Open GoogleCodeExporter opened 9 years ago
The project is dead for 2 years, I doubt that we'll see new working version
soon :(
Original comment by oyashirosama.hnnkni
on 13 Jun 2013 at 11:39
looks like there's a lot to update for the change from 1.0 to 1.1
It's probably way beyond me, just looking at the restful api docs.
Ugh. I probably never should have given up programming to focus on networking,
right?
Original comment by ak.hep...@gmail.com
on 14 Jun 2013 at 5:10
What's wrong? Will we receive a update?
Original comment by csahuqui...@gmail.com
on 19 Jun 2013 at 8:19
About to lose identi.ca support too when they transition to the pump.io
platform soon
http://identi.ca/doc/pumpio
Original comment by eugene...@gmail.com
on 27 Jun 2013 at 12:54
I'd consider this project not dead, but so well working there were no updates
required over the last two year. If there are other plugins for Twitter, let us
know!
Original comment by m...@jump-ing.de
on 16 Jul 2013 at 10:46
There is also https://code.google.com/p/prpltwtr/ which I actually switched to
from microblog-purple as it worked better but both are now in the same boat of
needing to be updated to the Twitter 1.1 API.
Original comment by bwmer...@gmail.com
on 17 Jul 2013 at 7:38
I made a patch for twitter API 1.1.
This uses json-glib-1.0 at least 0.12. Debian and Ubuntu users need to install
the package libjson-glib-dev to build.
I modified global.mak to link json-glib and it works on ubuntu 14.04, but will
probably work on most other systems aswell. Not tested on windows.
I replaced the default API URLs to the JSON-variants and changed all reply
parsing from XML to JSON.
If you are upgrading, it's best to remove your TwitterIM account and create it
again, so all API URLs are reset to the right ones. Though I doubt anyone is
using mbpidgin for twitter at the moment ;)
Original comment by biosera...@gmail.com
on 13 Jul 2014 at 1:02
Attachments:
Applied the patch, and am now back online.
Haven't seen any bugs creep in- the only downside is that "self" isn't hidden
in the timeline anymore, even if that option is checked.
Not a big deal, simple cosmetics, but this patch so far seems to do the trick
on ubuntu!
biosera ++
Original comment by ak.hep...@gmail.com
on 14 Jul 2014 at 4:22
Forgive my ignorance, but how would I use the patch to update Pidgin running
under Windows? I tried opening the patch file with 7-zip but it doesn't appear
to be an archive.
Original comment by mousebot...@gmail.com
on 14 Jul 2014 at 6:37
The patch file is for the source code. you'd have to have a build environment
all set up for it, and then using the gnu patch utility (or cygwin equivalent)
to apply that text file against the source, and then compile it.
However, the added difficulty for windows is the requirement for libjson-glib.
I haven't done a windows build for a handful of years, so I don't have the dev
environment anymore, and can't help walk you through it.
Original comment by ak.hep...@gmail.com
on 14 Jul 2014 at 7:43
Can't help with windows either, sorry.
Anyone else here who can build and share the DLLs?
Original comment by biosera...@gmail.com
on 14 Jul 2014 at 9:22
I do notice that when I log out of my accounts with mb enable, pidgin crashes.
I don't know if anybody else is seeing that, but i haven't had a chance to dig
into it yet.
Original comment by ak.hep...@gmail.com
on 17 Jul 2014 at 3:46
Applying the patch, 1 hunk failed. mbpurple-0.3.0.tar.gz
{{{
patching file microblog/twitter.c
Hunk #3 FAILED at 311.
Hunk #4 succeeded at 427 (offset -28 lines).
Hunk #5 succeeded at 915 (offset -28 lines).
Hunk #6 succeeded at 962 (offset -28 lines).
Hunk #7 succeeded at 988 (offset -28 lines).
Hunk #8 succeeded at 1029 (offset -28 lines).
Hunk #9 succeeded at 1158 (offset -28 lines).
Hunk #10 succeeded at 1177 (offset -28 lines).
1 out of 10 hunks FAILED -- saving rejects to file microblog/twitter.c.rej
patching file microblog/twitterim.c
Hunk #1 succeeded at 169 (offset -2 lines).
