Code::Blocks Forums

User forums => Nightly builds => Topic started by: killerbot on August 30, 2014, 05:30:14 pm

Title: The 30 August 2014 build (9884) is out.
Post by: killerbot on August 30, 2014, 05:30:14 pm
Get quick announcements through the RSS feed http://www.codeblocks.org/nightly/CodeBlock_RSS.xml

Before you use a nightly make sure you understand how it works (http://forums.codeblocks.org/index.php/topic,3232.0.html).

A link to the unicode windows wxWidget dll for Code::Blocks : http://sourceforge.net/projects/codeblocks/files/Binaries/Nightlies/Prerequisites/wxmsw28u_gcc_cb_wx2812_gcc481-TDM.7z

For those who might need this one (when no MingW installed on your system) : the mingw10m.dll : http://sourceforge.net/projects/codeblocks/files/Binaries/Nightlies/Prerequisites/mingwm10_gcc481-TDM.7z

The 30 August 2014 build is out.
  - Windows :
   http://sourceforge.net/projects/codeblocks/files/Binaries/Nightlies/2014/CB_20140830_rev9884_win32.7z
  - Linux :
   none

Resolved Fixed:


Regressions/Confirmed/Annoying/Common bugs:


Title: Re: The 30 August 2014 build (9884) is out.
Post by: shurick on August 31, 2014, 07:07:13 am
Packages for openSUSE (http://codeblocks.esy.es) (binaries and sources) for 32-bit and 64-bit.
Title: Re: The 30 August 2014 build (9884) is out.
Post by: mynameis on September 01, 2014, 06:48:13 am
i fonund that no ubuntu series  packages for a long time
Title: Re: The 30 August 2014 build (9884) is out.
Post by: edison on September 01, 2014, 07:07:19 am
I have try to build the svn build 9894, but the fortranproject was not present in the packages which downloaded from sf.

edit:
I know the reason now(it seems googlecode was banned in here).
Title: Re: The 30 August 2014 build (9884) is out.
Post by: xawari on September 01, 2014, 10:10:05 am
Let's check out last summer release...
Title: Re: The 30 August 2014 build (9884) is out.
Post by: Jenna on September 01, 2014, 01:55:39 pm
Quote from: mynameis on September 01, 2014, 06:48:13 am
i fonund that no ubuntu series  packages for a long time
You can try my (debian-)repo. They can work for ubuntu, but depends on the ubuntu-version.

By the way:
Debian packages (binaries and sources) for 32-bit and 64-bit systems can be found in my debian-repo (http://apt.jenslody.de/).
Fedora packages (binaries and sources) for 32-bit and 64-bit systems (fc19, fc20 and rawhide), RedHat/CentOS 5 and 6 packages (also 32-bit and 64-bit) and RedHat/CentOS 7 packages (only 64-bit at the moment) can be found in my rpm-repo (http://rpm.jenslody.de).
Title: Re: The 30 August 2014 build (9884) is out.
Post by: SteelRat on September 01, 2014, 03:33:43 pm
If stripping of whitespaces on saving is enabled and tab char used for indents, there is a little bug - when cursor is on blank line with one or more tabs i'm pressing Ctrl-S - tabs are stripped - i'm starting to type and line is filling by spaces to cursor position, but there must be tab characters.
Title: Re: The 30 August 2014 build (9884) is out.
Post by: oBFusCATed on September 01, 2014, 08:45:49 pm
Can you add a ticket in the tracker on sf.net?
Please include a bit more detailed explanation.
Title: Re: The 30 August 2014 build (9884) is out.
Post by: oBFusCATed on September 01, 2014, 09:15:05 pm
Quote
$ dpkg-buildpackage -us -uc
dpkg-buildpackage: source package codeblocks
dpkg-buildpackage: source version 13.12svn9513
dpkg-buildpackage: source distribution unstable
dpkg-buildpackage: source changed by Jens Lody <jens@codeblocks.org>
dpkg-buildpackage: host architecture amd64
 dpkg-source --before-build git
dpkg-checkbuilddeps: Unmet build dependencies: libstdc++6-4.3-dev | libstdc++6-4.4-dev | libstdc++6-4.5-dev | libstdc++6-4.6-dev
dpkg-buildpackage: warning: build dependencies/conflicts unsatisfied; aborting
dpkg-buildpackage: warning: (Use -d flag to override.)
obf@obf-VirtualBox ~/codeblocks/git $ sudo apt-get install libstdc++-dev
[sudo] password for obf:
Reading package lists... Done
Building dependency tree       
Reading state information... Done
Package libstdc++-dev is a virtual package provided by:
  libstdc++6-4.6-dev 4.6.4-6ubuntu2
  libstdc++6-4.4-dev 4.4.7-8ubuntu1
  libc++-dev 1.0~svn199600-1
  libstdc++6-4.7-dev 4.7.3-12ubuntu1
  libstdc++-4.8-dev 4.8.2-19ubuntu1
You should explicitly select one to install.

