnssutil3.so => /opt/libreoffice4.1/program/libnssutil3.so (0xb3be4000)
	libplc4.so => /opt/libreoffice4.1/program/libplc4.so (0xb3bdf000)
	libplds4.so => /opt/libreoffice4.1/program/libplds4.so (0xb3bdb000)
	libXau.so.6 => /usr/X11R7/lib/libXau.so.6 (0xb3bd8000)
which tells me very little, but someone with a bit more knowledge might be able to see a glaring issue.

User avatar
mikeb
Posts: 11297
Joined: Thu 23 Nov 2006, 13:56

#59 Post by mikeb »

Not a glaring issue, just a list of libraries in use which could be tested one by one...yes its as tedious as it sounds though so far several have been eliminated as a cause.

Its a bit like fumbling in the dark and otherwise any debug info for segemntation faults would come in the form of memory dumps though it might point to a specific libray...unfortunately i am unsure of how to obtain that either.

Gumboots needed

mike

ac2011
Posts: 134
Joined: Wed 09 Feb 2011, 08:22

#60 Post by ac2011 »

mikeb wrote:Not a glaring issue, just a list of libraries in use which could be tested one by one...yes its as tedious as it sounds though so far several have been eliminated as a cause.

Its a bit like fumbling in the dark and otherwise any debug info for segemntation faults would come in the form of memory dumps though it might point to a specific libray...unfortunately i am unsure of how to obtain that either.

Gumboots needed

mike
Got it :)

libgcc_s.so.1 from Carolina 1.1 allows me to run LibreOffice 4.1.4.2 on Fluppy. I haven't tested it extensively, but the program runs and seems fine on first inspection.

Here's the file. For safety, best to keep it separate to the rest of your system and invoke with, eg. LD_LIBRARY_PATH=/path/to/carolina_gcc_lib:$LD_LIBRARY_PATH and then launch LibreOffice.

Hope this helps others.
Attachments
libgcc_s.so.1_carolina.zip
(21.39 KiB) Downloaded 539 times

User avatar
mikeb
Posts: 11297
Joined: Thu 23 Nov 2006, 13:56

#61 Post by mikeb »

libgcc_s.so.1 from Carolina 1.1 allows me to run LibreOffice 4.1.4.2 on Fluppy. I haven't tested it extensively, but the program runs and seems fine on first inspection.
That was a strong suspicion and my test of it was not good as I was unsure if the pets I tested were even installing.

It is one of those segmentation fault inducing libs as its core to the compiling process and binaries built with gcc are supposed to match it in some way...not an area of expertise on my part.

Nice find mate.... I will do a retest using your library...or at least one thats ok on 4.12 if it does not get on. (or use LDLibrary as you suggest)

I like a bit of future proofing :)

You definately get a puppy tee shirt....

One thing I wonder is if altering libgcc would affect anything else ...indeed compiling comes to mind...but perhaps it could be added to the open office install so it gets used instead of the native one...I will play now we have something to work with :)

mike

User avatar
mikeb
Posts: 11297
Joined: Thu 23 Nov 2006, 13:56

#62 Post by mikeb »

Ok tested in 4.12.... added your libgcc.so.1 to /opt/libreoffice4.1/program and after a first stalled run seems to be fine. Gtk/glib spits out a few warnings but hey whats new :D.
Doing that seems to avoid the need to use LD_LIBRARY too.

Ok curiosity made me look at my copy of 3.4.6 ..sure enough it does have libgcc.so.1 but in ure/lib instead (layout is a bit different anyway) This was built directly from their debs by me as it happens.
It actually ran without it on puppy 4.12 but that library used for libreoffice 4.14 failed so I assume its an older version than yours.

As an aside renaming javaldx when java is not present saved 2 seconds of start up. Click to writer takes 4 seconds on a pentium 3...not bad at all really. That was run from a hard drive extraction rather than an lzma sfs which I normally use so I guess the size of the program benefits from not being compressed.

mystery solved sealed and delivered :)

mike

ac2011
Posts: 134
Joined: Wed 09 Feb 2011, 08:22

#63 Post by ac2011 »

Mine did that first stalled run too, then it's worked fine since.

I tend to keep the 'odd' libraries separate so I remember what they're for and, as you say, so they don't interfere with the rest of the system. So I might copy this one into /root/carolina_lib_for_libreoffice/ and then change defaultwordprocessor so that it does something like:

Code: Select all

export LD_LIBRARY_PATH=/root/carolina_lib_for_libreoffice:$LD_LIBRARY_PATH
libreoffice4.1 --writer "$@"
(or whatever the syntax is - I'm writing this from memory)

Nice to have this solved.

User avatar
mikeb
Posts: 11297
Joined: Thu 23 Nov 2006, 13:56

#64 Post by mikeb »

Does not make documents and accounts any more exiting though...wonder if there is a fix for that :D

mike

YoutIndia
Posts: 14
Joined: Tue 12 Feb 2013, 15:24

Works Fantastic on Puppy Precise 5.7.1 Thanks

#65 Post by YoutIndia »

Just a thankyou message. I am so much surprised that a lightweight distro like puppy could run the latest software like Libreoffice without any difficulty. So, I thought I owe you a thanks.

Thanks for sharing Libreoffice.

Post Reply