}}}
Original comment by rabes...@gmail.com
on 30 Jul 2014 at 7:09
Did you apply the patch to the svn checkout? It will most likely fail with the
packaged sources from the Downloads.
Original comment by biosera...@gmail.com
on 30 Jul 2014 at 7:17
I used a release version 0.3.0.
Well, I applied your patch semi-manually, and I am using it now. Do you plan to
release your version?
Original comment by rabes...@gmail.com
on 30 Jul 2014 at 9:19
I was actually thinking about a release for a while, so I just made a
repository at github with the code and my patch applied already:
https://github.com/fr0nk/microblog-purple2
There is also a release ready to download:
https://github.com/fr0nk/microblog-purple2/releases
Still no windows binaries, but I will try to build some in the coming days.
Can't be that difficult :) And maybe I'll fix a few other issues.
And FYI I've been using the patch for almost 3 weeks now on two computers and
had no problems yet.
@ak.hep: I can't reproduce the crash on sign out. But I suspect I introduced a
memory leak or two. I will try to fix them and maybe the crashes are gone then.
Original comment by biosera...@gmail.com
on 30 Jul 2014 at 10:51
Nice, thanks biosera! Sorry for asking, but could somebody tell me what I have
to with the tar.gz? I tried to tar -zxvf it, but I'm fairly to new to linux and
I don't really understand the next step to install it.
Original comment by sc0repi0...@gmail.com
on 31 Jul 2014 at 12:35
@sc0repi0...: Read how to make a package in your Linux distribution. If you are
using Arch Linux, you can wait until I create one.
Original comment by rabes...@gmail.com
on 31 Jul 2014 at 11:17
@biosera...
I've got a lot of client protocols open (yahoo, irc, msn, icq, sipe, mbpidgin,
facebook) so it's possible that the memory leak is exacerbated by having all
of these active.
I wish that the kde crash manager would let me save the crashinfo separately,
because then I could upload it...
Original comment by ak.hep...@gmail.com
on 1 Aug 2014 at 4:48
Arch Linux package: https://aur.archlinux.org/packages/microblog-purple2/ .
(The directory name "microblog-purple2-microblog-purple2-0.4" in the archive is
kind of strange.)
Original comment by rabes...@gmail.com
on 9 Aug 2014 at 5:03
@ak.hep: I've pushed a small change to my repo, freeing memory where possible.
That's all I could find and fix. Valgrind doesn't report many memory leaks,
which are caused by the plugin code (but I'm no expert). If this doesn't fix
your crashing issue, then I'm afraid I can't help you there. I'm not that good
a C programmer actually.
On a different note: I've finally got a working windows build environment. So
expect a Windows release soon.
Original comment by biosera...@gmail.com
on 9 Aug 2014 at 7:26
Building a plugin in windows is a major PITA :)
But it's finally done! I've built this on Windows 7 32bit. Hope it runs on all
other windows versions too.
Get it here: https://github.com/fr0nk/microblog-purple2/releases
Original comment by biosera...@gmail.com
on 12 Aug 2014 at 9:36
Unzipped into Pidgin Portable, running under Windows XP. Works OK so far.
Will transfer the stick to a Windows 8.1 (64-bit) machine tomorrow and see if
it still works OK.
Original comment by mousebot...@gmail.com
on 12 Aug 2014 at 10:51
@biosera - Built the new version this morning.
Was going fine, but then I tweeted once, and it crashed. Yikes!
SegvAnalysis:
Segfault happened at: 0x7f2f8a150725 <malloc_consolidate+277>: mov 0x8(%r12),%r13
PC (0x7f2f8a150725) ok
source "0x8(%r12)" (0xfe5f1b3a5190) not located in a known VMA region (needed readable region)!
destination "%r13" ok
SegvReason: reading unknown VMA
Original comment by ak.hep...@gmail.com
on 13 Aug 2014 at 5:29
Oops, sorry about that. I fixed it in the current release.
Please everyone update to the latest release.
Original comment by biosera...@gmail.com
on 13 Aug 2014 at 6:08
rebuilt, with test tweets (@ak_hepcat if you can believe that!) and so far
I haven't had a crash.