E: Package 'libstdc++-dev' has no installation candidate

Isn't it possible to make this command pick the appropriate libstdc++ library for my compiler (4.8.1 in my case)?
This is a mint-lastest installation inside virtual box.
Title: Re: The 30 August 2014 build (9884) is out.
Post by: cacb on September 01, 2014, 09:41:38 pm
Quote from: mynameis on September 01, 2014, 06:48:13 am
i fonund that no ubuntu series  packages for a long time

I am on Kubuntu, and I found it convenient to build C::B from Jens Lody's preconfigured source tarball. Works every time.
Just follow the few simple steps at http://forums.codeblocks.org/index.php/topic,18580.msg127254.html#msg127254
Title: Re: The 30 August 2014 build (9884) is out.
Post by: Jenna on September 01, 2014, 10:28:18 pm
Quote from: oBFusCATed on September 01, 2014, 09:15:05 pm
Quote
$ dpkg-buildpackage -us -uc
dpkg-buildpackage: source package codeblocks
dpkg-buildpackage: source version 13.12svn9513
dpkg-buildpackage: source distribution unstable
dpkg-buildpackage: source changed by Jens Lody <jens@codeblocks.org>
dpkg-buildpackage: host architecture amd64
 dpkg-source --before-build git
dpkg-checkbuilddeps: Unmet build dependencies: libstdc++6-4.3-dev | libstdc++6-4.4-dev | libstdc++6-4.5-dev | libstdc++6-4.6-dev
dpkg-buildpackage: warning: build dependencies/conflicts unsatisfied; aborting
dpkg-buildpackage: warning: (Use -d flag to override.)
obf@obf-VirtualBox ~/codeblocks/git $ sudo apt-get install libstdc++-dev
[sudo] password for obf:
Reading package lists... Done
Building dependency tree       
Reading state information... Done
Package libstdc++-dev is a virtual package provided by:
  libstdc++6-4.6-dev 4.6.4-6ubuntu2
  libstdc++6-4.4-dev 4.4.7-8ubuntu1
  libc++-dev 1.0~svn199600-1
  libstdc++6-4.7-dev 4.7.3-12ubuntu1
  libstdc++-4.8-dev 4.8.2-19ubuntu1
You should explicitly select one to install.

E: Package 'libstdc++-dev' has no installation candidate

Isn't it possible to make this command pick the appropriate libstdc++ library for my compiler (4.8.1 in my case)?
This is a mint-lastest installation inside virtual box.
It should work if you remove the line completely from debian/control.
g++ should depend on the correct libstdc++-dev automatically,
and g++ is a dependency of the build-essential-package which is an automatic dependency of the debian build-system.

I don't know if it also works for older revisions of debian-based distro's, but on wheezy it seems to work.
Title: Re: The 30 August 2014 build (9884) is out.
Post by: SteelRat on September 02, 2014, 08:40:50 am
Quote from: oBFusCATed on September 01, 2014, 08:45:49 pm
Can you add a ticket in the tracker on sf.net?
Please include a bit more detailed explanation.
Yes, i can. By the way, do you understand my english easily or with problems? =)
Title: Re: The 30 August 2014 build (9884) is out.
Post by: oBFusCATed on September 02, 2014, 11:48:36 pm
Your English is fine...
Title: Re: The 30 August 2014 build (9884) is out.
Post by: mynameis on September 03, 2014, 02:01:40 am
Quote from: cacb on September 01, 2014, 09:41:38 pm
Quote from: mynameis on September 01, 2014, 06:48:13 am
i fonund that no ubuntu series  packages for a long time

I am on Kubuntu, and I found it convenient to build C::B from Jens Lody's preconfigured source tarball. Works every time.
Just follow the few simple steps at http://forums.codeblocks.org/index.php/topic,18580.msg127254.html#msg127254

thanks a lot,it seems a little complex for me(a Linux learner),but i will try it :)
Title: Re: The 30 August 2014 build (9884) is out.
Post by: RomanV on September 03, 2014, 11:03:48 am
Avast antivirus is complaining there is a virus/malware in this binary (for Windows).
Title: Re: The 30 August 2014 build (9884) is out.
Post by: Jenna on September 03, 2014, 11:11:31 am
Quote from: RomanV on September 03, 2014, 11:03:48 am
Avast antivirus is complaining there is a virus/malware in this binary (for Windows).
Search the forum.
My (very personal) opinin:
there is a malware on your system called avast.
Title: Re: The 30 August 2014 build (9884) is out.
Post by: eckard_klotz on September 03, 2014, 12:44:42 pm
Hello Everybody.

