Re: Wrong method source pointer in Squeak5.2

I forgot to report, on Windows I experienced problem with stdout during Seaside install, because writeLimit inst. var. was set to nil.

Initializing the ivar to zero and restarting the execution somewhere in the execution stack did make it.

Le lun. 17 déc. 2018 à 15:03, Nicolas Cellier <[hidden email]> a écrit :

Hi Max,

I can reproduce on windows via cygwin shell (I have to export SHELLOPTS; set -o igncr  like in https://stackoverflow.com/questions/42072239/how-to-set-up-cygwin-to-use-dos-file-endings)

The Notifier opens, but I can't open the Debugger... It seems blocked on an AccessLock.

I have replaced

DebuggerMethodMap>>protected: aBlock
    "^(AccessLock ifNil: [ AccessLock := Mutex new ]) critical: aBlock"
    ^aBlock value

and then the test passed instead of failing...

I reverted above method to original, and now the test still passes...

Could it be that .changes file is opened with wrong encoding (not UTF8) thru some path, but any action re-opening the .changes with correct encoding makes it an Eisenbug?

Le lun. 17 déc. 2018 à 02:39, David T. Lewis <[hidden email]> a écrit :

Hmmm, this appears to be some kind of issue with SSL authorization
when accessing the github repository. It seems to be failing in

  MCGitHubReposistory class((MCGitBasedNetworkRepository class)>>projectDirectoryFrom:version:

from

  MCGitHubReposistory(MCGitBasedNetworkRepository)>>directory

from

  MCGitHubRepository(MCFileTreeRepository)>>allFileNames

and it is caused by a primitive failure in SqueakSSL>>primitiveSSLCreate.

Evaluating "SqueakSSL new" gives me the same error. So the problem is
probably something to do with the VM I am running, or with the runtime
packages that I have installed on my PC. I will follow up on this later.

Thanks,
Dave


On Sun, Dec 16, 2018 at 08:07:16PM -0500, David T. Lewis wrote:
> Hi Max,
>
> I'm on ubuntu 16.04 LTS.
>
> But it's probably some mistake that I am making. One thing I noticed - when
> I do the checkout of the Seaside version, I end up with a detached head, is
> this expected? Here is what I see on the terminal:
>
>   lewis@lewis-Gazelle-Pro:~/squeak/Squeak5.2/Seaside-TEST/Seaside$ git status
>   On branch master
>   Your branch is up-to-date with 'origin/master'.
>   nothing to commit, working directory clean
>   lewis@lewis-Gazelle-Pro:~/squeak/Squeak5.2/Seaside-TEST/Seaside$ git checkout -f master
>   Already on 'master'
>   Your branch is up-to-date with 'origin/master'.
>   lewis@lewis-Gazelle-Pro:~/squeak/Squeak5.2/Seaside-TEST/Seaside$ git checkout 9cb54a7b14cd254ef318294905c4e8dda8dd9f79
>   Note: checking out '9cb54a7b14cd254ef318294905c4e8dda8dd9f79'.
>   
>   You are in 'detached HEAD' state. You can look around, make experimental
>   changes and commit them, and you can discard any commits you make in this
>   state without impacting any branches by performing another checkout.
>   
>   If you want to create a new branch to retain commits you create, you may
>   do so (now or later) by using -b with the checkout command again. Example:
>   
>     git checkout -b <new-branch-name>
>   
>   HEAD is now at 9cb54a7... Merge pull request #1071 from jecisc/1070-Seaside-Tests-REST-is-never-referenced
>   lewis@lewis-Gazelle-Pro:~/squeak/Squeak5.2/Seaside-TEST/Seaside$
>
> I may not have time to follow up on this in the next couple of days,
> but the problem that you spotted does sound serious. It suggests that
> the changes file is getting corrupted, and that is not supposed to
> be possible.
>
> Dave
>
>
>
> On Sun, Dec 16, 2018 at 11:04:11PM +0100, Max Leske wrote:
> > (resending because of wrong subject, sorry)
> >
> > Hi Dave,
> >
> > I just retested loading in Ubuntu 17.04 (32-bits) without problems.
> > Maybe you didn't checkout the commit I referenced? I realise there's a
> > `cd Seaside` missing before the `git checkout`.
> >
> > What's the load error about? What system are you using exactly?
> >
> >
> > Cheers,
> > Max
> >
> > On 16 Dec 2018, at 20:12, [hidden email]
> > wrote:
> >
> > >I'm running into an error related to Metacello loading when I do the
> > >run.sh script in the recipe below, screen shot attached. I am running
> > >on Linux.
> > >
> > >It looks like I am missing something in the intial setup, can you give
> > >me a pointer?
> > >
> > >Thanks.
> > >Dave
> > >
> > >On Sun, Dec 16, 2018 at 06:14:47PM +0100, Max Leske wrote:
> > >
> > >>Hi,
> > >>
> > >>We have an issue in Seaside where a loaded test method holds a bad
> > >>source pointer. We only noticed because this method accesses the
> > >>debugger map.
> > >>
> > >>How to reproduce:
> > >>
> > >>1. clone SmalltalkCI:
> > >>  git clone [hidden email]:hpi-swa/smalltalkCI.git
> > >>2. clone Seaside:
> > >>  git clone [hidden email]:SeasideSt/Seaside.git;
> > >>  git checkout 9cb54a7b14cd254ef318294905c4e8dda8dd9f79
> > >>3. install Seaside in Squeak5.2:
> > >>  <path to SmalltalkCI>/run.sh --headful -s Squeak-5.2 <path to
> > >>Seaside>/.smalltalk.ston
> > >>4. Run the test WAPharoDebuggerTest>>testNamedTempAt and you'll see a
> > >>debugger pop up for an UndeclaredVariable. The source lookup is
> > >>performed in the changes file but in the middle of a chunk of binary
> > >>(font) data.
> > >>
> > >>I can can only speculate that some of that binary data introduces
> > >>random
> > >>chunks which messes with the offsets.
> > >>
> > >>We're adding a workaround for that particular test for now.
> > >>
> > >>
> > >>Let me know if I can help track down the issue.
> > >>
> > >>Cheers,
> > >>Max
>
> >