I'll keep poking at it.
Original comment by ak.hep...@gmail.com
on 13 Aug 2014 at 6:15
...And it just crashed again. Unfortunately, it didn't generate a core-dump,
so i can't see what's happening.
I guess i can run it in gdb?
Original comment by ak.hep...@gmail.com
on 13 Aug 2014 at 7:11
Program received signal SIGSEGV, Segmentation fault.
__GI___libc_free (mem=0x1) at malloc.c:2929
2929 malloc.c: No such file or directory.
(gdb) dns[10226]: nobody needs me... =(
Original comment by ak.hep...@gmail.com
on 13 Aug 2014 at 7:12
Please run it with pidgin -d and send me the output. Or the last few screens
full of output. There we should see what it was trying to do.
Original comment by biosera...@gmail.com
on 13 Aug 2014 at 7:14
hrm. kept crashing on startup until i disabled twitter auto-login, and then it
came up cleanly.
It's not logging back in right now, for whatever reason either.
I'll keep it logged out for a few minutes, then attempt to log back in.
Original comment by ak.hep...@gmail.com
on 13 Aug 2014 at 7:16
Here's a sanitized last few pages of output before the segv above
Original comment by ak.hep...@gmail.com
on 13 Aug 2014 at 8:00
Attachments:
Though I can't reproduce the crash, I at least know now that it happens in
twitter_decode_messages() or at least the problem starts there.
I think I'm doing something wrong reading arrays in the JSON response.
I'm away for a few days now, so it might take till next weekend until I can do
some more. In the meantime, try the attached patch. I removed a free that was
definitely not needed.
Original comment by biosera...@gmail.com
on 14 Aug 2014 at 9:28
Attachments:
http://twitter.com/ak_hepcat/status/500041993779544064
we'll see how long this goes.
Thanks for the patch
Original comment by ak.hep...@gmail.com
on 14 Aug 2014 at 10:12
FYI, there's still an odd bug somewhere in the freeing, i think.
Most of the time the crashes happen when I'm logging out of the client (but not
quitting it) but if it happens to survive a logout-login cycle, it'll crash in
the middle of a session.
It's not obvious when it'll happen, so it's more than a little harder to grab a
dump, but I'll keep trying.
--
For the windows build, i'm unable to see replies after the "/replies" command is given. The command is accepted, but nothing is printed.
(3.10.9, 32-bit build running on Win7-x64)
Original comment by ak.hep...@gmail.com
on 5 Sep 2014 at 6:00
Has anyone managed to get this working with Pidgin 2.10.10? I was using
microblog-purple2 0.4.1.1 downloaded from
https://github.com/fr0nk/microblog-purple2/releases successfully with 2.10.9
but it seems to hang 2.10.10.
What I did was:
1. Install the plugin files in 2.10.10.
2. Run 2.10.10.
3. Delete existing Twitter accounts.
4. Re-create a Twitter account.
5. Creation process works OK, fires up web browser where I'm invited to log in
and authorise the app. I do that and am presented with a PIN, which I copy.
6. I paste the PIN into the dialogue box and click on OK.
7. Pidgin then hangs and I have to force it to close.
Original comment by mousebot...@gmail.com
on 27 Oct 2014 at 10:43
I updated Pidgin 2.10.9->2.10.10, and it does not hang, and the plugin works.
Original comment by rabes...@gmail.com
on 27 Oct 2014 at 4:02
I upgraded again and this time it worked. It's perhaps worth detailing how I
did it this time in case it makes a difference. I should have said at the
outset that I'm actually using the PortableApps version of Pidgin.
Here's what I did:
1. Allowed the PortableApps update routine to download and install the new
version.
2. I DIDN'T run Pidgin - this may be important.
3. I extracted microblog-purple2-0.4.1.1-win32.zip to the installation again
(as I assumed the update had probably overwritten some of the files).
4. Started Pidgin and all worked as expected. This time I didn't have to
delete and re-create the Twitter accounts - they worked fine as they were.
Original comment by mousebot...@gmail.com
on 1 Nov 2014 at 11:21
Original issue reported on code.google.com by
smkel...@gmail.com
on 12 Jun 2013 at 10:16