I use Norton 360 from Symantec as virus-scanner and I can confirm that you get an arlarm with a nightly. And to be honest Roman and I are not the only people facing this problem. If you search through your forum, you will find some more comments around this.

My Norton calls the reason of detection "heuristic" and it is associated with the file "CbLauncher.exe" in the archive "CB_YYYYMMDD_revXXXX_win32.7z". Since this heuristic warning comes with every nightly since I use Norton 360 (end of last year), I replaced date by YYYYMMDD and revision-number by XXXX. I don't realy know, what "heuristic" realy means. But I think it means that the virus-scanner tries to estimate if a program may be a problem. This means not that the programm is realy corrupted but the scanner is not able to exclude the possibility.

Symantec offers to register such cases to avoid detecting trustable programs as a problem on its webside. Therfore I had to give them the associated download location. For some reasons (perhaps a problem between my ears whyle filling the online formular) they where not able to download. I tried to send them some more details but until now they didn't react. If you are intersted, I may send you the content of the emails. Perhaps we find to gether what I did wrong (except sending them a file they deteted as a potential virus).

How ever, in my case the problem is the file "CbLauncher.exe". When I download a nightly at my company, I have no problem since we use an other virus-scanner. Thus I unpack the download there and put it on a stick. In this unpacked state I can transfer it to my computer at home, where "CbLauncher.exe" will be still deleted by my Norton-360. But as I see everyday, for the normal use "CbLauncher.exe" is not neccessary. Thus I would propose to offer "CbLauncher.exe" in an own package as long as this virus-scanner problem is not solved.

Best regards,
Eckard.
Title: Re: The 30 August 2014 build (9884) is out.
Post by: SteelRat on September 03, 2014, 02:01:03 pm
Stupid heuristic analyzer. All anti-viruses are useless shit. Use firewalls and java-script blockers in browsers and you'll never catch a virus.
Title: Re: The 30 August 2014 build (9884) is out.
Post by: eckard_klotz on September 03, 2014, 05:36:39 pm
Hello SteelRat.

 ??? Stupid ???

I don't know. If you have to maintain a tool like a virus-scanner it is hard to define exactly that an unknown programm that looks in some parts similar to a virus is not a virus. I think there is no other posibility to register the software (or to let it register) to become the status as known what means trusted software. Since Code::Blocks is an open source project this should be no normaly no problem.

Best regards,
                  Eckard.
Title: Re: The 30 August 2014 build (9884) is out.
Post by: SteelRat on September 03, 2014, 05:52:40 pm
Yep, but it's nightly build, not official release. It does not need to be registered anywhere.
Title: Re: The 30 August 2014 build (9884) is out.
Post by: eckard_klotz on September 03, 2014, 07:32:13 pm
Hello SteelRat.

Since Code::Blocks is freeware it has to be registered no where even it is an oficial release.

To register it in a list of an virus-scanner provider as "False Positive detected" means to give him a chance to update his product. Otherwise every user with this virus-scanner may have problems to download a nightly, since it is not posible to download it without the "CbLauncher.exe".

I think today there are good reasons to have a virus-scanner. And I think it is not realy a good idea to skip it since it has problems with only one application.

Regards,
            Eckard.
Title: Re: The 30 August 2014 build (9884) is out.
Post by: oBFusCATed on September 03, 2014, 10:45:27 pm
Quote from: jens on September 01, 2014, 10:28:18 pm
It should work if you remove the line completely from debian/control.
g++ should depend on the correct libstdc++-dev automatically,
and g++ is a dependency of the build-essential-package which is an automatic dependency of the debian build-system.

I don't know if it also works for older revisions of debian-based distro's, but on wheezy it seems to work.
It worked, but it didn't pick the revision number correctly...
Title: Re: The 30 August 2014 build (9884) is out.
Post by: RomanV on September 05, 2014, 06:44:33 am
Quote from: jens on September 03, 2014, 11:11:31 am
Quote from: RomanV on September 03, 2014, 11:03:48 am
Avast antivirus is complaining there is a virus/malware in this binary (for Windows).
Search the forum.
My (very personal) opinin:
there is a malware on your system called avast.
I agree. I personally think it's False Positive result of Avast's heuristics. Avast only complained about this build. I usually install all nightly builds. And it was the first time Avast complained about the build.