Le lun. 17 déc. 2018 à 15:06, Nicolas Cellier <[hidden email]> a écrit :

I forgot to report, on Windows I experienced problem with stdout during Seaside install, because writeLimit inst. var. was set to nil.

Initializing the ivar to zero and restarting the execution somewhere in the execution stack did make it.

Ah no, it was not enough, I also did (self reopen) before being able to successfully proceed installation...

Le lun. 17 déc. 2018 à 15:03, Nicolas Cellier <[hidden email]> a écrit :

Hi Max,

I can reproduce on windows via cygwin shell (I have to export SHELLOPTS; set -o igncr  like in https://stackoverflow.com/questions/42072239/how-to-set-up-cygwin-to-use-dos-file-endings)

The Notifier opens, but I can't open the Debugger... It seems blocked on an AccessLock.

I have replaced

DebuggerMethodMap>>protected: aBlock
    "^(AccessLock ifNil: [ AccessLock := Mutex new ]) critical: aBlock"
    ^aBlock value

and then the test passed instead of failing...

I reverted above method to original, and now the test still passes...

Could it be that .changes file is opened with wrong encoding (not UTF8) thru some path, but any action re-opening the .changes with correct encoding makes it an Eisenbug?

Le lun. 17 déc. 2018 à 02:39, David T. Lewis <[hidden email]> a écrit :

Hmmm, this appears to be some kind of issue with SSL authorization
when accessing the github repository. It seems to be failing in

  MCGitHubReposistory class((MCGitBasedNetworkRepository class)>>projectDirectoryFrom:version:

from

  MCGitHubReposistory(MCGitBasedNetworkRepository)>>directory

from

  MCGitHubRepository(MCFileTreeRepository)>>allFileNames

and it is caused by a primitive failure in SqueakSSL>>primitiveSSLCreate.

Evaluating "SqueakSSL new" gives me the same error. So the problem is
probably something to do with the VM I am running, or with the runtime
packages that I have installed on my PC. I will follow up on this later.