But Avast itself is not top-level antivirus. I have many friends in IT security field which do not seriously think about Avast.
I posted my initial message because I just wanted other people to know about it to Code::Blocks community. Because I think some other users of Code::Blocks also use Avast and they may encounter the same false positive result of the scan.
Title: Re: The 30 August 2014 build (9884) is out.
Post by: White-Tiger on September 05, 2014, 01:18:18 pm
hey... avast! is quite good ;) There aren't that much more AV's that are better (and free)
And what is more important to me then the count of viruses/malware it detects is simply the performance. And avast! is also good with that one. Its features and feeling is quite good ;) (or at least was, it's been a few years since I last used an AV or software firewall)

And complaining about an AV because of false positives is also not a really good idea. They at least increase your security to some point ;) Plus you can easily bypass them if you're really sure everything is ok. Just use your AV right.
It's also the user of the AV who needs to take actions and report that false positive, not the developer of the software. Because you've got the AV, you've got the tools to report the false positive not the dev. (well some offer an online form without using their AV, but others don't. Also, why should a dev really care if only 1 AV got problems he didn't directly cause?)

P.S. avast! doesn't seem to report it anymore: https://www.virustotal.com/en/file/c892433b6092890716e76ee662877eb566ee43fdafc6e821ec9f602364c3f0ce/analysis/1409915475/
Title: Re: The 30 August 2014 build (9884) is out.
Post by: eckard_klotz on September 07, 2014, 06:32:49 pm
Hello Everybody.

I agree that the user of the anti virus software has to report the "false positive" detection to his av-scanner provider. But I think it would be helpful if there would be a topic in the forum of code::blocks, that can be used to post information about this. It may be useful, if every kind of anti virus software has its own sub-topic. Other users can see, what is already reported if the reporter posts the ticket-number.

In my case I found in the forum the post http://forums.codeblocks.org/index.php/topic,19182.0.html (http://forums.codeblocks.org/index.php/topic,19182.0.html), where I learned that afb45 already reported a similar "false positive" detection under the ticked number " submission [3491738]" in April. But his detection reported the detection of "Trojan.Gen.SMH" while in my case the "Suspicious.Cloud.7.F" was detected. For some reasons my first report to Symantec under the ticket-number "submission (3590276)" last month was not successful. Thus I reported it new today under the ticket-number "submission (3613580)".

I hope this information is helpful for other users of Symantec which have a similar problem.

Best regards,
                   Eckard.

Title: Re: The 30 August 2014 build (9884) is out.
Post by: stahta01 on September 07, 2014, 07:53:06 pm
Quote from: eckard_klotz on September 07, 2014, 06:32:49 pm
Hello Everybody.

I agree that the user of the anti virus software has to report the "false positive" detection to his av-scanner provider. But I think it would be helpful if there would be a topic in the forum of code::blocks, that can be used to post information about this. It may be useful, if every kind of anti virus software has its own sub-topic. Other users can see, what is already reported if the reporter posts the ticket-number.

In my case I found in the forum the post http://forums.codeblocks.org/index.php/topic,19182.0.html (http://forums.codeblocks.org/index.php/topic,19182.0.html), where I learned that afb45 already reported a similar "false positive" detection under the ticked number " submission [3491738]" in April. But his detection reported the detection of "Trojan.Gen.SMH" while in my case the "Suspicious.Cloud.7.F" was detected. For some reasons my first report to Symantec under the ticket-number "submission (3590276)" last month was not successful. Thus I reported it new today under the ticket-number "submission (3613580)".

I hope this information is helpful for other users of Symantec which have a similar problem.

Best regards,
                   Eckard.



I think it would be better to have only a single thread on the subject.
And, have a Wiki page on the subject with sub-pages to the Wikipage as needed.

But, I am NOT a CB Dev; they are the ones to decide.
And, I have forgot the little I learned on doing Wiki pages; so, I have no plans to start a Wiki page.
I would think subpages for each DLL or EXE in CB Blocks would be nice with page sections for each AV.

Tim S.
Title: Re: The 30 August 2014 build (9884) is out.
Post by: eckard_klotz on September 08, 2014, 04:04:44 pm
Hello Everybody.

This time I was successfull. Symantec answered:

Quote
In relation to submission [3613580].

Upon further analysis and investigation we have verified your submission and, as such, the detection(s) for the following file(s) will be removed from our products:

   854E5D01E60235E3ACFA0AFAD2AADC36 - cblauncher.exe


The updated detection(s) will be distributed in the next set of virus definitions, available via LiveUpdate or from our website at http://securityresponse.symantec.com/avcenter/defs.download.html

Decisions made by Symantec are subject to change if alterations to the Software are made over time or as classification criteria and/or the policy employed by Symantec changes over time to address the evolving landscape.

If you are a software vendor, why not take part in our whitelisting program?
To participate in this program, please complete the following form: https://submit.symantec.com/whitelist

First of all they agree that this was realy a wrong detection.

But second, they offer a more longlasting posibility to avoid such problems in the futute if you follow the link "https://submit.symantec.com/whitelist" you will reach a dialog that allows you to register Code::Blocks and its components in their white-list. But since this establishes a connection between symantec and the C::B project using this posibility is not the decision of a user it is a decision of the developers.

Thus, dear deveoplers of Code::Blocks please think about. As far as I understand it, with every new virus that behaves in some aspects like a component of Code::Blocks we the user have to report a new wrong detection again. With this offer you have the chance to make Code::Blocks known.

Best regards,
                  Eckard.



Title: Re: The 30 August 2014 build (9884) is out.
Post by: oBFusCATed on September 08, 2014, 07:12:22 pm
Bleh... so we'll have to upload every cb release we do to every av software vendor?
Title: Re: The 30 August 2014 build (9884) is out.
Post by: eckard_klotz on September 09, 2014, 10:43:26 am
Hello C::B Developers.

Quote
... so we'll have to upload every cb release we do to every av software vendor? ...

Some body has to do. OK, I agree that it is a great effort for the project to inform all possible av software publisher about every new nightly. And it may be easier, if the user is doing this to share this effort. But I still think that this topic should be supported with an own sub-forum to give us useres a central place to share the information, what files and realeases are allready reported to which vendor and with what result.


Best regards,
                  Eckard.
Title: Re: The 30 August 2014 build (9884) is out.
Post by: White-Tiger on September 09, 2014, 12:54:08 pm
Actually, uploading it to Google's VirusTotal should be enough :P
One of the benefits of VirusTotal are that results are shared among AV companies. This means if someone detects something but others do not, it's possible for those others to get samples and fix their detection... Well this was meant for non detected viruses, but I'm certain that it might also work with false positives^^

But I don't know how automatic that all works... or if false positives are handled at all... Still maybe better then nothing :P

Otherwise, uploading CB to every AV vendor on every release (nightly or not) can only be done by having it done automated... and even then it requires a lot of time (to upload it) unless you're using a server to do that :P

Anyway, I'm still saying the user is responsible for his AV, and every user should be able to handle false positives anyway or they should use a different AV. It's their PC that has a problem with it ;)

P.S. you'll never know if a report is indeed a false positive ;) Because the developers PC could be corrupted or the upload somehow was.. so best is to use VirtusTotal if unsure and then... well guessing if you want to trust it if only 1 or 2 report it and others don't... could be still infected :P
And false positives don't disappear by the first report, to have a false positive to disappear a lot of people have to report it. Why should the AV company trust the first one to report it? Why should it really be a false positive? You only know if you've got enough data.
I don't even think they'll trust developers blindly, it will just give them a hint.. also note their note about signing the executables... that costs a lot of money which Microsoft wants to receive just to have it signed. So not even near possible for Code::Blocks.
Title: Re: The 30 August 2014 build (9884) is out.
Post by: RomanV on September 09, 2014, 03:10:42 pm
BTW I submitted this build to VirusTotal. See the result:
https://www.virustotal.com/en/file/b187b0b7cf24dc67740f5e8d844bd0d43e6f81ecf7590a6630b9a0cf2b4d39bc/analysis/1410267798/

So it's clean. Strange, but even Avast shows it as clean. While on desktop Avast (with latest updates) shows:
Infection blocked

    URL
    hxxp://softlayer-dal.dl.sourceforge.net/project/codeblocks/Binaries/Nightlies/2014/CB_20140830_rev9884_win32.7z|CbLauncher.exe

    Infection
    Win32:Evo-gen [Susp]

Title: Re: The 30 August 2014 build (9884) is out.
Post by: eckard_klotz on September 09, 2014, 03:30:00 pm
Hello White-Tiger

Quote
And false positives don't disappear by the first report, to have a false positive to disappear a lot of people have to report it. Why should the AV company trust the first one to report it? Why should it really be a false positive? You only know if you've got enough data.

If you report a false positive detection you also have the podibility to upload the effacted file. Thats how I did it.

Best regards,
                   Eckard.

PS.: How ever, I miss in this discussion the comments of the developers and/or the forum admin. Even if I accept, that I have to do the report and upload for my av-software, I still think an own sub-forum to post information about this for other users would be a great help for all of us.
Title: Re: The 30 August 2014 build (9884) is out.
Post by: White-Tiger on September 09, 2014, 04:54:25 pm
Quote from: RomanV on September 09, 2014, 03:10:42 pm
[...]
So it's clean. Strange, but even Avast shows it as clean. While on desktop Avast (with latest updates) shows:
Infection blocked
[...]
well... avast! doesn't just do some on-access scan as most simple and free AV's do, it also checks HTTP,SMTP, etc. traffic and intercepts things before they even arrive on your HDD / PC
And I guess it's basically blocking the URL... so if you had the file on your local PC, it wouldn't even complain...