Thanks,
Dave


On Sun, Dec 16, 2018 at 08:07:16PM -0500, David T. Lewis wrote:
> Hi Max,
>
> I'm on ubuntu 16.04 LTS.
>
> But it's probably some mistake that I am making. One thing I noticed - when
> I do the checkout of the Seaside version, I end up with a detached head, is
> this expected? Here is what I see on the terminal:
>
>   lewis@lewis-Gazelle-Pro:~/squeak/Squeak5.2/Seaside-TEST/Seaside$ git status
>   On branch master
>   Your branch is up-to-date with 'origin/master'.
>   nothing to commit, working directory clean
>   lewis@lewis-Gazelle-Pro:~/squeak/Squeak5.2/Seaside-TEST/Seaside$ git checkout -f master
>   Already on 'master'
>   Your branch is up-to-date with 'origin/master'.
>   lewis@lewis-Gazelle-Pro:~/squeak/Squeak5.2/Seaside-TEST/Seaside$ git checkout 9cb54a7b14cd254ef318294905c4e8dda8dd9f79
>   Note: checking out '9cb54a7b14cd254ef318294905c4e8dda8dd9f79'.
>   
>   You are in 'detached HEAD' state. You can look around, make experimental
>   changes and commit them, and you can discard any commits you make in this
>   state without impacting any branches by performing another checkout.
>   
>   If you want to create a new branch to retain commits you create, you may
>   do so (now or later) by using -b with the checkout command again. Example:
>   
>     git checkout -b <new-branch-name>
>   
>   HEAD is now at 9cb54a7... Merge pull request #1071 from jecisc/1070-Seaside-Tests-REST-is-never-referenced
>   lewis@lewis-Gazelle-Pro:~/squeak/Squeak5.2/Seaside-TEST/Seaside$
>
> I may not have time to follow up on this in the next couple of days,
> but the problem that you spotted does sound serious. It suggests that
> the changes file is getting corrupted, and that is not supposed to
> be possible.
>
> Dave
>
>
>
> On Sun, Dec 16, 2018 at 11:04:11PM +0100, Max Leske wrote:
> > (resending because of wrong subject, sorry)
> >
> > Hi Dave,
> >
> > I just retested loading in Ubuntu 17.04 (32-bits) without problems.
> > Maybe you didn't checkout the commit I referenced? I realise there's a
> > `cd Seaside` missing before the `git checkout`.
> >
> > What's the load error about? What system are you using exactly?
> >
> >
> > Cheers,
> > Max
> >
> > On 16 Dec 2018, at 20:12, [hidden email]
> > wrote:
> >
> > >I'm running into an error related to Metacello loading when I do the
> > >run.sh script in the recipe below, screen shot attached. I am running
> > >on Linux.
> > >
> > >It looks like I am missing something in the intial setup, can you give
> > >me a pointer?
> > >
> > >Thanks.
> > >Dave
> > >
> > >On Sun, Dec 16, 2018 at 06:14:47PM +0100, Max Leske wrote:
> > >
> > >>Hi,
> > >>
> > >>We have an issue in Seaside where a loaded test method holds a bad
> > >>source pointer. We only noticed because this method accesses the
> > >>debugger map.
> > >>
> > >>How to reproduce:
> > >>
> > >>1. clone SmalltalkCI:
> > >>  git clone [hidden email]:hpi-swa/smalltalkCI.git
> > >>2. clone Seaside:
> > >>  git clone [hidden email]:SeasideSt/Seaside.git;
> > >>  git checkout 9cb54a7b14cd254ef318294905c4e8dda8dd9f79
> > >>3. install Seaside in Squeak5.2:
> > >>  <path to SmalltalkCI>/run.sh --headful -s Squeak-5.2 <path to
> > >>Seaside>/.smalltalk.ston
> > >>4. Run the test WAPharoDebuggerTest>>testNamedTempAt and you'll see a
> > >>debugger pop up for an UndeclaredVariable. The source lookup is
> > >>performed in the changes file but in the middle of a chunk of binary
> > >>(font) data.
> > >>
> > >>I can can only speculate that some of that binary data introduces
> > >>random
> > >>chunks which messes with the offsets.
> > >>
> > >>We're adding a workaround for that particular test for now.
> > >>
> > >>
> > >>Let me know if I can help track down the issue.
> > >>
> > >>Cheers,
> > >>Max
>
> >