It's actually weird that it behaves this differently... and it looks like the traffic filter isn't up-to-date^^

Anyway, did you try to manually check for updates for avast!? Maybe you're even using an older version :P VirusTotal is kinda up-to-date in that regard
Title: Re: The 30 August 2014 build (9884) is out.
Post by: oBFusCATed on September 09, 2014, 06:55:02 pm
Quote from: eckard_klotz on September 09, 2014, 03:30:00 pm
PS.: How ever, I miss in this discussion the comments of the developers and/or the forum admin.
You've asked for it: I wouldn't have bothered to do this even if I was using windows... I'm not so I can't care less.

If you don't trust us, then build everything from the sources...
Title: Re: The 30 August 2014 build (9884) is out.
Post by: eckard_klotz on September 09, 2014, 09:18:49 pm
Hello Developers.

Quote
If you don't trust us, then build everything from the sources...

For me this is not question of trusting your work. If I would not trust you I would not user Code::Blocks.  Don't ask me why av-tools detect C::B parts as potential viruses. Furthermore if you report them the "false positive" with an upload of the binary detected as suspicious, they agree that it was a "false positive" and set it on their white-list until they start to search for a new virus that behaves like a part of C::B. Then you get the next "false positive" and a new report has to be done.

Even this is a never ending story, we have to deal with it and this means when ever a "false positive" detection occures, somebody has to report it and share the report as well as the result with the community. The reason why I ask for a sub-forum is, that this makes it easier for the useres to share information about reporting "fasle positive" detections. If we the user have to do it, it would be extremly helpful to know if sombody already has done it for a specific detection-case and for a specific av-tool. An other advantage may be that this gives us the chance to collect some historical data. Symantec offers the posibility to add some additional information. Thus I noted down all available ticket-numbers I found here in the forum associated with cblauncher in the hope to animate them to find a more longlasting solution.

I don't want to blame you for the fact that av-tools detect C::B bianries as suspicious. My intention is to help you and other useres to deal in a constuctive way with this wrong detections.

Best regards,
                   Eckard.

Title: Re: The 30 August 2014 build (9884) is out.
Post by: oBFusCATed on September 10, 2014, 01:09:47 am
I don't see that much demand for a sub-forum, you can start a topic about this and if people start posting in it then we can pin it at the top.

Also, keep in mind that past experiences have shown that most users fail to find what they need (either without even searching for it or either because the forum's search functionality is not very good) and instead start a new topic at a random sub-forum.
Title: Re: The 30 August 2014 build (9884) is out.
Post by: RomanV on September 10, 2014, 08:29:03 am
Quote from: White-Tiger on September 09, 2014, 04:54:25 pm
well... avast! doesn't just do some on-access scan as most simple and free AV's do, it also checks HTTP,SMTP, etc. traffic and intercepts things before they even arrive on your HDD / PC
And I guess it's basically blocking the URL... so if you had the file on your local PC, it wouldn't even complain...

It's actually weird that it behaves this differently... and it looks like the traffic filter isn't up-to-date^^

Anyway, did you try to manually check for updates for avast!? Maybe you're even using an older version :P VirusTotal is kinda up-to-date in that regard
Did you read my message? I said I was using Avast with latest updates (program updates and virus definitions). The result is not only because of traffic interception. I turned off live shield of Avast. downloaded build's file (it was saved on my drive). Then I turned on shield. So I got the same message from Avast.

I repeat: I think it's false positive.
In reality what I did: it decided the following: I skip this build. I will wait for another build. That's all. Because I don't have time to find out how to turn off this Avast false positive. I should disable something in Avast heuristics I guess. But it may be useful in some situations. So I will wait to the next build.

And I think developers of Code::Blocks do not have to spend their time and submit nightly builds to different virus check sites. Nightly build - as it's said "nightly build". Cautious users must use "yearly" build or something. And for these rare build maybe virus check makes sense. But it's not the case for nightly builds. It's better to spend more time on development of new features or issue corrections.
Title: Re: The 30 August 2014 build (9884) is out.
Post by: eckard_klotz on September 11, 2014, 11:59:05 am
Hello Everybody.

Just to bring it to an end (or better temporary end):

I don't want to stress the points who has to do what and why any more since I learned in the last days that this is not really helpful as long as it does not result into a constructive solution. But pleas allow me to post the what I was able to do to solve my current anti virus problem just as an example for others and I hope that this is what all others expect from me if they write about things which are not the job of the project but by the users.


What did I learn:


RomanV, you decided to wait for the next nightly. As long nobody reports the false positive to your anti virus software provider this will not help you. Since your scanner does not detect a real infection of a file but only an alikeness with an infected one. This alikeness is part of the strategy used by the developer of a real virus to mask his bad software. Thus the best you can do, is finding out how you can report the false positive detection together with an upload of the file detected as suspicious. If you don't do it, who else should.

Best regards,
                  Eckard.

PS.: Dear Code::Blocks team. For some reasons I have the feeling that stressing this special problem is not really welcome what I can understand. Thus please accept my apologize if my posts may contain  unreasonable expectations or something like this. This was not my intention.
 
Title: Re: The 30 August 2014 build (9884) is out.
Post by: cacb on September 14, 2014, 12:17:37 pm
Based on Jens' source tarball found here (http://apt.jenslody.de/testing/pool/main/c/codeblocks/) , I have successfully built this Nightly (svn9884) under raspbian for the Raspberry PI. It took around 14 hours to complete, but it worked  :) It seems to me that this nightly runs faster than the old nightly that I got from someone else in January. I have not found up to date builds of C::B for raspbian elsewhere.

I'd be willing to upload the binaries to some place where others can access it, assuming that I don't break any rules by doing that. Please advice if you are interested.

Title: Re: The 30 August 2014 build (9884) is out.
Post by: edison on September 15, 2014, 06:20:51 pm
LLVM Clang does not support -mwindows any more, so please update the switch of GUI Application in options_clang.xml:

from:
value="$linker $libdirs -o $exe_output $link_objects $link_resobjects $link_options $libs -mwindows"/>

to:
value="$linker $libdirs -o $exe_output $link_objects $link_resobjects $link_options $libs -Wl,--subsystem,windows"/>
Title: Re: The 30 August 2014 build (9884) is out.
Post by: oBFusCATed on September 15, 2014, 09:23:11 pm
Patch?
Title: Re: The 30 August 2014 build (9884) is out.
Post by: edison on September 16, 2014, 02:08:55 am
Quote from: oBFusCATed on September 15, 2014, 09:23:11 pm
Patch?

Sorry, I dont how to make a patch(I am a rookie here), but the file is locale here:
src\devel\share\CodeBlocks\compilers\options_clang.xml

The content need to change is in line 188.

update:
src\plugins\compilergcc\options_clang.xml
Title: Re: The 30 August 2014 build (9884) is out.
Post by: oBFusCATed on September 16, 2014, 10:58:00 am
http://wiki.codeblocks.org/index.php?title=Creating_a_patch_to_submit_to_BerliOS_%28Patch_Tracker%29
Title: Re: The 30 August 2014 build (9884) is out.
Post by: edison on September 16, 2014, 03:22:50 pm
Quote from: oBFusCATed on September 16, 2014, 10:58:00 am
http://wiki.codeblocks.org/index.php?title=Creating_a_patch_to_submit_to_BerliOS_%28Patch_Tracker%29

I don't know how to use svn ... :

Quote
G:\cb_svn\src\plugins\compilergcc\resources\compilers>svn add options_clang.xml
svn: warning: W150002: 'G:\cb_svn\src\plugins\compilergcc\resources\compilers\op
tions_clang.xml' is already under version control
svn: E200009: Could not add all targets because some targets are already version
ed
svn: E200009: Illegal target for the requested operation
Title: Re: The 30 August 2014 build (9884) is out.
Post by: ollydbg on September 16, 2014, 03:51:05 pm
Quote from: edison on September 16, 2014, 03:22:50 pm
Quote from: oBFusCATed on September 16, 2014, 10:58:00 am
http://wiki.codeblocks.org/index.php?title=Creating_a_patch_to_submit_to_BerliOS_%28Patch_Tracker%29

I don't know how to use svn ... :

Quote
G:\cb_svn\src\plugins\compilergcc\resources\compilers>svn add options_clang.xml
svn: warning: W150002: 'G:\cb_svn\src\plugins\compilergcc\resources\compilers\op
tions_clang.xml' is already under version control
svn: E200009: Could not add all targets because some targets are already version
ed
svn: E200009: Illegal target for the requested operation
I think running the command "svn diff > my.patch" should be enough to create the patch file.
Title: Re: The 30 August 2014 build (9884) is out.
Post by: edison on September 16, 2014, 04:10:52 pm
Thanks, it works.

Can you review and submit it ?

Code
Index: options_clang.xml
===================================================================
--- options_clang.xml	(revision 9916)
+++ options_clang.xml	(working copy)
@@ -185,7 +185,7 @@
         <Command name="LinkNative"
                  value="$linker $libdirs -o $exe_output $link_objects $link_resobjects $link_options $libs -Wl,--subsystem,native"/>
         <Command name="LinkExe"
-                 value="$linker $libdirs -o $exe_output $link_objects $link_resobjects $link_options $libs -mwindows"/>
+                 value="$linker $libdirs -o $exe_output $link_objects $link_resobjects $link_options $libs -Wl,--subsystem,windows"/>
         <Command name="LinkDynamic"
                  value="$linker -shared -Wl,--output-def=$def_output -Wl,--out-implib=$static_output -Wl,--dll $libdirs $link_objects $link_resobjects -o $exe_output $link_options $libs"/>
     </if>
Title: Re: The 30 August 2014 build (9884) is out.
Post by: ollydbg on September 17, 2014, 07:44:28 am
Quote from: oBFusCATed on September 16, 2014, 10:58:00 am
http://wiki.codeblocks.org/index.php?title=Creating_a_patch_to_submit_to_BerliOS_%28Patch_Tracker%29
Update this wiki page to point to SF.
Title: Re: The 30 August 2014 build (9884) is out.
Post by: oBFusCATed on October 22, 2014, 10:27:42 pm
@edison: Which version of clang is the last one that supported the old switch?
@alpha: Can we make this parse the version of clang and then pass the appropriate option?
Title: Re: The 30 August 2014 build (9884) is out.
Post by: stahta01 on October 23, 2014, 04:57:46 am
Quote from: oBFusCATed on October 22, 2014, 10:27:42 pm
@edison: Which version of clang is the last one that supported the old switch?
@alpha: Can we make this parse the version of clang and then pass the appropriate option?

I think going with the first clang version to support the new switch is also an option.

Tim S.
Title: Re: The 30 August 2014 build (9884) is out.
Post by: Alpha on October 23, 2014, 06:23:51 am
Oops, must have missed this.  Yes, I should be able to add a quick regex.  Will do some testing this weekend to see what is necessary.
Title: Re: The 30 August 2014 build (9884) is out.
Post by: edison on October 23, 2014, 06:39:20 am
Quote from: oBFusCATed on October 22, 2014, 10:27:42 pm
@edison: Which version of clang is the last one that supported the old switch?

I don't know...

http://comments.gmane.org/gmane.comp.compilers.clang.devel/38114

Reid Kleckner said:
Quote
I doubt we ever recognized this argument.  This probably changed because we started rejecting unknown -m flags.
Title: Re: The 30 August 2014 build (9884) is out.
Post by: oBFusCATed on October 23, 2014, 07:16:51 am
Quote from: Alpha on October 23, 2014, 06:23:51 am
Oops, must have missed this.  Yes, I should be able to add a quick regex.  Will do some testing this weekend to see what is necessary.
Just keep in mind that dumpversion doesn't do anything useful - always returns 4.2.1 (at least on linux and osx)
Title: Re: The 30 August 2014 build (9884) is out.
Post by: Alpha on October 30, 2014, 02:32:45 am
From my experimentation, it seems neither of these flags are necessary.
Code
         <Command name="LinkExe"
-                 value="$linker $libdirs -o $exe_output $link_objects $link_resobjects $link_options $libs -mwindows"/>
+                 value="$linker $libdirs -o $exe_output $link_objects $link_resobjects $link_options $libs -Wl,--subsystem,windows"/>

I will soon commit the equivalent of (for all versions of clang)
Code
        <Command name="LinkExe"
                 value="$linker $libdirs -o $exe_output $link_objects $link_resobjects $link_options $libs"/>
if there are no reported issues with it.
Title: Re: The 30 August 2014 build (9884) is out.
Post by: edison on October 30, 2014, 09:40:33 am
Quote from: Alpha on October 30, 2014, 02:32:45 am
I will soon commit the equivalent of (for all versions of clang)
Code
        <Command name="LinkExe"
                 value="$linker $libdirs -o $exe_output $link_objects $link_resobjects $link_options $libs"/>
if there are no reported issues with it.

This will cause a console window below the gui window, make it not a "GUI application".
Title: Re: The 30 August 2014 build (9884) is out.
Post by: Alpha on October 30, 2014, 05:57:56 pm
Quote from: edison on October 30, 2014, 09:40:33 am
This will cause a console window below the gui window, make it not a "GUI application".
Oh dear... flawed test cases.  I am so embarrassed. :-[
I think I must be spending too much time under linux.

(Thank you.)